GOOGLE SCHOLAR · AUTO-SYNCED

Publications

All 88 records currently listed on Google Scholar are included. Author annotations are curated separately from research-topic tags. When no corresponding author is explicitly identified, the final author is marked as corresponding; alphabetical and team-authored reports are intentionally left unannotated.

88 publications 17,299 citations 48 h-index 70 i10-index
First / Co-First Author Corresponding Author Project Lead

88/88 detail records captured 88/88 complete author lists

Last synchronized 2026-10-04 13:09 UTC.

Select year
88 of 88 publications All years

2026

OpenCompass: A Universal Evaluation Platform for Large Language Models

Evaluation InfrastructureLLM EvaluationBenchmark Design

Maosong Cao, Kai Chen, Haodong Duan, Yixiao Fang, Tong Gao, Ge Jiaye, Mo Li, Hongwei Liu, Junnan Liu, Yuan Liu, Chengqi Lyu, Han Lyu, Ningsheng Ma, Zerun Ma, Yu Sun, Zhiyong Wu, Linchen Xiao, Jun Xu, Haochen Ye, Zhaohui Yu, Yike Yuan, Songyang Zhang, Yufeng Zhao, Fengzhe Zhou, Peiheng Zhou, Dongsheng Zhu, Lin Zhu, Jingming Zhuo

Technical Report, 2026

Full metadata
Publication date
2026/5/19
Journal
Technical Report

Description

In recent years, the field of artificial intelligence has undergone a paradigm shift from task-specific small-scale models to general-purpose large language models (LLMs). With the rapid iteration of LLMs, objective, quantitative, and comprehensive evaluation of their capabilities has become a critical link in advancing technological development. Currently, the mainstream static benchmark dataset-based evaluation methods face challenges such as the diversity of task types, inconsistent evaluation criteria, and fragmentation of data and processing workflows, making it difficult to efficiently conduct cross-domain and large-scale model evaluation. To address the aforementioned issues, this paper proposes and open-sources OpenCompass, a one-stop, scalable, and high-concurrency-supported general-purpose LLM evaluation platform. Adhering to the design philosophy of modularization and component decoupling, the platform boasts three core advantages: high compatibility, flexibility, and high concurrency. The core architecture of OpenCompass comprises five key components: the Configuration System, Task Partitioning Module, Execution and Scheduling Module, Task Execution Unit, and Result Visualization Module. Its workflow provides rule-based, LLM-as-a-Judge, and cascaded evaluators to adapt to the requirements of different task scenarios. Supporting mainstream benchmark datasets across multiple domains, including knowledge, reasoning, computation, science, language, code, etc., the platform offers a unified and efficient LLM evaluation tool for both academia and industry, facilitating the accurate identification of strengths and …

Paper
572 citations

MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Spatial ReasoningMulti-Image ReasoningMultimodal Evaluation

Sihan Yang, Runsen Xu, Yiman Xie, Sizhe Yang, Mo Li, Jingli Lin, Chenming Zhu, Xiaochen Chen, Haodong Duan, Xiangyu Yue, Dahua Lin, Tai Wang, Jiangmiao Pang

ICLR, 2026

Full metadata
Publication date
2025/5/29
Journal
ICLR, 2026

Description

Spatial intelligence is essential for multimodal large language models (MLLMs) operating in the complex physical world. Existing benchmarks, however, probe only single-image relations and thus fail to assess the multi-image spatial reasoning that real-world deployments demand. We introduce MMSI-Bench, a VQA benchmark dedicated to multi-image spatial intelligence. Six 3D-vision researchers spent more than 300 hours meticulously crafting 1,000 challenging, unambiguous multiple-choice questions from over 120,000 images, each paired with carefully designed distractors and a stepwise reasoning process. We conduct extensive experiments and evaluate 37 open-source and proprietary MLLMs, observing a wide gap: the strongest open-source model attains roughly 30\% accuracy and OpenAI's GPT-5 reasoning model reaches 40\%, while humans score 97\%. These results underscore the challenging nature of MMSI-Bench and the substantial headroom for future research. Leveraging the annotated reasoning processes, we also provide an automated error analysis pipeline that diagnoses four dominant failure modes, including (1) grounding errors,(2) overlap-matching and scene-reconstruction errors,(3) situation-transformation reasoning errors, and (4) spatial-logic errors, offering insights for advancing spatial intelligence.

Paper Code
190 citations

Visualprm: An effective process reward model for multimodal reasoning

Reward ModelingModel Adaptation

Weiyun Wang, Zhangwei Gao, Lianjie Chen, Zhe Chen, Jinguo Zhu, Xiangyu Zhao, Yangzhou Liu, Yue Cao, Shenglong Ye, Xizhou Zhu, Lewei Lu, Haodong Duan, Yu Qiao, Jifeng Dai, Wenhai Wang

ICLR, 2026

Full metadata
Publication date
2025/3/13
Journal
ICLR, 2026

Description

We introduce VisualPRM, an advanced multimodal Process Reward Model (PRM) with 8B parameters, which improves the reasoning abilities of existing Multimodal Large Language Models (MLLMs) across different model scales and families with Best-of-N (BoN) evaluation strategies. Specifically, our model improves the reasoning performance of three types of MLLMs and four different model scales. Even when applied to the highly capable InternVL2.5-78B, it achieves a 5.9-point improvement across seven multimodal reasoning benchmarks. Experimental results show that our model exhibits superior performance compared to Outcome Reward Models and Self-Consistency during BoN evaluation. To facilitate the training of multimodal PRMs, we construct a multimodal process supervision dataset VisualPRM400K using an automated data pipeline. For the evaluation of multimodal PRMs, we propose VisualProcessBench, a benchmark with human-annotated step-wise correctness labels, to measure the abilities of PRMs to detect erroneous steps in multimodal reasoning tasks. We hope that our work can inspire more future research and contribute to the development of MLLMs. Our model, data, and benchmark are released in https://internvl.github.io/blog/2025-03-13-VisualPRM/.

Paper
166 citations

Seed1.8 model card: Towards generalized real-world agency

Agentic SystemsFoundation Models

Bytedance Seed

Seed Technical Reports, 2026

Full metadata
Publication date
2026/3/21
Journal
Seed Technical Reports, 2026

Description

We present Seed1.8, a foundation model aimed at generalized real-world agency: going beyond single-turn prediction to multi-turn interaction, tool use, and multi-step execution. Seed1.8 keeps strong LLM and vision-language performance while supporting a unified agentic interface-search, code generation and execution, and GUI interaction. For deployment, it offers latency- and cost-aware inference, including configurable thinking modes and optimized visual encoding for images and video. We report evaluations on standard benchmarks and application-aligned workflows spanning foundational skills, multimodal understanding, and agentic behavior. Seed1.8 is released to support further research and development on interactive, real-world use cases.

Paper
79 citations

LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?

Spatial ReasoningVisual ReasoningMultimodal Evaluation

Kexian Tang, Junyao Gao, Yanhong Zeng, Haodong Duan, Yanan Sun, Zhening Xing, Wenran Liu, Kaifeng Lyu, Kai Chen

ECCV, 2026

Full metadata
Publication date
2025/3/25
Journal
arXiv preprint arXiv:2503.19990

Description

Many real-world applications of spatial intelligence, such as robotic control, autonomous driving, and automated assembly, require spatial reasoning across multiple sequential steps. However, the extent to which current Multimodal Large Language Models (MLLMs) possess this capability remains largely unexplored. Inspired by LEGO construction, a recreational activity that critically relies on multi-step spatial reasoning, we introduce LEGO-Puzzles: a benchmark designed to systematically evaluate the spatial reasoning capabilities of MLLMs from basic spatial understanding to multi-step planning. LEGO-Puzzles contains two task sets. The Elementary set covers 11 visual question-answering (VQA) tasks with 1,100 carefully curated samples to test elementary spatial reasoning skills that are cruical for LEGO assembly. The Planning set directly requires the model to generate a step-by-step plan for assembling a target LEGO structure, where the tasks are organized into subsets with different planning horizons ranging up to 8. Our evaluation of 29 state-of-the-art MLLMs shows that even the strongest models struggle with elementary reasoning tasks in LEGO construction, falling at least 20% behind human performance. The planning accuracy also quickly drops to 0% as the number of planning steps increases, whereas our human participants solve all the tasks perfectly. Switching the output format from multiple choice to image generation degrades model performance even further, leading to zero accuracy even for planning 3 steps. Overall, LEGO-Puzzles reveals critical limitations in current MLLMs' spatial reasoning capabilities and highlights the …

Paper Code
58 citations

Mmbench-gui: Hierarchical multi-platform evaluation framework for gui agents

Agentic SystemsMultimodal EvaluationBenchmark Design

Xuehui Wang, Zhenyu Wu, JingJing Xie, Zichen Ding, Bowen Yang, Zehao Li, Zhaoyang Liu, Qingyun Li, Xuan Dong, Zhe Chen, Weiyun Wang, Xiangyu Zhao, Jixuan Chen, Haodong Duan, Tianbao Xie, Chenyu Yang, Shiqian Su, Yue Yu, Yuan Huang, Yiqian Liu, Xiao Zhang, Yanting Zhang, Xiangyu Yue, Weijie Su, Xizhou Zhu, Wei Shen, Jifeng Dai, Wenhai Wang

CVPR, 2026

Full metadata
Publication date
2025/7/25
Journal
CVPR, 2026

Description

We introduce MMBench-GUI, a hierarchical benchmark for evaluating GUI automation agents across Windows, macOS, Linux, iOS, Android, and Web platforms. It comprises four levels: GUI Content Understanding, Element Grounding, Task Automation, and Task Collaboration, covering essential skills for GUI agents. In addition, we propose a novel Efficiency-Quality Area (EQA) metric to assess GUI agent execution efficiency in online automation scenarios. Through MMBench-GUI, we identify accurate visual grounding as a critical determinant of overall task success, emphasizing the substantial benefits of modular frameworks that integrate specialized grounding modules. Furthermore, to achieve reliable GUI automation, an agent requires strong task planning and cross-platform generalization abilities, with long-context memory, a broad action space, and long-term reasoning playing a critical role. More important, task efficiency remains a critically underexplored dimension, and all models suffer from substantial inefficiencies, with excessive redundant steps even when tasks are ultimately completed. The integration of precise localization, effective planning, and early stopping strategies is indispensable to enable truly efficient and scalable GUI automation. Our benchmark code, evaluation data, and running environment will be publicly available at https://github.com/open-compass/MMBench-GUI.

Paper
53 citations

Wildclawbench: A benchmark for real-world, long-horizon agent evaluation

Agentic SystemsBenchmark Design

Shuangrui Ding, Xuanlang Dai, Long Xing, Shengyuan Ding, Ziyu Liu, Yang JingYi, Penghui Yang, Zhixiong Zhang, Xilin Wei, Xinyu Fang, Yubo Ma, Haodong Duan, Jing Shao, Jiaqi Wang, Dahua Lin, Kai Chen, Yuhang Zang

arXiv preprint arXiv:2605.10912, 2026

Full metadata
Publication date
2026/5/11
Journal
arXiv preprint arXiv:2605.10912

Description

Large language and vision-language models increasingly power agents that act on a user's behalf through command-line interface (CLI) harnesses. However, most agent benchmarks still rely on synthetic sandboxes, short-horizon tasks, mock-service APIs, and final-answer checks, leaving open whether agents can complete realistic long-horizon work in the runtimes where they are deployed. This work presents WildClawBench, a native-runtime benchmark of 60 human-authored, bilingual, multimodal tasks spanning six thematic categories. Each task averages roughly 8 minutes of wall-clock time and over 20 tool calls, and runs inside a reproducible Docker container hosting an actual CLI agent harness (OpenClaw, Claude Code, Codex, or Hermes Agent) with access to real tools rather than mock services. Grading is hybrid, combining deterministic rule-based checks, environment-state auditing of side effects, and an LLM/VLM judge for semantic verification. Across 19 frontier models, the best, Claude Opus 4.7, reaches only 62.2% overall under OpenClaw, while every other model stays below 60%, and switching harness alone shifts a single model by up to 18 points. These results show that long-horizon, native-runtime agent evaluation remains a far-from-resolved task for current frontier models. We release the tasks, code, and containerized tooling to support reproducible evaluation.

Paper
50 citations

Spatial-ssrl: Enhancing spatial understanding via self-supervised reinforcement learning

Spatial ReasoningReinforcement LearningSelf-Supervised Learning

Yuhong Liu, Beichen Zhang, Yuhang Zang, Yuhang Cao, Long Xing, Xiaoyi Dong, Haodong Duan, Dahua Lin, Jiaqi Wang

CVPR, 2026

Full metadata
Publication date
2025/10/31
Journal
CVPR, 2026

Description

Spatial understanding remains a weakness of Large Vision-Language Models (LVLMs). Existing supervised fine-tuning (SFT) and recent reinforcement learning with verifiable rewards (RLVR) pipelines depend on costly supervision, specialized tools, or constrained environments that limit scale. We introduce Spatial-SSRL, a self-supervised RL paradigm that derives verifiable signals directly from ordinary RGB or RGB-D images. Spatial-SSRL automatically formulates five pretext tasks that capture 2D and 3D spatial structure: shuffled patch reordering, flipped patch recognition, cropped patch inpainting, regional depth ordering, and relative 3D position prediction. These tasks provide ground-truth answers that are easy to verify and require no human or LVLM annotation. Training on our tasks substantially improves spatial reasoning while preserving general visual capabilities. On seven spatial understanding benchmarks in both image and video settings, Spatial-SSRL delivers average accuracy gains of 4.63%(3B) and 3.89%(7B) over the Qwen2. 5-VL baselines. Our results show that simple, intrinsic supervision enables RLVR at scale and provides a practical route to stronger spatial intelligence in LVLMs.

Paper
35 citations

OPT-BENCH: Evaluating LLM Agent on Large-Scale Search Spaces Optimization Problems

Agentic SystemsScientific ReasoningLLM Evaluation

Xiaozhe Li, Jixuan Chen, Xinyu Fang, Shengyuan Ding, Haodong Duan, Qingwen Liu, Kai Chen

ACL Findings, 2026

Full metadata
Publication date
2025/6/12
Journal
ACL Findings, 2026

Description

Large Language Models (LLMs) have shown remarkable capabilities in solving diverse tasks. However, their proficiency in iteratively optimizing complex solutions through learning from previous feedback remains insufficiently explored. To bridge this gap, we present OPT-BENCH, a comprehensive benchmark designed to evaluate LLM agents on large-scale search space optimization problems. OPT-BENCH includes 20 real-world machine learning tasks sourced from Kaggle and 10 classical NP problems, offering a diverse and challenging environment for assessing LLM agents on iterative reasoning and solution refinement. To enable rigorous evaluation, we introduce OPT-Agent, an end-to-end optimization framework that emulates human reasoning when tackling complex problems by generating, validating, and iteratively improving solutions through leveraging historical feedback. Through extensive experiments on 9 state-of-the-art LLMs from 6 model families, we analyze the effects of optimization iterations, temperature settings, and model architectures on solution quality and convergence. Our results demonstrate that incorporating historical context significantly enhances optimization performance across both ML and NP tasks. All datasets, code, and evaluation tools are open-sourced to promote further research in advancing LLM-driven optimization and iterative reasoning. Project page: \href{https://github.com/OliverLeeXZ/OPT-BENCH}{https://github.com/OliverLeeXZ/OPT-BENCH}.

Paper
15 citations

Np-engine: Empowering optimization reasoning in large language models with verifiable synthetic np problems

Scientific ReasoningFoundation Models

Xiaozhe Li, Xinyu Fang, Shengyuan Ding, Linyang Li, Haodong Duan, Qingwen Liu, Kai Chen

ACL Findings, 2026

Full metadata
Publication date
2025/10/18
Journal
ACL Findings, 2026

Description

Large Language Models (LLMs) have shown strong reasoning capabilities, with models like OpenAI's O-series and DeepSeek R1 excelling at tasks such as mathematics, coding, logic, and puzzles through Reinforcement Learning with Verifiable Rewards (RLVR). However, their ability to solve more complex optimization problems - particularly NP-hard tasks - remains underexplored. To bridge this gap, we propose NP-ENGINE, the first comprehensive framework for training and evaluating LLMs on NP-hard problems. NP-ENGINE covers 10 tasks across five domains, each equipped with (i) a controllable instance generator, (ii) a rule-based verifier, and (iii) a heuristic solver that provides approximate optimal solutions as ground truth. This generator-verifier-heuristic pipeline enables scalable and verifiable RLVR training under hierarchical difficulties. We also introduce NP-BENCH, a benchmark derived from NP-ENGINE-DATA, specifically designed to evaluate LLMs' ability to tackle NP-hard level reasoning problems, focusing not only on feasibility but also on solution quality. Additionally, we present QWEN2.5-7B-NP, a model trained via zero-RLVR with curriculum learning on Qwen2.5-7B-Instruct, which significantly outperforms GPT-4o on NP-BENCH and achieves SOTA performance with the same model size. Beyond in-domain tasks, we demonstrate that RLVR training on NP-ENGINE-DATA enables strong out-of-domain (OOD) generalization to reasoning tasks (logic, puzzles, math, and knowledge), as well as non-reasoning tasks such as instruction following. We also observe a scaling trend: increasing task diversity improves OOD …

Paper
14 citations

ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning

Agentic SystemsReward ModelingMultimodal Generation

Shengyuan Ding, Xinyu Fang, Ziyu Liu, Yuhang Zang, Yuhang Cao, Xiangyu Zhao, Haodong Duan, Xiaoyi Dong, Jianze Liang, Bin Wang, Conghui He, Dahua Lin, Jiaqi Wang

CVPR, 2026

Full metadata
Publication date
2025/12/4
Journal
CVPR, 2026

Description

Reward models are critical for aligning vision-language systems with human preferences, yet current approaches suffer from hallucination, weak visual grounding, and an inability to use tools for verification, limiting their reliability on complex multimodal reasoning tasks. We present** ARM-Thinker**, an** A** gentic multimodal** R** eward** M** odel that autonomously invokes external tools (eg, image cropping, doc page retrieval) to ground judgments in verifiable evidence, replacing static, non-interactive reward scoring. This enables the model to verify fine-grained visual details, cross-reference multi-page evidence, and validate reasoning claims, which are capabilities absent in existing reward models. We train ARM-Thinker with multi-stage reinforcement learning, jointly optimizing tool-calling decisions and judgment accuracy. To evaluate agentic reward modeling, we introduce** ARMBench-VL**, comprising three benchmarks that assess fine-grained visual grounding (image-level tools), multi-page document understanding (retrieval tools), and instruction following (text-level verification). ARM-Thinker achieves+ 16.2% average improvement on reward modeling benchmarks,+ 9.6% on tool-use tasks, and outperforms baselines on multimodal math and logical reasoning benchmarks. Our results demonstrate that agentic capabilities significantly enhance both accuracy and interpretability of reward models.

Paper
12 citations

FARS: A fully automated research system deployed at scale

Agentic SystemsScientific Reasoning

Qiong Tang, Tianxiang Sun, Xiangkun Hu, Xiangyang Liu, Yiran Chen, Yunfan Shao, Bobo Li, Changze Lv, Cheng Xu, Chengsong Huang, Chunyang Li, Dizhan Xue, Hao Bai, Haodong Duan, Hengquan Guo, Hongyang He, Hongyi Chen, Hui Shen, Jiahao Yuan, Jiankai Sun, Jikang Cheng, Jinfeng Xu, Jingqi Tong, Jingye Chen, Jinxiu Liu, Jixuan Leng, Junchi Yu, Kaixun Jiang, Kun Xiang, Kunpeng Yao, Lang Feng, Liangqi Yuan, Longsen Gao, Meng Li, Qi Jia, Qiushi Sun, Shengyuan Ding, Shizhan Gong, Siru Zhong, Terry Jingchen Zhang, Tianle Gu, Tianyi Liang, Weijie Liu, Weikai Yang, Weizhi Fei, Xin Wang, Xinpeng Liu, Xuanwen Ding, Yihong Tang, Yuanli Wang, Yukun Jiang, Yuming Yang, Zhengbao He, Zhikai Chen, Zhikun Xu, Zhuang Li, Zihao Huang

Technical Report, arXiv preprint arXiv:2606.31651, 2026

Full metadata
Publication date
2026/6/30
Journal
Technical Report, arXiv preprint arXiv:2606.31651

Description

Recent automated research systems show that language-model agents can generate hypotheses, run experiments, and write complete manuscripts, but most evidence still comes from selected examples, human-framed topics, or a few pre-defined research tasks. We present FARS (Fully Automated Research System), a fully automated AI-for-AI research system designed to operate across research topics at scale. FARS autonomously generates and advances projects through ideation, planning, experimentation, and writing, using stage-specific agents coordinated through a shared workspace that records proposals, code, logs, results, and manuscripts. In its first public deployment, FARS produced 166 complete research papers spanning 67 fine-grained AI/ML topics while preserving intermediate artifacts as an auditable corpus rather than a curated set of successes. We evaluate this corpus with 282 structured reviews from volunteer reviewers covering 140 papers, including overall ratings, sub-scores, integrity checks, and LLM-use disclosure. The reviews indicate that FARS can produce review-worthy and occasionally strong AI/ML research artifacts in a large-scale public deployment, while also exposing recurring failure modes in narrow experimental scope, methodological limitations, and integrity issues.

Paper
7 citations

RISE-Video: Can Video Generators Decode Implicit World Rules?

Multimodal GenerationVideo UnderstandingVisual Reasoning

Mingxin Liu, Shuran Ma, Shibei Meng, Xiangyu Zhao, Zicheng Zhang, Shaofeng Zhang, Zhihang Zhong, Peixian Chen, Haoyu Cao, Xing Sun, Haodong Duan, Xue Yang

NeurIPS D&B, 2026

Full metadata
Publication date
2026/2/5
Journal
arXiv preprint arXiv:2602.05986

Description

While generative video models have achieved remarkable visual fidelity, their capacity to internalize and reason over implicit world rules remains a critical yet under-explored frontier. To bridge this gap, we present RISE-Video, a pioneering reasoning-oriented benchmark for Text-Image-to-Video (TI2V) synthesis that shifts the evaluative focus from surface-level aesthetics to deep cognitive reasoning. RISE-Video comprises 467 meticulously human-annotated samples spanning eight rigorous categories, providing a structured testbed for probing model intelligence across diverse dimensions, ranging from commonsense and spatial dynamics to specialized subject domains. Our framework introduces a multi-dimensional evaluation protocol consisting of four metrics: \textit{Reasoning Alignment}, \textit{Temporal Consistency}, \textit{Physical Rationality}, and \textit{Visual Quality}. To further support scalable evaluation, we propose an automated pipeline leveraging Large Multimodal Models (LMMs) to emulate human-centric assessment. Extensive experiments on 11 state-of-the-art TI2V models reveal pervasive deficiencies in simulating complex scenarios under implicit constraints, offering critical insights for the advancement of future world-simulating generative models.

Paper
7 citations

Think Visually, Reason Textually: Vision-Language Synergy in ARC

Visual ReasoningFoundation Models

Beichen Zhang, Yuhang Zang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, Jiaqi Wang

CVPR, 2026

Full metadata
Publication date
2025/11/19
Journal
CVPR, 2026

Description

Abstract reasoning from minimal examples remains a core unsolved problem for frontier foundation models such as GPT-5 and Grok 4. These models still fail to infer structured transformation rules from a handful of examples, which is a key hallmark of human intelligence. The Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) provides a rigorous testbed for this capability, demanding conceptual rule induction and transfer to novel tasks. Most existing methods treat ARC-AGI as a purely textual reasoning task, overlooking the fact that humans rely heavily on visual abstraction when solving such puzzles. However, our pilot experiments reveal a paradox: naively rendering ARC-AGI grids as images degrades performance due to imprecise rule execution. This leads to our central hypothesis that vision and language possess complementary strengths across distinct reasoning stages: vision supports global pattern abstraction and verification, whereas language specializes in symbolic rule formulation and precise execution. Building on this insight, we introduce two synergistic strategies: (1) Vision-Language Synergy Reasoning (VLSR), which decomposes ARC-AGI into modality-aligned subtasks; and (2) Modality-Switch Self-Correction (MSSC), which leverages vision to verify text-based reasoning for intrinsic error correction. Extensive experiments demonstrate that our approach yields up to a 4.33\% improvement over text-only baselines across diverse flagship models and multiple ARC-AGI tasks. Our findings suggest that unifying visual abstraction with linguistic reasoning is a crucial step toward achieving generalizable …

Paper
5 citations

Videozerobench: Probing the limits of video mllms with spatio-temporal evidence verification

Video UnderstandingMultimodal EvaluationBenchmark Design

Jiahao Meng, Tan Yue, Qi Xu, Haochen Wang, Zhongwei Ren, Weisong Liu, Yuhao Wang, Renrui Zhang, Yunhai Tong, Haodong Duan

arXiv preprint arXiv:2604.01569, 2026

Full metadata
Publication date
2026/4/2
Journal
arXiv preprint arXiv:2604.01569

Description

Recent video multimodal large language models achieve impressive results across various benchmarks. However, current evaluations suffer from two critical limitations: (1) inflated scores can mask deficiencies in fine-grained visual understanding and reasoning, and (2) answer correctness is often measured without verifying whether models identify the precise spatio-temporal evidence supporting their predictions. To address this, we present VideoZeroBench, a hierarchical benchmark designed for challenging long-video question answering that rigorously verifies spatio-temporal evidence. It comprises 500 manually annotated questions across 13 domains, paired with temporal intervals and spatial bounding boxes as evidence. To disentangle answering generation, temporal grounding, and spatial grounding, we introduce a five-level evaluation protocol that progressively tightens evidence requirements. Experiments show that even Gemini-3-Pro correctly answers fewer than 17% of questions under the standard end-to-end QA setting (Level-3). When grounding constraints are imposed, performance drops sharply: No model exceeds 1% accuracy when both correct answering and accurate spatio-temporal localization are required (Level-5), with most failing to achieve any correct grounded predictions. These results expose a significant gap between surface-level answer correctness and genuine evidence-based reasoning, revealing that grounded video understanding remains a bottleneck for long-video QA. We further analyze performance across minimal evidence spans, atomic abilities, and inference paradigms, providing insights for future …

Paper
5 citations

MM-HELIX: Boosting Multimodal Long-Chain Reflective Reasoning with Holistic Platform and Adaptive Hybrid Policy Optimization

Reinforcement LearningModel Adaptation

Xiangyu Zhao, Junming Lin, Tianhao Liang, Yifan Zhou, Wenhao Chai, Yuzhe Gu, Weiyun Wang, Kai Chen, Gen Luo, Wenwei Zhang, Junchi Yan, Hua Yang, Haodong Duan, Xue Yang

ICLR, 2026

Full metadata
Publication date
2025/10/9
Journal
ICLR, 2026

Description

While current Multimodal Large Language Models (MLLMs) have demonstrated proficiency in reasoning tasks such as mathematics and logic, their capacity for long-chain reflective reasoning, a prerequisite for solving complex real-world problems, remains largely underexplored. In this work, we first conduct an extensive empirical investigation to evaluate this capability. Leveraging a carefully designed data synthesis engine, we construct MM-HELIX, a multimodal benchmark consisting 1,260 samples of 42 challenging synthetic tasks that require iterative thinking and backtracking. Empirical results on this benchmark reveal that existing MLLMs exhibit significant performance deficits in long-chain reflective reasoning. To address this limitation, we generate post-training data and further explore learning paradigms for exploiting such data. We first develop the Step-Elicited Response Generation pipeline to create MM-HELIX-100K, a large-scale dataset of 100k high-quality, reflective reasoning traces for instruction-tuning stage. Given that standard Reinforcement Learning fails on complex tasks due to sparse reward signals and catastrophic forgetting after Supervised Fine-Tuning, we propose Adaptive Hybrid Policy Optimization (AHPO), a novel training strategy that dynamically unifies offline supervision and online optimization into a single stage. This strategy enables the model to learn from expert data when rewards are sparse and conduct independent exploration once proficient. When applied to the Qwen2. 5-VL-7B baseline, our method achieves a+ 18.6\% accuracy improvement on MM-HELIX benchmark and demonstrates strong …

Paper
4 citations

Project imaging-x: A survey of 1000+ open-access medical imaging datasets for foundation model development

Medical AIBenchmark DesignData-Centric Learning

Zhongying Deng, Cheng Tang, Ziyan Huang, Jiashi Lin, Ying Chen, Junzhi Ning, Chenglong Ma, Jiyao Liu, Wei Li, Yinghao Zhu, Shujian Gao, Yanyan Huang, Sibo Ju, Yanzhou Su, Pengcheng Chen, Wenhao Tang, Tianbin Li, Haoyu Wang, Yuanfeng Ji, Hui Sun, Shaobo Min, Liang Peng, Feilong Tang, Haochen Xue, Rulin Zhou, Chaoyang Zhang, Wenjie Li, Shaohao Rui, Weijie Ma, Xingyue Zhao, Yibin Wang, Kun Yuan, Zhaohui Lu, Shujun Wang, Jinjie Wei, Lihao Liu, Dingkang Yang, Lin Wang, Yulong Li, Haolin Yang, Yiqing Shen, Lequan Yu, Xiaowei Hu, Yun Gu, Yicheng Wu, Benyou Wang, Minghui Zhang, Angelica I Aviles-Rivero, Qi Gao, Hongming Shan, Xiaoyu Ren, Fang Yan, Hongyu Zhou, Haodong Duan, Maosong Cao, Shanshan Wang, Bin Fu, Xiaomeng Li, Zhi Hou, Chunfeng Song, Lei Bai, Yuan Cheng, Yuandong Pu, Xiang Li, Wenhai Wang, Hao Chen, Jiaxin Zhuang, Songyang Zhang, Huiguang He, Mengzhang Li, Bohan Zhuang, Zhian Bai, Rongshan Yu, Liansheng Wang, Yukun Zhou, Xiaosong Wang, Xin Guo, Guanbin Li, Xiangru Lin, Dakai Jin, Mianxin Liu, Wenlong Zhang, Qi Qin, Conghui He, Yuqiang Li, Ye Luo, Nanqing Dong, Jie Xu, Wenqi Shao, Bo Zhang, Qiujuan Yan, Yihao Liu, Jun Ma, Zhi Lu, Yuewen Cao, Zongwei Zhou, Jianming Liang, Shixiang Tang, Qi Duan, Dongzhan Zhou, Chen Jiang, Yuyin Zhou, Yanwu Xu, Jiancheng Yang, Shaoting Zhang, Xiaohong Liu, Siqi Luo, Yi Xin, Chaoyu Liu, Haochen Wen, Xin Chen, Alejandro Lozano, Min Woo Sun, Yuhui Zhang, Yue Yao, Xiaoxiao Sun, Serena Yeung-Levy, Xia Li, Jing Ke, Chunhui Zhang, Zongyuan Ge, Ming Hu, Jin Ye, Zhifeng Li, Yirong Chen, Yu Qiao, Junjun He

arXiv preprint arXiv:2603.27460, 2026

Full metadata
Publication date
2026/3/29
Journal
arXiv preprint arXiv:2603.27460

Description

Foundation models have demonstrated remarkable success across diverse domains and tasks, primarily due to the thrive of large-scale, diverse, and high-quality datasets. However, in the field of medical imaging, the curation and assembling of such medical datasets are highly challenging due to the reliance on clinical expertise and strict ethical and privacy constraints, resulting in a scarcity of large-scale unified medical datasets and hindering the development of powerful medical foundation models. In this work, we present the largest survey to date of medical image datasets, covering over 1,000 open-access datasets with a systematic catalog of their modalities, tasks, anatomies, annotations, limitations, and potential for integration. Our analysis exposes a landscape that is modest in scale, fragmented across narrowly scoped tasks, and unevenly distributed across organs and modalities, which in turn limits the utility of existing medical image datasets for developing versatile and robust medical foundation models. To turn fragmentation into scale, we propose a metadata-driven fusion paradigm (MDFP) that integrates public datasets with shared modalities or tasks, thereby transforming multiple small data silos into larger, more coherent resources. Building on MDFP, we release an interactive discovery portal that enables end-to-end, automated medical image dataset integration, and compile all surveyed datasets into a unified, structured table that clearly summarizes their key characteristics and provides reference links, offering the community an accessible and comprehensive repository. By charting the current terrain and offering a principled …

Paper
3 citations

Workflow-gym: Towards long-horizon evaluation of computer-use agentic tasks in real-world professional fields

Agentic SystemsBenchmark Design

Liya Zhu, Jingzhe Ding, Jian Zhang, Jianbo Xue, Shihao Liang, Ge Zhang, Yi Zhu, Duju Zeng, Xiang Gao, Qingshui Gu, Mailun Gao, Huimin Che, Yan Zhao, Peiheng Zhou, Haojun Wang, Chaobo Xian, Lili Le, Chi Wu, Yiwei Liu, Shengda Long, Jiale Yang, Fangzhi Xu, Sijin Wu, Haodong Duan, Chao He, Zhaojian Li, Minchao Wang, Huan Zhou, Jiani Hou, Chuqian Yu, Weiran Shi, Hongwan Gao, Jiamin Chen, Guanhong Chen, Tingqin Luo, Kaiyuan Zhang, Zhixin Yao, Qing Hua, Yuhao Jiang, Jin Chen, Pu Chen, Zhenyu Hu, Xingyu Li, Zhengxuan Jiang, Meng Cao, Tianfeng Long, Haozhe Wang, Mingzhang Wang, Yichen Zhang, Yiming Dai, Chenchen Zhang, Jiaying Wang, Xinying Liu, Xingzu Liu, Lingling Zhang, Xinjie Chen, Yujia Qin, Wangchunshu Zhou, Zhiyong Wu, Yang Liu, Jiaheng Liu, Lei Zhang, Shen Yan, Wenhao Huang, Zaiyuan Wang, Xiaolong Chang

arXiv preprint arXiv:2606.11042, 2026

Full metadata
Publication date
2026/6/9
Journal
arXiv preprint arXiv:2606.11042

Description

Recent years have witnessed the rapid evolution of AI agents toward handling increasingly complex, real-world tasks. However, existing benchmarks rarely evaluate whether agents can operate graphical user interfaces to complete long-horizon, high-value professional workflows across diverse domains. Current GUI benchmarks still predominantly focus on general-purpose software, relatively simple applications, and short-horizon tasks, leaving it largely unknown whether modern agents can follow user instructions to autonomously operate domain-specific professional software and accomplish economically valuable work in an end-to-end manner. To bridge this gap, we introduce Workflow-GYM, a benchmark for long-horizon GUI tasks centered on professional domains and specialized software environments. Through extensive experiments on state-of-the-art models, we find that even the strongest models achieve only slightly above 30% success rates, highlighting that professional long-horizon GUI workflows remain highly challenging for current GUI agents. Further analysis reveals that current agents struggle to maintain long-horizon workflow consistency, frequently exhibiting workflow stage omission, error propagation, objective drift, and insufficient understanding of professional software environments. Our findings provide important insights into the limitations of current agent systems and suggest key directions for the next generation of GUI-agent research.

Paper
3 citations

Can Retrieval Heads See Images? Multimodal Retrieval Heads in Long-Context Vision-Language Models

Long-Context ModelingLLM EvaluationFoundation Models

Aaron Branson Cigres Li, Zhaowei Wang, Yu Zhao, Yiming Du, Haobo Li, Xiyu Ren, Ginny Wong, Simon See, Lishu Luo, Haodong Duan, Pasquale Minervini, Yangqiu Song

EMNLP, 2026

Full metadata
Publication date
2026/5/26
Journal
EMNLP, 2026

Description

Large vision-language models increasingly rely on long-context modeling to reason over documents, hour-level videos, and long-horizon agent trajectories, requiring them to locate relevant evidence across interleaved text and images. Prior work has studied this behavior using retrieval heads in large language models, but its copy-based criterion does not directly apply when evidence appears in images. We introduce a multimodal retrieval head detection method that scores attention from question tokens to textual or visual evidence. With this method, we show that multimodal retrieval heads are sparse, intrinsic, and causally important: only 4.4-10.2% of attention heads account for 50% of the positive retrieval-score mass, and masking the top-5% selected heads drops MMLongBench-Doc from 48.2% to 5.7% and SlideVQA from 71.2% to 8.9%, while random-head masking is far less damaging. Further analysis shows that these heads are partly shared across modalities yet remain dynamic within each modality, with image retrieval heads changing more than text retrieval heads as context length and haystack modality change. Without further training, we find that these heads can also be used directly to rank visually rich documents: on MMDocIR, Qwen3-VL-8B selected-head scoring improves Recall@1 by 7.7/7.4 macro/micro points for page retrieval and 6.3/6.8 points for layout retrieval over the strongest reported baseline.

Paper
2 citations

PosterIQ: A Design Perspective Benchmark for Poster Understanding and Generation

Document IntelligenceMultimodal GenerationBenchmark Design

Yuheng Feng, Wen Zhang, Haodong Duan, Xingxing Zou

CVPR, 2026

Full metadata
Publication date
2026/3/25
Journal
CVPR, 2026

Description

We present PosterIQ, a design-driven benchmark for poster understanding and generation, annotated across composition structure, typographic hierarchy, and semantic intent. It includes 7,765 image-annotation instances and 822 generation prompts spanning real, professional, and synthetic cases. To bridge visual design cognition and generative modeling, we define tasks for layout parsing, text-image correspondence, typography/readability and font perception, design quality assessment, and controllable, composition-aware generation with metaphor. We evaluate state-of-the-art MLLMs and diffusion-based generators, finding persistent gaps in visual hierarchy, typographic semantics, saliency control, and intention communication; commercial models lead on high-level reasoning but act as insensitive automatic raters, while generators render text well yet struggle with composition-aware synthesis. Extensive analyses show PosterIQ is both a quantitative benchmark and a diagnostic tool for design reasoning, offering reproducible, task-specific metrics. We aim to catalyze models' creativity and integrate human-centred design principles into generative vision-language systems.

Paper
2 citations

Mibench: Evaluating lmms on multimodal interaction

Multimodal EvaluationBenchmark DesignFoundation Models

Yu Miao, Zequn Yang, Yake Wei, Ziheng Chen, Haotian Ni, Haodong Duan, Kai Chen, Di Hu

arXiv preprint arXiv:2603.13427, 2026

Full metadata
Publication date
2026/3/13
Journal
arXiv preprint arXiv:2603.13427

Description

In different multimodal scenarios, it needs to integrate and utilize information across modalities in a specific way based on the demands of the task. Different integration ways between modalities are referred to as "multimodal interaction". How well a model handles various multimodal interactions largely characterizes its multimodal ability. In this paper, we introduce MIBench, a comprehensive benchmark designed to evaluate the multimodal interaction capabilities of Large Multimodal Models (LMMs), which formulates each instance as a (con_v , con_t, task) triplet with contexts from vision and text, necessitating that LMMs employ correct forms of multimodal interaction to effectively complete the task. MIBench assesses models from three key aspects: the ability to source information from vision-centric or text-centric cues, and the ability to generate new information from their joint synergy. Each interaction capability is evaluated hierarchically across three cognitive levels: Recognition, Understanding, and Reasoning. MIBench comprises over 10,000 vision-text context pairs spanning 32 distinct tasks. Evaluation of state-of-the-art LMMs show that: (1) LMMs' ability on multimodal interaction remains constrained, despite the scaling of model parameters and training data; (2) they are easily distracted by textual modalities when processing vision information; (3) they mostly possess a basic capacity for multimodal synergy; and (4) natively trained multimodal models show noticeable deficits in fundamental interaction ability. We expect that these observations can serve as a reference for developing LMMs with more enhanced multimodal ability in the future.

Paper
1 citations

Training Long-Context Vision-Language Models Effectively with Generalization Beyond 128K Context

Long-Context ModelingModel AdaptationFoundation Models

Zhaowei Wang, Lishu Luo, Haodong Duan, Weiwei Liu, Sijin Wu, Ji Luo, Shen Yan, Shuai Peng, Sihang Yuan, Chaoyi Huang, Yi Lin, Yangqiu Song

NeurIPS, 2026

Full metadata
Publication date
2026/5/13
Journal
arXiv preprint arXiv:2605.13831

Description

Long-context modeling is becoming a core capability of modern large vision-language models (LVLMs), enabling sustained context management across long-document understanding, video analysis, and multi-turn tool use in agentic workflows. Yet practical training recipes remain insufficiently explored, particularly for designing and balancing long-context data mixtures. In this work, we present a systematic study of long-context continued pre-training for LVLMs, extending a 7B model from 32K to 128K context with extensive ablations on long-document data. We first show that long-document VQA is substantially more effective than OCR transcription. Building on this observation, our ablations further yield three key findings: i) for sequence-length distribution, balanced data outperforms target-length-focused data (e.g., 128K), suggesting that long-context ability requires generalizable key-information retrieval across various lengths and positions; ii) retrieval remains the primary bottleneck, favoring retrieval-heavy mixtures with modest reasoning data for task diversity; and iii) pure long-document VQA largely preserves short-context capabilities, suggesting that instruction-formatted long data reduces the need for short-data mixing. Based on these findings, we introduce MMProLong, obtained by long-context continued pre-training from Qwen2.5-VL-7B with only a 5B-token budget. MMProLong improves long-document VQA scores by 7.1% and maintains strong performance at 256K and 512K contexts beyond its 128K training window, without additional training. It further generalizes to webpage-based multimodal needle retrieval, long-context vision …

Paper
1 citations

Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games

Agentic SystemsMultimodal EvaluationLLM Evaluation

Shengyuan Ding, Xilin Wei, Xinyu Fang, Haodong Duan, Dahua Lin, Jiaqi Wang, Yuhang Zang

EMNLP, 2026

Full metadata
Publication date
2026/6/17
Journal
EMNLP, 2026

Description

Deploying multimodal foundation models as closed-loop policies increasingly requires conditioning actions on observations that are no longer visible. However, existing benchmarks either expose the full state, conflate hidden-state reconstruction with other agent skills, or test recall only after an episode has ended. We introduce RNG-Bench (Reconstructive Non-Markov Games), a benchmark suite designed to isolate a base model's ability to reconstruct past observations and act on them during multi-step interaction. RNG-Bench includes two complementary games: Matching Pairs, where card identities briefly revealed at specific locations must later be recalled, and 3D Maze, where egocentric views must be integrated into a spatial map. Both games are evaluated under a unified harness with three controlled difficulty axes: grid size, visual pattern, and observation modality. The benchmark further introduces a head-to-head duel protocol to control for instance-level variance and a Memory Gap metric that disentangles forgetting from poor action selection. The hardest configurations require contexts of roughly 128K tokens and 350 image inputs per episode, and remain far from saturated by frontier MLLMs. Memory Gap analysis shows that most residual errors stem from forgetting earlier observations rather than from suboptimal decision making. Finally, fine-tuning Qwen3.5-9B on optimal-policy rollouts and filtered model demonstrations improves performance on RNG-Bench and transfers to existing benchmarks without degrading general multimodal capability.

Paper
0 citations

V-ICAL Bench: Evaluating Video In-Context Learning for Multimodal Agents in Interactive Environments

Agentic SystemsVideo UnderstandingMultimodal Evaluation

Ziqian Fan, Shibo Xu, Junjie Li, Xiangyu Zhao, Shengyuan Ding, Yifan Yang, Zhenjie Yang, Haodong Duan, Yue Zhou, Zhihang Zhong, Xue Yang

arXiv preprint arXiv:2609.15683, 2026

Full metadata
Publication date
2026/9/14
Journal
arXiv preprint arXiv:2609.15683

Description

While In-Context Learning (ICL) enables models to adapt from exemplars without parameter updates, multimodal ICL remains largely underexplored, particularly regarding video demonstrations in interactive environments. For multimodal agents, learning from videos presents unique challenges: they must translate in-context demonstrations into executable policies, ground these policies in novel visual states, and iteratively refine actions based on environmental feedback. We introduce V-ICAL, a novel benchmark designed to evaluate video-based ICL for multimodal agents. Comprising 342 interactive tasks across 37 environments, V-ICAL utilizes human-curated demonstration videos as task-specific behavioral exemplars, evaluating agents through sustained interaction from a target initialization. The benchmark seamlessly connects in-context knowledge induction with core agentic capabilities, including state grounding, temporal memory, planning, and adaptation in dynamic environments. Extensive evaluations across 19 state-of-the-art multimodal agents reveal significant limitations: the best-performing model, Seed-2.1-Pro, achieves a score of only 54.4/100, while other leading models (e.g., Gemini-3.1-Pro, GPT-5.6) fail to surpass 50, far below the human baseline of 83.6. Controlled comparisons further demonstrate that current agents struggle to reliably translate video exemplars into effective policies, failing to yield consistent performance gains. Ultimately, V-ICAL exposes a critical gap in the ICL capabilities of multimodal agents, underscoring an urgent need for future research.

Paper
0 citations

2025

InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency

Model EfficiencyFoundation Models

Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, Zhaokai Wang, Zhe Chen, Hongjie Zhang, Ganlin Yang, Haomin Wang, Qi Wei, Jinhui Yin, Wenhao Li, Erfei Cui, Guanzhou Chen, Zichen Ding, Changyao Tian, Zhenyu Wu, Jingjing Xie, Zehao Li, Bowen Yang, Yuchen Duan, Xuehui Wang, Songze Li, Xiangyu Zhao, Haodong Duan, Nianchen Deng, Bin Fu, Yinan He, Yi Wang, Conghui He, Botian Shi, Junjun He, Yingtong Xiong, Han Lv, Lijun Wu, Wenqi Shao, Kaipeng Zhang, Huipeng Deng, Biqing Qi, Jiaye Ge, Qipeng Guo, Wenwei Zhang, Wanli Ouyang, Limin Wang, Min Dou, Xizhou Zhu, Tong Lu, Dahua Lin, Jifeng Dai, Bowen Zhou, Weijie Su, Kai Chen, Yu Qiao, Wenhai Wang, Gen Luo

arXiv preprint arXiv:2508.18265, 2025

Full metadata
Publication date
2025/8/25
Journal
arXiv preprint arXiv:2508.18265

Description

We introduce InternVL 3.5, a new family of open-source multimodal models that significantly advances versatility, reasoning capability, and inference efficiency along the InternVL series. A key innovation is the Cascade Reinforcement Learning (Cascade RL) framework, which enhances reasoning through a two-stage process: offline RL for stable convergence and online RL for refined alignment. This coarse-to-fine training strategy leads to substantial improvements on downstream reasoning tasks, e.g., MMMU and MathVista. To optimize efficiency, we propose a Visual Resolution Router (ViR) that dynamically adjusts the resolution of visual tokens without compromising performance. Coupled with ViR, our Decoupled Vision-Language Deployment (DvD) strategy separates the vision encoder and language model across different GPUs, effectively balancing computational load. These contributions collectively enable InternVL3.5 to achieve up to a +16.0\% gain in overall reasoning performance and a 4.05 inference speedup compared to its predecessor, i.e., InternVL3. In addition, InternVL3.5 supports novel capabilities such as GUI interaction and embodied agency. Notably, our largest model, i.e., InternVL3.5-241B-A28B, attains state-of-the-art results among open-source MLLMs across general multimodal, reasoning, text, and agentic tasks -- narrowing the performance gap with leading commercial models like GPT-5. All models and code are publicly released.

Paper
1647 citations

Visual-RFT: Visual Reinforcement Fine-Tuning

Reinforcement LearningReward ModelingModel Adaptation

Ziyu Liu, Zeyi Sun, Yuhang Zang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, Jiaqi Wang

IEEE/CVF International Conference on Computer Vision (ICCV), 2025

Full metadata
Publication date
2025/3/3
Journal
ICCV, 2025

Description

Reinforcement Fine-Tuning (RFT) in Large Reasoning Models like OpenAI ol learns from feedback on its answers, which is especially useful in applications when fine-tuning data is scarce. Recent open-source work like DeepSeekR1 demonstrates that reinforcement learning with verifiable reward is possibly one key direction in reproducing o1. While the R1-style model has demonstrated success in language models, its application in multi-modal domains remains under-explored. This work introduces Visual Reinforcement Fine-Tuning (Visual-RFT), which further extends the application areas of RFT on visual tasks. Specifically, Visual-RFT first uses Large Vision-Language Models (LVLMs) to generate multiple responses containing reasoning tokens and final answers for each input, and then uses our proposed visual perception verifiable reward functions to update the model via the policy optimization algorithm …

Paper Code
623 citations

OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding?

Video UnderstandingMultimodal EvaluationLLM Evaluation

Yifei Li, Junbo Niu, Ziyang Miao, Chunjiang Ge, Yuanhang Zhou, Qihao He, Xiaoyi Dong, Haodong Duan, Shuangrui Ding, Rui Qian, Pan Zhang, Yuhang Zang, Yuhang Cao, Conghui He, Jiaqi Wang

CVPR, 2025

Full metadata
Publication date
2025/1/9
Journal
CVPR, 2025

Description

Temporal Awareness—the ability to reason dynamically based on the timestamp when a question is raised—is the key distinction between offline and online video LLMs. Unlike offline models, which rely on complete videos for static, post hoc analysis, online models process video streams incrementally and dynamically adapt their responses based on the timestamp at which the question is posed. Despite its significance, temporal awareness has not been adequately evaluated in existing benchmarks. To fill this gap, we present OVO-Bench (Online-VideO-Benchmark), a novel video benchmark that emphasizes the importance of timestamps for advanced online video understanding capability benchmarking. OVO-Bench evaluates the ability of video LLMs to reason and respond to events occurring at specific timestamps under three distinct scenarios: (1) Backward tracing: trace back to past events to answer the …

Paper
164 citations

Seed2.0 model card: Towards intelligence frontier for real-world complexity

Agentic SystemsFoundation Models

Bytedance Seed

Technical report, Technical report, Bytedance, 2025. URL https://lf3-static …

Full metadata
Publication date
2026/2
Publisher
Technical report, Technical report, Bytedance, 2025. URL https://lf3-static. bytednsdoc. com

Description

We present Seed2.0, a model series that takes a meaningful step toward solving complex, real-world tasks. Our approach begins with identifying users' genuine needs and constructing a reliable, forward-looking evaluation system by selecting and abstracting benchmarks grounded in these needs and in realistic, complex scenarios. Guided by this evaluation system, Seed2.0 targets two persistent challenges, long-tail knowledge and complex instruction following, substantially improving the model's reliability on intricate, long-horizon tasks. Beyond these, Seed2.0 delivers world-leading reasoning intelligence, visual understanding, and search capabilities that address the most common needs of a broad user base. Through extensive real-world use cases documented in this model card, we demonstrate that Seed2.0 begins to exhibit the ability to handle initial complex real-world tasks, delivering greater value to hundreds of millions of users.

Paper
138 citations

Intern-s1: A scientific multimodal foundation model

Scientific ReasoningFoundation Models

Lei Bai, Zhongrui Cai, Yuhang Cao, Maosong Cao, Weihan Cao, Chiyu Chen, Haojiong Chen, Kai Chen, Pengcheng Chen, Ying Chen, Yongkang Chen, Yu Cheng, Pei Chu, Tao Chu, Erfei Cui, Ganqu Cui, Long Cui, Ziyun Cui, Nianchen Deng, Ning Ding, Nanqing Dong, Peijie Dong, Shihan Dou, Sinan Du, Haodong Duan, Caihua Fan, Ben Gao, Changjiang Gao, Jianfei Gao, Songyang Gao, Yang Gao, Zhangwei Gao, Jiaye Ge, Qiming Ge, Lixin Gu, Yuzhe Gu, Aijia Guo, Qipeng Guo, Xu Guo, Conghui He, Junjun He, Yili Hong, Siyuan Hou, Caiyu Hu, Hanglei Hu, Jucheng Hu, Ming Hu, Zhouqi Hua, Haian Huang, Junhao Huang, Xu Huang, Zixian Huang, Zhe Jiang, Lingkai Kong, Linyang Li, Peiji Li, Pengze Li, Shuaibin Li, Tianbin Li, Wei Li, Yuqiang Li, Dahua Lin, Junyao Lin, Tianyi Lin, Zhishan Lin, Hongwei Liu, Jiangning Liu, Jiyao Liu, Junnan Liu, Kai Liu, Kaiwen Liu, Kuikun Liu, Shichun Liu, Shudong Liu, Wei Liu, Xinyao Liu, Yuhong Liu, Zhan Liu, Lu

arXiv preprint arXiv:2508.15763, 2025

Full metadata
Publication date
2025/8/21
Journal
arXiv preprint arXiv:2508.15763

Description

In recent years, a plethora of open-source foundation models have emerged, achieving remarkable progress in some widely attended fields, with performance being quite close to that of closed-source models. However, in high-value but more challenging scientific professional fields, either the fields still rely on expert models, or the progress of general foundation models lags significantly compared to those in popular areas, far from sufficient for transforming scientific research and leaving substantial gap between open-source models and closed-source models in these scientific domains. To mitigate this gap and explore a step further toward Artificial General Intelligence (AGI), we introduce Intern-S1, a specialized generalist equipped with general understanding and reasoning capabilities with expertise to analyze multiple science modal data. Intern-S1 is a multimodal Mixture-of-Experts (MoE) model with 28 billion activated parameters and 241 billion total parameters, continually pre-trained on 5T tokens, including over 2.5T tokens from scientific domains. In the post-training stage, Intern-S1 undergoes offline and then online reinforcement learning (RL) in InternBootCamp, where we propose Mixture-of-Rewards (MoR) to synergize the RL training on more than 1000 tasks simultaneously. Through integrated innovations in algorithms, data, and training systems, Intern-S1 achieved top-tier performance in online RL training. On comprehensive evaluation benchmarks, Intern-S1 demonstrates competitive performance on general reasoning tasks among open-source models and significantly outperforms open-source models in scientific domains …

Paper
117 citations

NeedleBench: Evaluating LLM Retrieval and Reasoning Across Varying Information Densities

Long-Context ModelingLLM EvaluationBenchmark Design

Mo Li, Songyang Zhang, Taolin Zhang, Haodong Duan, Yunxin Liu, Kai Chen

TMLR, 2025

Full metadata
Publication date
2025
Journal
TMLR, 2025

Description

The capability of large language models to handle long-context information is crucial across various real-world applications. Existing evaluation methods often rely either on real-world long texts, making it difficult to exclude the influence of models' inherent knowledge, or introduce irrelevant filler content to artificially achieve target lengths, reducing assessment effectiveness. To address these limitations, we introduce NeedleBench, a synthetic framework for assessing retrieval and reasoning performance in bilingual long-context tasks with adaptive context lengths. NeedleBench systematically embeds key data points at varying depths to rigorously test model capabilities. Tasks are categorized into two scenarios: information-sparse, featuring minimal relevant details within extensive irrelevant text to simulate simple retrieval tasks; and information-dense (the Ancestral Trace Challenge), where relevant information is continuously distributed throughout the context to simulate complex reasoning tasks. Our experiments reveal that although recent reasoning models like Deepseek-R1 and OpenAI's o3 excel in mathematical reasoning, they struggle with continuous retrieval and reasoning in information-dense scenarios, even at shorter context lengths. We also characterize a phenomenon termed 'under-thinking', where models prematurely conclude reasoning despite available information. NeedleBench thus provides critical insights and targeted tools essential for evaluating and improving LLMs' long-context capabilities. All resources are available at OpenCompass: https://github.com/open-compass/opencompass.

Paper
101 citations

Envisioning beyond the pixels: Benchmarking reasoning-informed visual editing

Multimodal GenerationVisual ReasoningMultimodal Evaluation

Xiangyu Zhao, Peiyuan Zhang, Kexian Tang, Xiaorong Zhu, Hao Li, Wenhao Chai, Zicheng Zhang, Renqiu Xia, Guangtao Zhai, Junchi Yan, Hua Yang, Xue Yang, Haodong Duan

NeurIPS D&B Oral, 2025

Full metadata
Publication date
2025/4/3
Journal
NeurIPS D&B Oral, 2025

Description

Large Multi-modality Models (LMMs) have made significant progress in visual understanding and generation, but they still face challenges in General Visual Editing, particularly in following complex instructions, preserving appearance consistency, and supporting flexible input formats. To study this gap, we introduce RISEBench, the first benchmark for evaluating Reasoning-Informed viSual Editing (RISE). RISEBench focuses on four key reasoning categories: Temporal, Causal, Spatial, and Logical Reasoning. We curate high-quality test cases for each category and propose an robust evaluation framework that assesses Instruction Reasoning, Appearance Consistency, and Visual Plausibility with both human judges and the LMM-as-a-judge approach. We conducted experiments evaluating nine prominent visual editing models, comprising both open-source and proprietary models. The evaluation results demonstrate that current models face significant challenges in reasoning-based editing tasks. Even the most powerful model evaluated, GPT-image-1, achieves an accuracy of merely 28.8%. RISEBench effectively highlights the limitations of contemporary editing models, provides valuable insights, and indicates potential future directions for the field of reasoning-aware visual editing. Our code and data have been released at https://github. com/PhoenixZ810/RISEBench.

Paper Code
94 citations

InternLM-XComposer2.5-Reward: A Simple Yet Effective Multi-Modal Reward Model

Reward ModelingModel AdaptationFoundation Models

Yuhang Zang, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Ziyu Liu, Shengyuan Ding, Shenxi Wu, Yubo Ma, Haodong Duan, Wenwei Zhang, Kai Chen, Dahua Lin, Jiaqi Wang

ACL Findings, 2025

Full metadata
Publication date
2025/1/21
Journal
ACL Findings, 2025

Description

Despite the promising performance of Large Vision Language Models (LVLMs) in visual understanding, they occasionally generate incorrect outputs. While reward models (RMs) with reinforcement learning or test-time scaling offer the potential for improving generation quality, a critical gap remains: publicly available multi-modal RMs for LVLMs are scarce, and the implementation details of proprietary models are often unclear. We bridge this gap with InternLM-XComposer2. 5-Reward (IXC-2.5-Reward), a simple yet effective multi-modal reward model that aligns LVLMs with human preferences. To ensure the robustness and versatility of IXC-2.5-Reward, we set up a high-quality multi-modal preference corpus spanning text, image, and video inputs across diverse domains, such as instruction following, general understanding, text-rich documents, mathematical reasoning, and video understanding. IXC-2.5-Reward achieves excellent results on the latest multi-modal reward model benchmark and shows competitive performance on text-only reward model benchmarks. We further demonstrate three key applications of IXC-2.5-Reward:(1) Providing a supervisory signal for RL training. We integrate IXC-2.5-Reward with Proximal Policy Optimization (PPO) yields IXC-2.5-Chat, which shows consistent improvements in instruction following and multi-modal open-ended dialogue;(2) Selecting the best response from candidate responses for test-time scaling; and (3) Filtering outlier or noisy samples from existing image and video instruction tuning training data.

Paper
82 citations

VideoRoPE: What Makes for Good Video Rotary Position Embedding?

Long-Context ModelingVideo Understanding

Xilin Wei, Xiaoran Liu, Yuhang Zang, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Jian Tong, Haodong Duan, Qipeng Guo, Jiaqi Wang, Xipeng Qiu, Dahua Lin

ICML Oral, 2025

Full metadata
Publication date
2025/2/7
Journal
ICML Oral, 2025

Description

While Rotary Position Embedding (RoPE) and its variants are widely adopted for their long-context capabilities, the extension of the 1D RoPE to video, with its complex spatio-temporal structure, remains an open challenge. This work first introduces a comprehensive analysis that identifies four key characteristics essential for the effective adaptation of RoPE to video, which have not been fully considered in prior work. As part of our analysis, we introduce a challenging V-NIAH-D (Visual Needle-In-A-Haystack with Distractors) task, which adds periodic distractors into V-NIAH. The V-NIAH-D task demonstrates that previous RoPE variants, lacking appropriate temporal dimension allocation, are easily misled by distractors. Based on our analysis, we introduce \textbf{VideoRoPE}, with a \textit{3D structure} designed to preserve spatio-temporal relationships. VideoRoPE features \textit{low-frequency temporal allocation} to mitigate periodic oscillations, a \textit{diagonal layout} to maintain spatial symmetry, and \textit{adjustable temporal spacing} to decouple temporal and spatial indexing. VideoRoPE consistently surpasses previous RoPE variants, across diverse downstream tasks such as long video retrieval, video understanding, and video hallucination. Our code will be available at \href{https://github.com/Wiselnn570/VideoRoPE}{https://github.com/Wiselnn570/VideoRoPE}.

Paper
75 citations

Large multimodal models evaluation: a survey

Multimodal EvaluationBenchmark DesignFoundation Models

Zicheng Zhang, Junying Wang, Farong Wen, Yijin Guo, Xiangyu Zhao, Xinyu Fang, Shengyuan Ding, Ziheng Jia, Jiahao Xiao, Ye Shen, Yushuo Zheng, Xiaorong Zhu, Yalun Wu, Ziheng Jiao, Wei Sun, Zijian Chen, Kaiwei Zhang, Kang Fu, Yuqin Cao, Ming Hu, Yue Zhou, Xuemei Zhou, Juntai Cao, Wei Zhou, Jinyu Cao, Ronghui Li, Donghao Zhou, Yuan Tian, Xiangyang Zhu, Chunyi Li, Haoning Wu, Xiaohong Liu, Junjun He, Yu Zhou, Hui Liu, Lin Zhang, Zesheng Wang, Huiyu Duan, Yingjie Zhou, Xiongkuo Min, Qi Jia, Dongzhan Zhou, Wenlong Zhang, Jiezhang Cao, Xue Yang, Junzhi Yu, Songyang Zhang, Haodong Duan, Guangtao Zhai

Science China Information Sciences, 2025 68 (12), 221301

Full metadata
Publication date
2025/12
Source
Science China Information Sciences, 2025
Volume
68
Issue
12
Pages
221301
Publisher
Science China Press

Description

As large multimodal models (LMMs) advance rapidly across diverse multimodal understanding and generation tasks, the need for systematic and reliable evaluation frameworks becomes increasingly critical. To address this need, this survey provides a structured overview of LMM evaluation, centered around two main axes: multimodal evaluation for understanding and generation. (1) For understanding, a dual-perspective framework is introduced to distinguish benchmarks between general capabilities, which emphasize common tasks, and specialized capabilities, which reflect expert-level competence in domain-specific fields. (2) For generation, evaluation is organized by output modality, including image, video, audio, and 3D content. (3) From a community perspective, this survey further highlights authoritative leaderboards and foundational tools that have been instrumental in establishing a comprehensive …

Paper
69 citations

Visual agentic reinforcement fine-tuning

Agentic SystemsReinforcement LearningModel Adaptation

Ziyu Liu, Yuhang Zang, Yushan Zou, Zijian Liang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2505.14246, 2025

Full metadata
Publication date
2025/5/20
Journal
arXiv preprint arXiv:2505.14246

Description

A key trend in Large Reasoning Models (e.g., OpenAI's o3) is the native agentic ability to use external tools such as web browsers for searching and writing/executing code for image manipulation to think with images. In the open-source research community, while significant progress has been made in language-only agentic abilities such as function calling and tool integration, the development of multi-modal agentic capabilities that involve truly thinking with images, and their corresponding benchmarks, are still less explored. This work highlights the effectiveness of Visual Agentic Reinforcement Fine-Tuning (Visual-ARFT) for enabling flexible and adaptive reasoning abilities for Large Vision-Language Models (LVLMs). With Visual-ARFT, open-source LVLMs gain the ability to browse websites for real-time information updates and write code to manipulate and analyze input images through cropping, rotation, and other image processing techniques. We also present a Multi-modal Agentic Tool Bench (MAT) with two settings (MAT-Search and MAT-Coding) designed to evaluate LVLMs' agentic search and coding abilities. Our experimental results demonstrate that Visual-ARFT outperforms its baseline by +18.6% F1 / +13.0% EM on MAT-Coding and +10.3% F1 / +8.7% EM on MAT-Search, ultimately surpassing GPT-4o. Visual-ARFT also achieves +29.3 F1% / +25.9% EM gains on existing multi-hop QA benchmarks such as 2Wiki and HotpotQA, demonstrating strong generalization capabilities. Our findings suggest that Visual-ARFT offers a promising path toward building robust and generalizable multimodal agents.

Paper
58 citations

Mm-ifengine: Towards multimodal instruction following

Model AlignmentFoundation Models

Shengyuan Ding, Shenxi Wu, Xiangyu Zhao, Yuhang Zang, Haodong Duan, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Dahua Lin, Jiaqi Wang

ICCV, 2025

Full metadata
Publication date
2025/4/10
Journal
ICCV, 2025

Description

The Instruction Following (IF) ability measures how well Multi-modal Large Language Models (MLLMs) understand exactly what users are telling them and whether they are doing it right. Existing multimodal instruction following training data is scarce, the benchmarks are simple with atomic instructions, and the evaluation strategies are imprecise for tasks demanding exact output constraints. To address this, we present MM-IFEngine, an effective pipeline to generate high-quality image-instruction pairs. Our MMIFEngine pipeline yields large-scale, diverse, and highquality training data MM-IFInstruct-23k, which is suitable for Supervised Fine-Tuning (SFT) and extended as MM-IFDPO-23k for Direct Preference Optimization (DPO). We further introduce MM-IFEval, a challenging and diverse multi-modal instruction-following benchmark that includes (1) both compose-level constraints for output responses and perception …

Paper
55 citations

Mia-dpo: Multi-image augmented direct preference optimization for large vision-language models

Multi-Image ReasoningModel AlignmentModel Adaptation

Ziyu Liu, Yuhang Zang, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Haodong Duan, Conghui He, Yuanjun Xiong, Dahua Lin, Jiaqi Wang

ICLR, 2025

Full metadata
Publication date
2024/10/23
Journal
ICLR, 2025

Description

Visual preference alignment involves training Large Vision-Language Models (LVLMs) to predict human preferences between visual inputs. This is typically achieved by using labeled datasets of chosen/rejected pairs and employing optimization algorithms like direct preference optimization (DPO). Existing visual alignment methods, primarily designed for single-image scenarios, struggle to effectively handle the complexity of multi-image tasks due to the scarcity of diverse training data and the high cost of annotating chosen/rejected pairs. We present Multi-Image Augmented Direct Preference Optimization (MIA-DPO), a visual preference alignment approach that effectively handles multi-image inputs. MIA-DPO mitigates the scarcity of diverse multi-image training data by extending single-image data with unrelated images arranged in grid collages or pic-in-pic formats, significantly reducing the costs associated with multi-image data annotations. Our observation reveals that attention values of LVLMs vary considerably across different images. We use attention values to identify and filter out rejected responses the model may have mistakenly focused on. Our attention-aware selection for constructing the chosen/rejected pairs without relying on (i) human annotation,(ii) extra data, and (iii) external models or APIs. MIA-DPO is compatible with various architectures and outperforms existing methods on five multi-image benchmarks, achieving an average performance boost of 3.0% on LLaVA-v1. 5 and 4.3% on the recent InternLM-XC2. 5. Moreover, MIA-DPO has a minimal effect on the model's ability to understand single images.

Paper
50 citations

A survey of scientific large language models: From data foundations to agent frontiers

Agentic SystemsScientific ReasoningLLM Evaluation

Ming Hu, Chenglong Ma, Wei Li, Wanghan Xu, Jiamin Wu, Jucheng Hu, Tianbin Li, Guohang Zhuang, Jiaqi Liu, Yingzhou Lu, Ying Chen, Chaoyang Zhang, Cheng Tan, Jie Ying, Guocheng Wu, Shujian Gao, Pengcheng Chen, Jiashi Lin, Haitao Wu, Lulu Chen, Fengxiang Wang, Yuanyuan Zhang, Xiangyu Zhao, Feilong Tang, Encheng Su, Junzhi Ning, Xinyao Liu, Ye Du, Changkai Ji, Pengfei Jiang, Cheng Tang, Ziyan Huang, Jiyao Liu, Jiaqi Wei, Yuejin Yang, Xiang Zhang, Guangshuai Wang, Yue Yang, Huihui Xu, Ziyang Chen, Yizhou Wang, Chen Tang, Jianyu Wu, Yuchen Ren, Siyuan Yan, Zhonghua Wang, Zhongxing Xu, Shiyan Su, Shangquan Sun, Runkai Zhao, Zhisheng Zhang, Dingkang Yang, Jinjie Wei, Jiaqi Wang, Jiahao Xu, Jiangtao Yan, Wenhao Tang, Hongze Zhu, Yu Liu, Fudi Wang, Yiqing Shen, Yuanfeng Ji, Yanzhou Su, Tong Xie, Hongming Shan, Chun-Mei Feng, Zhi Hou, Diping Song, Lihao Liu, Yanyan Huang, Lequan Yu, Bin Fu, Shujun Wang, Xiaomeng Li, Hu

arXiv preprint arXiv:2508.21148, 2025

Full metadata
Publication date
2025/8/28
Journal
arXiv preprint arXiv:2508.21148

Description

Scientific Large Language Models (Sci-LLMs) are transforming how knowledge is represented, integrated, and applied in scientific research, yet their progress is shaped by the complex nature of scientific data. This survey presents a comprehensive, data-centric synthesis that reframes the development of Sci-LLMs as a co-evolution between models and their underlying data substrate. We formulate a unified taxonomy of scientific data and a hierarchical model of scientific knowledge, emphasizing the multimodal, cross-scale, and domain-specific challenges that differentiate scientific corpora from general natural language processing datasets. We systematically review recent Sci-LLMs, from general-purpose foundations to specialized models across diverse scientific disciplines, alongside an extensive analysis of over 270 pre-/post-training datasets, showing why Sci-LLMs pose distinct demands -- heterogeneous, multi-scale, uncertainty-laden corpora that require representations preserving domain invariance and enabling cross-modal reasoning. On evaluation, we examine over 190 benchmark datasets and trace a shift from static exams toward process- and discovery-oriented assessments with advanced evaluation protocols. These data-centric analyses highlight persistent issues in scientific data development and discuss emerging solutions involving semi-automated annotation pipelines and expert validation. Finally, we outline a paradigm shift toward closed-loop systems where autonomous agents based on Sci-LLMs actively experiment, validate, and contribute to a living, evolving knowledge base. Collectively, this work provides a …

Paper
40 citations

Image Quality Assessment: From Human to Machine Preference

Model AlignmentImage UnderstandingMultimodal Evaluation

Chunyi Li, Yuan Tian, Xiaoyue Ling, Zicheng Zhang, Haodong Duan, Haoning Wu, Ziheng Jia, Xiaohong Liu, Xiongkuo Min, Guo Lu, Weisi Lin, Guangtao Zhai

CVPR Spotlight, 2025

Full metadata
Publication date
2025/3/13
Journal
CVPR Spotlight, 2025

Description

Image Quality Assessment (IQA) based on human subjective preferences has undergone extensive research in the past decades. However, with the development of communication protocols, the visual data consumption volume of machines has gradually surpassed that of humans. For machines, the preference depends on downstream tasks such as segmentation and detection, rather than visual appeal. Considering the huge gap between human and machine visual systems, this paper proposes the topic: Image Quality Assessment for Machine Vision for the first time. Specifically, we (1) defined the subjective preferences of machines, including downstream tasks, test models, and evaluation metrics; (2) established the Machine Preference Database (MPD), which contains 2.25M fine-grained annotations and 30k reference/distorted image pair instances; (3) verified the performance of mainstream IQA algorithms …

Paper
37 citations

Mg-llava: Towards multi-granularity visual instruction tuning

Model AlignmentImage UnderstandingModel Adaptation

Xiangyu Zhao, Xiangtai Li, Haodong Duan, Haian Huang, Yining Li, Kai Chen, Hua Yang

IEEE Transactions on Circuits and Systems for Video Technology, 2025

Full metadata
Publication date
2025/12/12
Journal
IEEE Transactions on Circuits and Systems for Video Technology
Publisher
IEEE

Description

Multi-modal large language models (MLLMs) have made significant strides in various visual understanding tasks. However, the majority of these models are constrained to process low-resolution images, which limits their effectiveness in perception tasks that necessitate detailed visual information. In our study, we present MG-LLaVA, an innovative MLLM that enhances the model’s visual processing capabilities by incorporating a multi-granularity vision flow, which includes low-resolution, high-resolution, and object-centric features. We propose the integration of an additional high-resolution visual encoder to capture fine-grained details, which are then fused with base visual features through a Conv-Gate fusion network. To further refine the model’s object recognition abilities, we incorporate object-level features derived from bounding boxes identified by offline detectors. Being trained solely on publicly available …

Paper
34 citations

Omnialign-v: Towards enhanced alignment of mllms with human preference

Model AlignmentReward ModelingData-Centric Learning

Xiangyu Zhao, Shengyuan Ding, Zicheng Zhang, Haian Huang, Maosong Cao, Weiyun Wang, Jiaqi Wang, Xinyu Fang, Wenhai Wang, Guangtao Zhai, Haodong Duan, Hua Yang, Kai Chen

ACL, 2025

Full metadata
Publication date
2025/2/25
Journal
ACL, 2025

Description

Recent advancements in open-source multi-modal large language models (MLLMs) have primarily focused on enhancing foundational capabilities, leaving a significant gap in human preference alignment. This paper introduces OmniAlign-V, a comprehensive dataset of 200K high-quality training samples featuring diverse images, complex questions, and varied response formats to improve MLLMs’ alignment with human preferences. We also present MM-AlignBench, a human-annotated benchmark specifically designed to evaluate MLLMs’ alignment with human values. Experimental results show that finetuning MLLMs with OmniAlign-V, using Supervised Fine-Tuning (SFT) or Direct Preference Optimization (DPO), significantly enhances human preference alignment while maintaining or enhancing performance on standard VQA benchmarks, preserving their fundamental capabilities.

Paper Code
33 citations

Towards Storage-Efficient Visual Document Retrieval: An Empirical Study on Reducing Patch-Level Embeddings

Document IntelligenceModel Efficiency

Yubo Ma, Jinsong Li, Yuhang Zang, Xiaobao Wu, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Haodong Duan, Jiaqi Wang, Yixin Cao, Aixin Sun

ACL Findings, 2025

Full metadata
Publication date
2025/6/5
Journal
ACL Findings, 2025

Description

Despite the strong performance of ColPali/ColQwen2 in Visualized Document Retrieval (VDR), its patch-level embedding approach leads to excessive memory usage. This empirical study investigates methods to reduce patch embeddings per page while minimizing performance degradation. We evaluate two token-reduction strategies: token pruning and token merging. Regarding token pruning, we surprisingly observe that a simple random strategy outperforms other sophisticated pruning methods, though still far from satisfactory. Further analysis reveals that pruning is inherently unsuitable for VDR as it requires removing certain page embeddings without query-specific information. Turning to token merging (more suitable for VDR), we search for the optimal combinations of merging strategy across three dimensions and develops Light-ColPali/ColQwen2. It maintains 98.2% of retrieval performance with only 11.8% of original memory usage, and preserves 94.6% effectiveness at 2% memory footprint. We expect our empirical findings and resulting Light-ColPali/ColQwen2 offer valuable insights and establish a competitive baseline for future efficient-VDR research.

Paper
31 citations

Redundancy Principles for MLLMs Benchmarks

Benchmark DesignModel EfficiencyMultimodal Evaluation

Zicheng Zhang, Xiangyu Zhao, Xinyu Fang, Chunyi Li, Xiaohong Liu, Xiongkuo Min, Haodong Duan, Kai Chen, Guangtao Zhai

ACL, 2025

Full metadata
Publication date
2025/1/27
Journal
ACL, 2025

Description

With the rapid iteration of Multi-modality Large Language Models (MLLMs) and the evolving demands of the field, the number of benchmarks produced annually has surged into the hundreds. The rapid growth has inevitably led to significant redundancy among benchmarks. Therefore, it is crucial to take a step back and critically assess the current state of redundancy and propose targeted principles for constructing effective MLLM benchmarks. In this paper, we focus on redundancy from three key perspectives: 1) Redundancy of benchmark capability dimensions, 2) Redundancy in the number of test questions, and 3) Cross-benchmark redundancy within specific domains. Through the comprehensive analysis over hundreds of MLLMs’ performance across more than 20 benchmarks, we aim to quantitatively measure the level of redundancy lies in existing MLLM evaluations, provide valuable insights to guide the future development of MLLM benchmarks, and offer strategies to refine and address redundancy issues effectively.

Paper Code
28 citations

Condor: Enhance LLM Alignment with Knowledge-Driven Data Synthesis and Refinement

Data-Centric LearningModel AdaptationModel Alignment

Maosong Cao, Taolin Zhang, Mo Li, Chuyu Zhang, Yunxin Liu, Haodong Duan, Songyang Zhang, Kai Chen

ACL, 2025

Full metadata
Publication date
2025/1/21
Journal
ACL, 2025

Description

The quality of Supervised Fine-Tuning (SFT) data plays a critical role in enhancing the conversational capabilities of Large Language Models (LLMs). However, the availability of high-quality human-annotated SFT data has become a significant bottleneck for LLMs, necessitating a greater reliance on synthetic training data. In this work, we introduce Condor, a two-stage synthetic data generation framework that incorporates World Knowledge Trees and Self-Reflection Refinement to produce highquality SFT data at scale. Our experimental results demonstrate that a base model finetuned on only 20K Condor-generated samples achieves superior performance compared to instruct model trained with RLHF. The additional refinement stage in Condor further enables iterative self-improvement for LLMs at various scales (up to 72B), validating the effectiveness of our approach. Furthermore, our investigation into the scaling of synthetic data in posttraining reveals substantial unexplored potential for performance improvements, opening promising avenues for future research. 1 2

Paper Code
27 citations

Creation-MMBench: Assessing Context-Aware Creative Intelligence in MLLMs

Multimodal GenerationMultimodal EvaluationBenchmark Design

Xinyu Fang, Zhijian Chen, Kai Lan, Lixin Ma, Shengyuan Ding, Yingji Liang, Xiangyu Zhao, Farong Wen, Zicheng Zhang, Guofeng Zhang, Haodong Duan, Kai Chen, Dahua Lin

ICCV, 2025, 447-456

Full metadata
Publication date
2025
Conference
ICCV, 2025
Pages
447-456

Description

Creativity is a fundamental aspect of intelligence, involving the ability to generate novel and appropriate solutions across diverse contexts. While Large Language Models (LLMs) have been extensively evaluated for their creative capabilities, the assessment of Multimodal Large Language Models (MLLMs) in this domain remains largely unexplored. To address this gap, we introduce CreationMMBench, a multimodal benchmark specifically designed to evaluate the creative capabilities of MLLMs in realworld, image-based tasks. The benchmark comprises 765 test cases spanning 51 fine-grained tasks. To ensure rigorous evaluation, we define instance-specific evaluation criteria for each test case, guiding the assessment of both general response quality and factual consistency with visual inputs. Experimental results reveal that current open-source MLLMs significantly underperform compared to proprietary models in …

Paper
22 citations

Booststep: Boosting mathematical capability of large language models via improved single-step reasoning

Scientific ReasoningFoundation Models

Beichen Zhang, Yuhong Liu, Xiaoyi Dong, Yuhang Zang, Pan Zhang, Haodong Duan, Yuhang Cao, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2501.03226, 2025

Full metadata
Publication date
2025/1/6
Journal
arXiv preprint arXiv:2501.03226

Description

Large language models (LLMs) have demonstrated impressive ability in solving complex mathematical problems with multi-step reasoning and can be further enhanced with well-designed in-context learning (ICL) examples. However, this potential is often constrained by two major challenges in ICL: granularity mismatch and irrelevant information. We observe that while LLMs excel at decomposing mathematical problems, they often struggle with reasoning errors in fine-grained steps. Moreover, ICL examples retrieved at the question level may omit critical steps or even mislead the model with irrelevant details. To address this issue, we propose BoostStep, a method that enhances reasoning accuracy through step-aligned ICL, a novel mechanism that carefully aligns retrieved reference steps with the corresponding reasoning steps. Additionally, BoostStep incorporates an effective "first-try" strategy to deliver exemplars highly relevant to the current state of reasoning. BoostStep is a flexible and powerful method that integrates seamlessly with chain-of-thought (CoT) and tree search algorithms, refining both candidate selection and decision-making. Empirical results show that BoostStep improves GPT-4o's CoT performance by 4.6% across mathematical benchmarks, significantly surpassing traditional few-shot learning's 1.2%. Moreover, it can achieve an additional 7.5\% gain combined with tree search. Surprisingly, it enhances state-of-the-art LLMs to solve challenging math problems using simpler examples. It improves DeepSeek-R1-671B's performance on AIME by 2.2%, leveraging simple examples only from the MATH dataset.

Paper
21 citations

Affordance Benchmark for MLLMs

Spatial ReasoningMultimodal EvaluationBenchmark Design

Junying Wang, Wenzhe Li, Yalun Wu, Yingji Liang, Yijin Guo, Chunyi Li, Haodong Duan, Zicheng Zhang, Guangtao Zhai

arXiv preprint arXiv:2506.00893, 2025

Full metadata
Publication date
2025/6/1
Journal
arXiv preprint arXiv:2506.00893

Description

Affordance theory suggests that environments inherently provide action possibilities shaping perception and behavior. While Multimodal Large Language Models (MLLMs) achieve strong performance in vision-language tasks, their ability to perceive affordance, which is crucial for intuitive and safe interactions, remains underexplored. To address this, we introduce A4Bench, a novel benchmark designed to evaluate the affordance perception abilities of MLLMs across two dimensions: 1) Constitutive Affordance, assessing understanding of inherent object properties through 1,282 question-answer pairs spanning nine sub-disciplines, and 2) Transformative Affordance, probing dynamic and contextual nuances (e.g., misleading, time-dependent, cultural, or individual-specific affordance) with 718 challenging question-answer pairs. We evaluate 17 MLLMs (nine proprietary and eight open-source) and compare them to …

Paper
11 citations

Information Density Principle for MLLM Benchmarks

Long-Context ModelingMultimodal EvaluationBenchmark Design

Chunyi Li, Xiaozhe Li, Zicheng Zhang, Yuan Tian, Ziheng Jia, Xiaohong Liu, Xiongkuo Min, Jia Wang, Haodong Duan, Kai Chen, Guangtao Zhai

ICCV, 2025

Full metadata
Publication date
2025/3/13
Journal
ICCV, 2025

Description

With the emergence of Multimodal Large Language Models (MLLMs), hundreds of benchmarks have been developed to ensure the reliability of MLLMs in downstream tasks. However, the evaluation mechanism itself may not be reliable. For developers of MLLMs, questions remain about which benchmark to use and whether the test results meet their requirements. Therefore, we propose a critical principle of Information Density, which examines how much insight a benchmark can provide for the development of MLLMs. We characterize it from four key dimensions: (1) Fallacy, (2 Difficulty, (3) Redundancy, (4) Diversity. Through a comprehensive analysis of more than 10,000 samples, we measured the information density of 19 MLLM benchmarks. Experiments show that using the latest benchmarks in testing can provide more insight compared to previous ones, but there is still room for improvement in their …

Paper
10 citations

ATLAS: A High-Difficulty, Multidisciplinary Benchmark for Frontier Scientific Reasoning

Scientific ReasoningLLM EvaluationBenchmark Design

Hongwei Liu, Junnan Liu, Shudong Liu, Haodong Duan, Yuqiang Li, Mao Su, Xiaohong Liu, Guangtao Zhai, Xinyu Fang, Qianhong Ma, Taolin Zhang, Zihan Ma, Yufeng Zhao, Peiheng Zhou, Linchen Xiao, Wenlong Zhang, Shijie Zhou, Xingjian Ma, Siqi Sun, Jiaye Ge, Meng Li, Yuhong Liu, Jianxin Dong, Jiaying Li, Hui Wu, Hanwen Liang, Jintai Lin, Yanting Wang, Jie Dong, Tong Zhu, Tianfan Fu, Conghui He, Qi Zhang, Songyang Zhang, Lei Bai, Kai Chen

arXiv preprint arXiv:2511.14366, 2025

Full metadata
Publication date
2025/11/18
Journal
arXiv preprint arXiv:2511.14366

Description

The rapid advancement of Large Language Models (LLMs) has led to performance saturation on many established benchmarks, questioning their ability to distinguish frontier models. Concurrently, existing high-difficulty benchmarks often suffer from narrow disciplinary focus, oversimplified answer formats, and vulnerability to data contamination, creating a fidelity gap with real-world scientific inquiry. To address these challenges, we introduce ATLAS (AGI-Oriented Testbed for Logical Application in Science), a large-scale, high-difficulty, and cross-disciplinary evaluation suite composed of approximately 800 original problems. Developed by domain experts (PhD-level and above), ATLAS spans seven core scientific fields: mathematics, physics, chemistry, biology, computer science, earth science, and materials science. Its key features include: (1) High Originality and Contamination Resistance, with all questions newly created or substantially adapted to prevent test data leakage; (2) Cross-Disciplinary Focus, designed to assess models' ability to integrate knowledge and reason across scientific domains; (3) High-Fidelity Answers, prioritizing complex, open-ended answers involving multi-step reasoning and LaTeX-formatted expressions over simple multiple-choice questions; and (4) Rigorous Quality Control, employing a multi-stage process of expert peer review and adversarial testing to ensure question difficulty, scientific value, and correctness. We also propose a robust evaluation paradigm using a panel of LLM judges for automated, nuanced assessment of complex answers. Preliminary results on leading models demonstrate ATLAS's …

Paper
9 citations

Spark: Synergistic policy and reward co-evolving framework

Reward ModelingModel Adaptation

Ziyu Liu, Yuhang Zang, Shengyuan Ding, Yuhang Cao, Xiaoyi Dong, Haodong Duan, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2509.22624, 2025

Full metadata
Publication date
2025/9/26
Journal
arXiv preprint arXiv:2509.22624

Description

Recent Large Language Models (LLMs) and Large Vision-Language Models (LVLMs) increasingly use Reinforcement Learning (RL) for post-pretraining, such as RL with Verifiable Rewards (RLVR) for objective tasks and RL from Human Feedback (RLHF) for subjective tasks. However, RLHF incurs high costs and potential reward-policy mismatch due to reliance on human preferences, while RLVR still wastes supervision by discarding rollouts and correctness signals after each update. To address these challenges, we introduce the Synergistic Policy And Reward Co-Evolving Framework (SPARK), an efficient, on-policy, and stable method that builds on RLVR. Instead of discarding rollouts and correctness data, SPARK recycles this valuable information to simultaneously train the model itself as a generative reward model. This auxiliary training uses a mix of objectives, such as pointwise reward score, pairwise comparison, and evaluation conditioned on further-reflection responses, to teach the model to evaluate and improve its own responses. Our process eliminates the need for a separate reward model and costly human preference data. SPARK creates a positive co-evolving feedback loop: improved reward accuracy yields better policy gradients, which in turn produce higher-quality rollouts that further refine the reward model. Our unified framework supports test-time scaling via self-reflection without external reward models and their associated costs. We show that SPARK achieves significant performance gains on multiple LLM and LVLM models and multiple reasoning, reward models, and general benchmarks. For example, SPARK-VL-7B …

Paper
9 citations

GOBench: Benchmarking Geometric Optics Generation and Understanding of MLLMs

Spatial ReasoningMultimodal GenerationMultimodal Evaluation

Xiaorong Zhu, Ziheng Jia, Jiarui Wang, Xiangyu Zhao, Haodong Duan, Xiongkuo Min, Jia Wang, Zicheng Zhang, Guangtao Zhai

ACMMM Datasets, 2025

Full metadata
Publication date
2025/6/1
Journal
ACMMM Datasets, 2025

Description

The rapid evolution of Multi-modality Large Language Models (MLLMs) is driving significant advancements in visual understanding and generation. Nevertheless, a comprehensive assessment of their capabilities, concerning the fine-grained physical principles especially in geometric optics, remains underexplored. To address this gap, we introduce GOBench, the first benchmark to systematically evaluate MLLMs' ability across two tasks: 1) Generating Optically Authentic Imagery and 2) Understanding Underlying Optical Phenomena. We curate high-quality prompts of geometric optical scenarios and use MLLMs to construct the GOBench-Gen-1k dataset. We then organize subjective experiments to assess the generated imagery based on Optical Authenticity, Aesthetic Quality, and Instruction Fidelity, revealing MLLMs' generation flaws that violate optical principles. For the understanding task, we apply crafted …

Paper
4 citations

Gdi-bench: a benchmark for general document intelligence with vision and reasoning decoupling

Document IntelligenceVisual ReasoningBenchmark Design

Siqi Li, Yufan Shen, Xiangnan Chen, Jiayi Chen, Hengwei Ju, Haodong Duan, Song Mao, Hongbin Zhou, Bo Zhang, Bin Fu, Pinlong Cai, Licheng Wen, Botian Shi, Yong Liu, Xinyu Cai, Yu Qiao

arXiv preprint arXiv:2505.00063, 2025

Full metadata
Publication date
2025/4/30
Journal
arXiv preprint arXiv:2505.00063

Description

The rapid advancement of multimodal large language models (MLLMs) has profoundly impacted the document domain, creating a wide array of application scenarios. This progress highlights the need for a comprehensive benchmark to evaluate these models' capabilities across various document-specific tasks. However, existing benchmarks often fail to locate specific model weaknesses or guide systematic improvements. To bridge this gap, we introduce a General Document Intelligence Benchmark (GDI-Bench), featuring 2.3k images across 9 key scenarios and 19 document-specific tasks. By decoupling visual complexity and reasoning complexity, the GDI-Bench structures graded tasks that allow performance assessment by difficulty, aiding in model weakness identification and optimization guidance. We evaluate various open-source and closed-source models on GDI-Bench, conducting decoupled analyses in the visual and reasoning domains, revealing their strengths and weaknesses. To address the diverse tasks and domains in the GDI-Bench, we propose a GDI-Model that mitigates catastrophic forgetting during the supervised fine-tuning (SFT) process through an intelligence-preserving training strategy, thereby reinforcing the inherent weaknesses of the base model. Our model achieves state-of-the-art performance on previous benchmarks and the GDI-Bench. Both our benchmark and models are or will be open-sourced on https://huggingface.co/GDIBench.

Paper
4 citations

SciEvalKit: An Open-source Evaluation Toolkit for Scientific General Intelligence

Scientific ReasoningEvaluation InfrastructureLLM Evaluation

Yiheng Wang, Yixin Chen, Shuo Li, Yifan Zhou, Bo Liu, Hengjian Gao, Jiakang Yuan, Jia Bu, Wanghan Xu, Yuhao Zhou, Xiangyu Zhao, Zhiwang Zhou, Fengxiang Wang, Haodong Duan, Songyang Zhang, Jun Yao, Han Deng, Yizhou Wang, Jiabei Xiao, Jiaqi Liu, Encheng Su, Yujie Liu, Weida Wang, Junchi Yao, Shenghe Zheng, Haoran Sun, Runmin Ma, Xiangchao Yan, Bo Zhang, Dongzhan Zhou, Shufei Zhang, Peng Ye, Xiaosong Wang, Shixiang Tang, Wenlong Zhang, Lei Bai

arXiv preprint arXiv:2512.22334, 2025

Full metadata
Publication date
2025/12/26
Journal
arXiv preprint arXiv:2512.22334

Description

We introduce SciEvalKit, a unified benchmarking toolkit designed to evaluate AI models for science across a broad range of scientific disciplines and task capabilities. Unlike general-purpose evaluation platforms, SciEvalKit focuses on the core competencies of scientific intelligence, including Scientific Multimodal Perception, Scientific Multimodal Reasoning, Scientific Multimodal Understanding, Scientific Symbolic Reasoning, Scientific Code Generation, Science Hypothesis Generation and Scientific Knowledge Understanding. It supports six major scientific domains, spanning from physics and chemistry to astronomy and materials science. SciEvalKit builds a foundation of expert-grade scientific benchmarks, curated from real-world, domain-specific datasets, ensuring that tasks reflect authentic scientific challenges. The toolkit features a flexible, extensible evaluation pipeline that enables batch evaluation across models and datasets, supports custom model and dataset integration, and provides transparent, reproducible, and comparable results. By bridging capability-based evaluation and disciplinary diversity, SciEvalKit offers a standardized yet customizable infrastructure to benchmark the next generation of scientific foundation models and intelligent agents. The toolkit is open-sourced and actively maintained to foster community-driven development and progress in AI4Science.

Paper
2 citations

2024

MMBench: Is Your Multi-modal Model an All-around Player?

Multimodal EvaluationBenchmark DesignVisual Reasoning

Yuan Liu, Haodong Duan, Yuanhan Zhang, Bo Li, Songyang Zhang, Wangbo Zhao, Yike Yuan, Jiaqi Wang, Conghui He, Ziwei Liu, Kai Chen, Dahua Lin

European Conference on Computer Vision (ECCV), 2024

Full metadata
Publication date
2023/7/12
Journal
ECCV Oral, 2024

Description

Large vision-language models (VLMs) have recently achieved remarkable progress, exhibiting impressive multimodal perception and reasoning abilities. However, effectively evaluating these large VLMs remains a major challenge, hindering future development in this domain. Traditional benchmarks like VQAv2 or COCO Caption provide quantitative performance measurements but lack fine-grained ability assessment and robust evaluation metrics. Meanwhile, subjective benchmarks, such as OwlEval, offer comprehensive evaluations of a model’s abilities by incorporating human labor, which is not scalable and may display significant bias. In response to these challenges, we propose MMBench, a bilingual benchmark for assessing the multi-modal capabilities of VLMs. MMBench methodically develops a comprehensive evaluation pipeline, primarily comprised of the following key features: 1. MMBench is …

Paper Code
3015 citations

Are We on the Right Way for Evaluating Large Vision-Language Models?

Benchmark DesignMultimodal EvaluationVisual Reasoning

Lin Chen, Jinsong Li, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Zehui Chen, Haodong Duan, Jiaqi Wang, Yu Qiao, Dahua Lin, Feng Zhao

Conference on Neural Information Processing Systems (NeurIPS), 2024

Full metadata
Publication date
2024/3/29
Journal
NeurIPS, 2024

Description

Large vision-language models (LVLMs) have recently achieved rapid progress, sparking numerous studies to evaluate their multi-modal capabilities. However, we dig into current evaluation works and identify two primary issues: 1) Visual content is unnecessary for many samples. The answers can be directly inferred from the questions and options, or the world knowledge embedded in LLMs. This phenomenon is prevalent across current benchmarks. For instance, GeminiPro achieves 42.7% on the MMMU benchmark without any visual input, and outperforms the random choice baseline across six benchmarks near 24% on average. 2) Unintentional data leakage exists in LLM and LVLM training. LLM and LVLM could still answer some visual-necessary questions without visual content, indicating the memorizing of these samples within large-scale training data. For example, Sphinx-X-MoE gets 43.6% on MMMU without accessing images, surpassing its LLM backbone with 17.9%. Both problems lead to misjudgments of actual multi-modal gains and potentially misguide the study of LVLM. To this end, we present MMStar, an elite vision-indispensable multi-modal benchmark comprising 1,500 samples meticulously selected by humans. MMStar benchmarks 6 core capabilities and 18 detailed axes, aiming to evaluate LVLMs' multi-modal capacities with carefully balanced and purified samples. These samples are first roughly selected from current benchmarks with an automated pipeline, human review is then involved to ensure each curated sample exhibits visual dependency, minimal data leakage, and requires advanced multi-modal capabilities …

Paper Code
1294 citations

Internlm2 technical report

Foundation ModelsModel Adaptation

Zheng Cai, Maosong Cao, Haojiong Chen, Kai Chen, Keyu Chen, Xin Chen, Xun Chen, Zehui Chen, Zhi Chen, Pei Chu, Xiaoyi Dong, Haodong Duan, Qi Fan, Zhaoye Fei, Yang Gao, Jiaye Ge, Chenya Gu, Yuzhe Gu, Tao Gui, Aijia Guo, Qipeng Guo, Conghui He, Yingfan Hu, Ting Huang, Tao Jiang, Penglong Jiao, Zhenjiang Jin, Zhikai Lei, Jiaxing Li, Jingwen Li, Linyang Li, Shuaibin Li, Wei Li, Yining Li, Hongwei Liu, Jiangning Liu, Jiawei Hong, Kaiwen Liu, Kuikun Liu, Xiaoran Liu, Chengqi Lv, Haijun Lv, Kai Lv, Li Ma, Runyuan Ma, Zerun Ma, Wenchang Ning, Linke Ouyang, Jiantao Qiu, Yuan Qu, Fukai Shang, Yunfan Shao, Demin Song, Zifan Song, Zhihao Sui, Peng Sun, Yu Sun, Huanze Tang, Bin Wang, Guoteng Wang, Jiaqi Wang, Jiayu Wang, Rui Wang, Yudong Wang, Ziyi Wang, Xingjian Wei, Qizhen Weng, Fan Wu, Yingtong Xiong, Chao Xu, Ruiliang Xu, Hang Yan, Yirong Yan, Xiaogui Yang, Haochen Ye, Huaiyuan Ying, Jia Yu, Jing Yu, Yuhang Zang

arXiv preprint arXiv:2403.17297, 2024

Full metadata
Publication date
2024/3/26
Journal
arXiv preprint arXiv:2403.17297

Description

The evolution of Large Language Models (LLMs) like ChatGPT and GPT-4 has sparked discussions on the advent of Artificial General Intelligence (AGI). However, replicating such advancements in open-source models has been challenging. This paper introduces InternLM2, an open-source LLM that outperforms its predecessors in comprehensive evaluations across 6 dimensions and 30 benchmarks, long-context modeling, and open-ended subjective evaluations through innovative pre-training and optimization techniques. The pre-training process of InternLM2 is meticulously detailed, highlighting the preparation of diverse data types including text, code, and long-context data. InternLM2 efficiently captures long-term dependencies, initially trained on 4k tokens before advancing to 32k tokens in pre-training and fine-tuning stages, exhibiting remarkable performance on the 200k ``Needle-in-a-Haystack" test. InternLM2 is further aligned using Supervised Fine-Tuning (SFT) and a novel Conditional Online Reinforcement Learning from Human Feedback (COOL RLHF) strategy that addresses conflicting human preferences and reward hacking. By releasing InternLM2 models in different training stages and model sizes, we provide the community with insights into the model's evolution.

Paper
801 citations

VLMEvalKit: An Open-Source Toolkit for Evaluating Large Multi-Modality Models

Evaluation InfrastructureMultimodal Evaluation

Haodong Duan, Junming Yang, Yuxuan Qiao, Xinyu Fang, Lin Chen, Yuan Liu, Xiaoyi Dong, Yuhang Zang, Pan Zhang, Jiaqi Wang, Dahua Lin, Kai Chen

ACM International Conference on Multimedia (MM), 2024

Full metadata
Publication date
2024/7/16
Journal
MMOpen, 2024

Description

We present VLMEvalKit: an open-source toolkit for evaluating large multi-modality models based on PyTorch. The toolkit aims to provide a user-friendly and comprehensive framework for researchers and developers to evaluate existing multi-modality models and publish reproducible evaluation results. In VLMEvalKit, we implement over 70 different large multi-modality models, including both proprietary APIs and open-source models, as well as more than 20 different multi-modal benchmarks. By implementing a single interface, new models can be easily added to the toolkit, while the toolkit automatically handles the remaining workloads, including data preparation, distributed inference, prediction post-processing, and metric calculation. Although the toolkit is currently mainly used for evaluating large vision-language models, its design is compatible with future updates that incorporate additional modalities, such as …

Paper Code
660 citations

ShareGPT4Video: Improving Video Understanding and Generation with Better Captions

Video UnderstandingData-Centric LearningMultimodal Generation

Lin Chen, Xilin Wei, Jinsong Li, Xiaoyi Dong, Pan Zhang, Yuhang Zang, Zehui Chen, Haodong Duan, Bin Lin, Zhenyu Tang, Li Yuan, Yu Qiao, Dahua Lin, Feng Zhao, Jiaqi Wang

Conference on Neural Information Processing Systems (NeurIPS), Datasets & Benchmarks Track, 2024

Full metadata
Publication date
2024/6/6
Journal
NeurIPS D&B, 2024

Description

We present the ShareGPT4Video series, aiming to facilitate the video understanding of large video-language models (LVLMs) and the video generation of text-to-video models (T2VMs) via dense and precise captions. The series comprises: 1) ShareGPT4Video, 40K GPT4V annotated dense captions of videos with various lengths and sources, developed through carefully designed data filtering and annotating strategy. 2) ShareCaptioner-Video, an efficient and capable captioning model for arbitrary videos, with 4.8 M high-quality aesthetic videos annotated by it. 3) ShareGPT4Video-8B, a simple yet superb LVLM that reached SOTA performance on three advancing video benchmarks. To achieve this, taking aside the non-scalable costly human annotators, we find using GPT4V to caption video with a naive multi-frame or frame-concatenation input strategy leads to less detailed and sometimes temporal-confused results. We argue the challenge of designing a high-quality video captioning strategy lies in three aspects: 1) Inter-frame precise temporal change understanding. 2) Intra-frame detailed content description. 3) Frame-number scalability for arbitrary-length videos. To this end, we meticulously designed a differential video captioning strategy, which is stable, scalable, and efficient for generating captions for videos with arbitrary resolution, aspect ratios, and length. Based on it, we construct ShareGPT4Video, which contains 40K high-quality videos spanning a wide range of categories, and the resulting captions encompass rich world knowledge, object attributes, camera movements, and crucially, detailed and precise temporal descriptions of …

Paper Code
540 citations

Internlm-xcomposer2: Mastering free-form text-image composition and comprehension in vision-language large model

Multimodal GenerationFoundation Models

Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Bin Wang, Linke Ouyang, Xilin Wei, Songyang Zhang, Haodong Duan, Maosong Cao, Wenwei Zhang, Yining Li, Hang Yan, Yang Gao, Xinyue Zhang, Wei Li, Jingwen Li, Kai Chen, Conghui He, Xingcheng Zhang, Yu Qiao, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2401.16420, 2024

Full metadata
Publication date
2024/1/29
Journal
arXiv preprint arXiv:2401.16420

Description

We introduce InternLM-XComposer2, a cutting-edge vision-language model excelling in free-form text-image composition and comprehension. This model goes beyond conventional vision-language understanding, adeptly crafting interleaved text-image content from diverse inputs like outlines, detailed textual specifications, and reference images, enabling highly customizable content creation. InternLM-XComposer2 proposes a Partial LoRA (PLoRA) approach that applies additional LoRA parameters exclusively to image tokens to preserve the integrity of pre-trained language knowledge, striking a balance between precise vision understanding and text composition with literary talent. Experimental results demonstrate the superiority of InternLM-XComposer2 based on InternLM2-7B in producing high-quality long-text multi-modal content and its exceptional vision-language understanding performance across various benchmarks, where it not only significantly outperforms existing multimodal models but also matches or even surpasses GPT-4V and Gemini Pro in certain assessments. This highlights its remarkable proficiency in the realm of multimodal understanding. The InternLM-XComposer2 model series with 7B parameters are publicly available at https://github.com/InternLM/InternLM-XComposer.

Paper
471 citations

Internlm-xcomposer-2.5: A versatile large vision language model supporting long-contextual input and output

Long-Context ModelingFoundation Models

Pan Zhang, Xiaoyi Dong, Yuhang Zang, Yuhang Cao, Rui Qian, Lin Chen, Qipeng Guo, Haodong Duan, Bin Wang, Linke Ouyang, Songyang Zhang, Wenwei Zhang, Yining Li, Yang Gao, Peng Sun, Xinyue Zhang, Wei Li, Jingwen Li, Wenhai Wang, Hang Yan, Conghui He, Xingcheng Zhang, Kai Chen, Jifeng Dai, Yu Qiao, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2407.03320, 2024

Full metadata
Publication date
2024/7/3
Journal
arXiv preprint arXiv:2407.03320

Description

We present InternLM-XComposer-2.5 (IXC-2.5), a versatile large-vision language model that supports long-contextual input and output. IXC-2.5 excels in various text-image comprehension and composition applications, achieving GPT-4V level capabilities with merely 7B LLM backend. Trained with 24K interleaved image-text contexts, it can seamlessly extend to 96K long contexts via RoPE extrapolation. This long-context capability allows IXC-2.5 to excel in tasks requiring extensive input and output contexts. Compared to its previous 2.0 version, InternLM-XComposer-2.5 features three major upgrades in vision-language comprehension: (1) Ultra-High Resolution Understanding, (2) Fine-Grained Video Understanding, and (3) Multi-Turn Multi-Image Dialogue. In addition to comprehension, IXC-2.5 extends to two compelling applications using extra LoRA parameters for text-image composition: (1) Crafting Webpages and (2) Composing High-Quality Text-Image Articles. IXC-2.5 has been evaluated on 28 benchmarks, outperforming existing open-source state-of-the-art models on 16 benchmarks. It also surpasses or competes closely with GPT-4V and Gemini Pro on 16 key tasks. The InternLM-XComposer-2.5 is publicly available at https://github.com/InternLM/InternLM-XComposer.

Paper
331 citations

ProSA: Assessing and Understanding the Prompt Sensitivity of LLMs

LLM EvaluationBenchmark DesignModel Alignment

Jingming Zhuo, Songyang Zhang, Xinyu Fang, Haodong Duan, Dahua Lin, Kai Chen

Empirical Methods in Natural Language Processing (EMNLP), Findings, 2024

Full metadata
Publication date
2024/10/16
Journal
EMNLP Findings, 2024

Description

Large language models (LLMs) have demonstrated impressive capabilities across various tasks, but their performance is highly sensitive to the prompts utilized. This variability poses challenges for accurate assessment and user satisfaction. Current research frequently overlooks instance-level prompt variations and their implications on subjective evaluations. To address these shortcomings, we introduce ProSA, a framework designed to evaluate and comprehend prompt sensitivity in LLMs. ProSA incorporates a novel sensitivity metric, PromptSensiScore, and leverages decoding confidence to elucidate underlying mechanisms. Our extensive study, spanning multiple tasks, uncovers that prompt sensitivity fluctuates across datasets and models, with larger models exhibiting enhanced robustness. We observe that few-shot examples can alleviate this sensitivity issue, and subjective evaluations are also susceptible to prompt sensitivities, particularly in complex, reasoning-oriented tasks. Furthermore, our findings indicate that higher model confidence correlates with increased prompt robustness. We believe this work will serve as a helpful tool in studying prompt sensitivity of LLMs. The project is released at: https://github. com/open-compass/ProSA.

Paper Code
313 citations

MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding

Video UnderstandingLong-Context ModelingMultimodal Evaluation

Xinyu Fang, Kangrui Mao, Haodong Duan, Xiangyu Zhao, Yining Li, Dahua Lin, Kai Chen

Conference on Neural Information Processing Systems (NeurIPS), Datasets & Benchmarks Track, 2024

Full metadata
Publication date
2024/6/20
Journal
NeurIPS D&B, 2024

Description

The advent of large vision-language models (LVLMs) has spurred research into their applications in multi-modal contexts, particularly in video understanding. Traditional VideoQA benchmarks, despite providing quantitative metrics, often fail to encompass the full spectrum of video content and inadequately assess models' temporal comprehension. To address these limitations, we introduce MMBench-Video, a quantitative benchmark designed to rigorously evaluate LVLMs' proficiency in video understanding. MMBench-Video incorporates lengthy videos from YouTube and employs free-form questions, mirroring practical use cases. The benchmark is meticulously crafted to probe the models' temporal reasoning skills, with all questions human-annotated according to a carefully constructed ability taxonomy. We employ GPT-4 for automated assessment, demonstrating superior accuracy and robustness over earlier LLM-based evaluations. Utilizing MMBench-Video, we have conducted comprehensive evaluations that include both proprietary and open-source LVLMs for images and videos. MMBench-Video stands as a valuable resource for the research community, facilitating improved evaluation of LVLMs and catalyzing progress in the field of video understanding.

Paper Code
242 citations

InternLM-XComposer2-4KHD: A Pioneering Large Vision-Language Model Handling Resolutions from 336 Pixels to 4K HD

Long-Context ModelingDocument IntelligenceFoundation Models

Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Bin Wang, Linke Ouyang, Songyang Zhang, Haodong Duan, Wenwei Zhang, Yining Li, Hang Yan, Yang Gao, Zhe Chen, Xinyue Zhang, Wei Li, Jingwen Li, Wenhai Wang, Kai Chen, Conghui He, Xingcheng Zhang, Jifeng Dai, Yu Qiao, Dahua Lin, Jiaqi Wang

Conference on Neural Information Processing Systems (NeurIPS), 2024

Full metadata
Publication date
2024/4/9
Journal
NeurIPS, 2024

Description

The Large Vision-Language Model (LVLM) field has seen significant advancements, yet its progression has been hindered by challenges in comprehending fine-grained visual content due to limited resolution. Recent efforts have aimed to enhance the high-resolution understanding capabilities of LVLMs, yet they remain capped at approximately 15001500 pixels and constrained to a relatively narrow resolution range. This paper represents InternLM-XComposer2-4KHD, a groundbreaking exploration into elevating LVLM resolution capabilities up to 4K HD (3840× 1600) and beyond. Concurrently, considering the ultra-high resolution may not be necessary in all scenarios, it supports a wide range of diverse resolutions from 336 pixels to 4K standard, significantly broadening its scope of applicability. Specifically, this research advances the patch division paradigm by introducing a novel extension: dynamic resolution with automatic patch configuration. It maintains the training image aspect ratios while automatically varying patch counts and configuring layouts based on a pre-trained Vision Transformer (ViT)(336336), leading to dynamic training resolution from 336 pixels to 4K standard. Our research demonstrates that scaling training resolution up to 4K HD leads to consistent performance enhancements without hitting the ceiling of potential improvements. InternLM-XComposer2-4KHD shows superb capability that matches or even surpasses GPT-4V and Gemini Pro in 10 of the 16 benchmarks.

Paper Code
239 citations

GMAI-MMBench: A Comprehensive Multimodal Evaluation Benchmark Towards General Medical AI

Medical AIMultimodal EvaluationBenchmark Design

Pengcheng Chen, Jin Ye, Guoan Wang, Yanjun Li, Zhongying Deng, Wei Li, Tianbin Li, Haodong Duan, Ziyan Huang, Yanzhou Su, Benyou Wang, Shaoting Zhang, Bin Fu, Jianfei Cai, Bohan Zhuang, Eric J Seibel, Junjun He, Yu Qiao

Conference on Neural Information Processing Systems (NeurIPS), Datasets & Benchmarks Track, 2024

Full metadata
Publication date
2024/8/6
Journal
NeurIPS D&B, 2024

Description

Large Vision-Language Models (LVLMs) are capable of handling diverse data types such as imaging, text, and physiological signals, and can be applied in various fields. In the medical field, LVLMs have a high potential to offer substantial assistance for diagnosis and treatment. Before that, it is crucial to develop benchmarks to evaluate LVLMs' effectiveness in various medical applications. Current benchmarks are often built upon specific academic literature, mainly focusing on a single domain, and lacking varying perceptual granularities. Thus, they face specific challenges, including limited clinical relevance, incomplete evaluations, and insufficient guidance for interactive LVLMs. To address these limitations, we developed the GMAI-MMBench, the most comprehensive general medical AI benchmark with well-categorized data structure and multi-perceptual granularity to date. It is constructed from 284 datasets across 38 medical image modalities, 18 clinical-related tasks, 18 departments, and 4 perceptual granularities in a Visual Question Answering (VQA) format. Additionally, we implemented a lexical tree structure that allows users to customize evaluation tasks, accommodating various assessment needs and substantially supporting medical AI research and applications. We evaluated 50 LVLMs, and the results show that even the advanced GPT-4o only achieves an accuracy of 53.96\%, indicating significant room for improvement. Moreover, we identified five key insufficiencies in current cutting-edge LVLMs that need to be addressed to advance the development of better medical applications. We believe that GMAI-MMBench will stimulate …

Paper Code
175 citations

MathBench: Evaluating the Theory and Application Proficiency of LLMs with a Hierarchical Mathematics Benchmark

Scientific ReasoningLLM EvaluationBenchmark Design

Hongwei Liu, Zilong Zheng, Yuxuan Qiao, Haodong Duan, Zhiwei Fei, Fengzhe Zhou, Wenwei Zhang, Songyang Zhang, Dahua Lin, Kai Chen

Annual Meeting of the Association for Computational Linguistics (ACL), Findings, 2024

Full metadata
Publication date
2024/5/20
Journal
ACL Findings, 2024

Description

Recent advancements in large language models (LLMs) have showcased significant improvements in mathematics. However, traditional math benchmarks like GSM8k offer a unidimensional perspective, which fall short in providing a holistic assessment of the LLMs’ math capabilities. To address this gap, we introduce MathBench, a new benchmark that rigorously assesses the mathematical capabilities of large language models. MathBench spans a wide range of mathematical disciplines, offering a detailed evaluation of both theoretical understanding and practical problem-solving skills. The benchmark progresses through five distinct stages, from basic arithmetic to college mathematics, and is structured to evaluate models at various depths of knowledge. Each stage includes theoretical questions and application problems, allowing us to measure a model’s mathematical proficiency and its ability to apply concepts in practical scenarios. MathBench aims to enhance the evaluation of LLMs’ mathematical abilities, providing a nuanced view of their knowledge understanding levels and problem solving skills in a bilingual context.

Paper Code
167 citations

Mme-survey: A comprehensive survey on evaluation of multimodal llms

Multimodal EvaluationLLM EvaluationBenchmark Design

Chaoyou Fu, Yi-Fan Zhang, Shukang Yin, Bo Li, Xinyu Fang, Sirui Zhao, Haodong Duan, Xing Sun, Ziwei Liu, Liang Wang, Caifeng Shan, Ran He

arXiv preprint arXiv:2411.15296, 2024

Full metadata
Publication date
2024/11/22
Source
arXiv preprint arXiv:2411.15296

Description

As a prominent direction of Artificial General Intelligence (AGI), Multimodal Large Language Models (MLLMs) have garnered increased attention from both industry and academia. Building upon pre-trained LLMs, this family of models further develops multimodal perception and reasoning capabilities that are impressive, such as writing code given a flow chart or creating stories based on an image. In the development process, evaluation is critical since it provides intuitive feedback and guidance on improving models. Distinct from the traditional train-eval-test paradigm that only favors a single task like image classification, the versatility of MLLMs has spurred the rise of various new benchmarks and evaluation methods. In this paper, we aim to present a comprehensive survey of MLLM evaluation, discussing four key aspects: 1) the summarised benchmarks types divided by the evaluation capabilities, including foundation capabilities, model self-analysis, and extented applications; 2) the typical process of benchmark counstruction, consisting of data collection, annotation, and precautions; 3) the systematic evaluation manner composed of judge, metric, and toolkit; 4) the outlook for the next benchmark. This work aims to offer researchers an easy grasp of how to effectively evaluate MLLMs according to different needs and to inspire better evaluation methods, thereby driving the progress of MLLM research.

Paper
121 citations

BotChat: Evaluating LLMs' Capabilities of Having Multi-Turn Dialogues

LLM EvaluationAgentic SystemsModel Alignment

Haodong Duan, Jueqi Wei, Chonghua Wang, Hongwei Liu, Yixiao Fang, Songyang Zhang, Dahua Lin, Kai Chen

Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), 2024

Full metadata
Publication date
2023/10/20
Journal
NAACL Findings, 2024

Description

In the realm of modern Large Language Models (LLMs), facilitating high-quality, multi-turn dialogues with humans represents a cornerstone feature. However, human-based evaluation of such a capability involves substantial manual effort. This study offers a formative assessment of current LLMs’ proficiency in emulating human-like, multi-turn conversations using an LLM-centric approach. The evaluation encompasses three key elements in the evaluation pipeline: utterance generation, evaluation protocol, and judgement, and we delve deeply into each aspect. GPT-4, both as an utterance generator and as a judge, exhibits exceptional performance. As a generator, GPT-4 crafts dialogues indistinguishable from human interactions in terms of style and flow. When judging, it shows a heightened alignment with human evaluative standards and consistency. Conversely, other LLMs face challenges in producing quality multi-turn dialogues, hindered by inadequate instruction-following abilities, a propensity for prolix utterances, and overall limited capabilities. Notably, generating extensive dialogues (eg, spanning tens of turns) remains a formidable task for most LLMs, particularly in Chinese contexts. We hope that our work can serve as a valuable resource for evaluating the multi-turn chatting capabilities of LLMs. Related resources are available at https://github. com/open-compass/BotChat.

Paper Code
103 citations

InternLM-XComposer2.5-OmniLive: A Comprehensive Multimodal System for Long-term Streaming Video and Audio Interactions

Long-Context ModelingVideo UnderstandingFoundation Models

Pan Zhang, Xiaoyi Dong, Yuhang Cao, Yuhang Zang, Rui Qian, Xilin Wei, Lin Chen, Yifei Li, Junbo Niu, Shuangrui Ding, Qipeng Guo, Haodong Duan, Xin Chen, Han Lv, Zheng Nie, Min Zhang, Bin Wang, Wenwei Zhang, Xinyue Zhang, Jiaye Ge, Wei Li, Jingwen Li, Zhongying Tu, Conghui He, Xingcheng Zhang, Kai Chen, Yu Qiao, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2412.09596, 2024

Full metadata
Publication date
2024/12/12
Journal
arXiv preprint arXiv:2412.09596

Description

Creating AI systems that can interact with environments over long periods, similar to human cognition, has been a longstanding research goal. Recent advancements in multimodal large language models (MLLMs) have made significant strides in open-world understanding. However, the challenge of continuous and simultaneous streaming perception, memory, and reasoning remains largely unexplored. Current MLLMs are constrained by their sequence-to-sequence architecture, which limits their ability to process inputs and generate responses simultaneously, akin to being unable to think while perceiving. Furthermore, relying on long contexts to store historical data is impractical for long-term interactions, as retaining all information becomes costly and inefficient. Therefore, rather than relying on a single foundation model to perform all functions, this project draws inspiration from the concept of the Specialized Generalist AI and introduces disentangled streaming perception, reasoning, and memory mechanisms, enabling real-time interaction with streaming video and audio input. The proposed framework InternLM-XComposer2.5-OmniLive (IXC2.5-OL) consists of three key modules: (1) Streaming Perception Module: Processes multimodal information in real-time, storing key details in memory and triggering reasoning in response to user queries. (2) Multi-modal Long Memory Module: Integrates short-term and long-term memory, compressing short-term memories into long-term ones for efficient retrieval and improved accuracy. (3) Reasoning Module: Responds to queries and executes reasoning tasks, coordinating with the perception and …

Paper
62 citations

Prism: A Framework for Decoupling and Assessing the Capabilities of VLMs

Visual ReasoningMultimodal EvaluationModel Efficiency

Yuxuan Qiao, Haodong Duan, Xinyu Fang, Junming Yang, Lin Chen, Songyang Zhang, Jiaqi Wang, Dahua Lin, Kai Chen

Conference on Neural Information Processing Systems (NeurIPS), 2024

Full metadata
Publication date
2024/6/20
Journal
NeurIPS, 2024

Description

Vision Language Models (VLMs) demonstrate remarkable proficiency in addressing a wide array of visual questions, which requires strong perception and reasoning faculties. Assessing these two competencies independently is crucial for model refinement, despite the inherent difficulty due to the intertwined nature of seeing and reasoning in existing VLMs. To tackle this issue, we present Prism, an innovative framework designed to disentangle the perception and reasoning processes involved in visual question solving. Prism comprises two distinct stages: a perception stage that utilizes a VLM to extract and articulate visual information in textual form, and a reasoning stage that formulates responses based on the extracted visual information using a Large Language Model (LLM). This modular design enables the systematic comparison and assessment of both proprietary and open-source VLM for their perception and reasoning strengths. Our analytical framework provides several valuable insights, underscoring Prism's potential as a cost-effective solution for vision-language tasks. By combining a streamlined VLM focused on perception with a powerful LLM tailored for reasoning, Prism achieves superior results in general vision-language tasks while substantially cutting down on training and operational expenses. Quantitative evaluations show that Prism, when configured with a vanilla 2B LLaVA and freely accessible GPT-3.5, delivers performance on par with VLMs larger on the rigorous multimodal benchmark MMStar.

Paper Code
53 citations

Compassjudger-1: All-in-one judge model helps model evaluation and evolution

Evaluation InfrastructureLLM EvaluationBenchmark Design

Maosong Cao, Alexander Lam, Haodong Duan, Hongwei Liu, Songyang Zhang, Kai Chen

arXiv preprint arXiv:2410.16256, 2024

Full metadata
Publication date
2024/10/21
Journal
arXiv preprint arXiv:2410.16256

Description

Efficient and accurate evaluation is crucial for the continuous improvement of large language models (LLMs). Among various assessment methods, subjective evaluation has garnered significant attention due to its superior alignment with real-world usage scenarios and human preferences. However, human-based evaluations are costly and lack reproducibility, making precise automated evaluators (judgers) vital in this process. In this report, we introduce \textbf{CompassJudger-1}, the first open-source \textbf{all-in-one} judge LLM. CompassJudger-1 is a general-purpose LLM that demonstrates remarkable versatility. It is capable of: 1. Performing unitary scoring and two-model comparisons as a reward model; 2. Conducting evaluations according to specified formats; 3. Generating critiques; 4. Executing diverse tasks like a general LLM. To assess the evaluation capabilities of different judge models under a unified setting, we have also established \textbf{JudgerBench}, a new benchmark that encompasses various subjective evaluation tasks and covers a wide range of topics. CompassJudger-1 offers a comprehensive solution for various evaluation tasks while maintaining the flexibility to adapt to diverse requirements. Both CompassJudger and JudgerBench are released and available to the research community athttps://github.com/open-compass/CompassJudger. We believe that by open-sourcing these tools, we can foster collaboration and accelerate progress in LLM evaluation methodologies.

Paper
52 citations

Ada-LEval: Evaluating Long-Context LLMs with Length-Adaptable Benchmarks

Long-Context ModelingLLM EvaluationBenchmark Design

Chonghua Wang, Haodong Duan, Songyang Zhang, Dahua Lin, Kai Chen

Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), 2024

Full metadata
Publication date
2024/4/9
Journal
NAACL, 2024

Description

Recently, the large language model (LLM) community has shown increasing interest in enhancing LLMs’ capability to handle extremely long documents. As various long-text techniques and model architectures emerge, the precise and detailed evaluation of models’ long-text capabilities has become increasingly important. Existing long-text evaluation benchmarks, such as L-Eval and LongBench, construct long-text test sets based on open-source datasets, focusing mainly on QA and summarization tasks. These datasets include test samples of varying lengths (from 2k to 32k+) entangled together, making it challenging to assess model capabilities across different length ranges. Moreover, they do not cover the ultralong settings (100k+ tokens) that the latest LLMs claim to achieve. In this paper, we introduce Ada-LEval, a length-adaptable benchmark for evaluating the long-context understanding of LLMs. Ada-LEval includes two challenging subsets, TSort and BestAnswer, which enable a more reliable evaluation of LLMs’ long context capabilities. These benchmarks support intricate manipulation of the length of test cases, and can easily produce text samples up to 128k tokens. We evaluate 4 state-of-the-art closed-source API models and 6 open-source models with Ada-LEval. The evaluation results demonstrate the limitations of current LLMs, especially in ultra-long-context settings. Our code is available at https://github. com/open-compass/Ada-LEval.

Paper Code
46 citations

2023

Internlm-xcomposer: A vision-language large model for advanced text-image comprehension and composition

Image UnderstandingFoundation Models

Pan Zhang, Xiaoyi Dong, Bin Wang, Yuhang Cao, Chao Xu, Linke Ouyang, Zhiyuan Zhao, Haodong Duan, Songyang Zhang, Shuangrui Ding, Wenwei Zhang, Hang Yan, Xinyue Zhang, Wei Li, Jingwen Li, Kai Chen, Conghui He, Xingcheng Zhang, Yu Qiao, Dahua Lin, Jiaqi Wang

arXiv preprint arXiv:2309.15112, 2023

Full metadata
Publication date
2023/9/26
Journal
arXiv preprint arXiv:2309.15112

Description

We propose InternLM-XComposer, a vision-language large model that enables advanced image-text comprehension and composition. The innovative nature of our model is highlighted by three appealing properties: 1) Interleaved Text-Image Composition: InternLM-XComposer can effortlessly generate coherent and contextual articles that seamlessly integrate images, providing a more engaging and immersive reading experience. Simply provide a writing instruction, and our system will generate the corresponding manuscript. It can intelligently identify the areas in the text where images would enhance the content and automatically insert the most appropriate visual candidates. 2) Comprehension with Rich Multilingual Knowledge: The text-image comprehension is empowered by training on an extensive multi-modal multilingual database with carefully crafted strategies, resulting in a deep understanding of visual content. 3) State-of-the-art Performance: Our model consistently achieves state-of-the-art results across various mainstream benchmarks for vision-language foundational models, including MME Benchmark, MMBench, MMBench-CN, Seed-Bench, CCBench (Chinese Cultural Benchmark), QBench and Tiny LVLM. Owing to the absence of established metrics for quantitatively assessing text-image composition, we have devised a robust evaluation procedure that comprises both human and GPT4-Vision (GPT4-V) to ensure reliability. Notably, our InternLM-XComposer achieves competitive text-image composition scores compared to public solutions, including GPT4-V and GPT3.5. Collectively, InternLM-XComposer seamlessly blends …

Paper
399 citations

Internlm: A multilingual language model with progressively enhanced capabilities

Foundation ModelsModel Adaptation

InternLM Team

InternLM Project, https://github.com/InternLM/InternLM, 2023

Full metadata
Publication date
2023
Source
InternLM Project, https://github.com/InternLM/InternLM

Description

We present InternLM, a multilingual foundational language model with 104B parameters. InternLM is pre-trained on a large corpora with 1.6 T tokens with a multi-phase progressive process, and then fine-tuned to align with human preferences. We also developed a training system called Uniscale-LLM for efficient large language model training. The evaluation on a number of benchmarks shows that InternLM achieves state-of-the-art performance in multiple aspects, including knowledge understanding, reading comprehension, mathematics, and coding. With such well-rounded capabilities, InternLM achieves outstanding performances on comprehensive exams, including MMLU, AGIEval, C-Eval and GAOKAO-Bench, without resorting to external tools. On these benchmarks, InternLM not only significantly outperforms open-source models, but also obtains superior performance compared to ChatGPT. Also, InternLM demonstrates excellent capability of understanding Chinese language and Chinese culture, which makes it a suitable foundation model to support Chinese-oriented language applications. This manuscript gives a detailed study of our results, with benchmarks and examples across a diverse set of knowledge domains and tasks 1.

Paper
328 citations

JourneyDB: A Benchmark for Generative Image Understanding

Image UnderstandingMultimodal GenerationMultimodal Evaluation

Keqiang Sun, Junting Pan, Yuying Ge, Hao Li, Haodong Duan, Xiaoshi Wu, Renrui Zhang, Aojun Zhou, Zipeng Qin, Yi Wang, Jifeng Dai, Yu Qiao, Limin Wang, Hongsheng Li

Conference on Neural Information Processing Systems (NeurIPS), Datasets & Benchmarks Track, 2023

Full metadata
Publication date
2024/2/13
Journal
NeurIPS D&B, 2023

Description

While recent advancements in vision-language models have had a transformative impact on multi-modal comprehension, the extent to which these models possess the ability to comprehend generated images remains uncertain. Synthetic images, in comparison to real data, encompass a higher level of diversity in terms of both content and style, thereby presenting significant challenges for the models to fully grasp. In light of this challenge, we introduce a comprehensive dataset, referred to as JourneyDB, that caters to the domain of generative images within the context of multi-modal visual understanding. Our meticulously curated dataset comprises 4 million distinct and high-quality generated images, each paired with the corresponding text prompts that were employed in their creation. Furthermore, we additionally introduce an external subset with results of another 22 text-to-image generative models, which makes JourneyDB a comprehensive benchmark for evaluating the comprehension of generated images. On our dataset, we have devised four benchmarks to assess the performance of generated image comprehension in relation to both content and style interpretation. These benchmarks encompass prompt inversion, style retrieval, image captioning, and visual question answering. Lastly, we evaluate the performance of state-of-the-art multi-modal models when applied to the JourneyDB dataset, providing a comprehensive analysis of their strengths and limitations in comprehending generated content. We anticipate that the proposed dataset and benchmarks will facilitate further research in the field of generative content understanding. The …

Paper Code
257 citations

Self-supervised Action Representation Learning from Partial Spatio-Temporal Skeleton Sequences

Skeleton RecognitionAction RecognitionSelf-Supervised Learning

Yujie Zhou, Haodong Duan, Anyi Rao, Bing Su, Jiaqi Wang

AAAI Conference on Artificial Intelligence (AAAI), 2023

Full metadata
Publication date
2023/2/17
Journal
AAAI, 2023

Description

Self-supervised learning has demonstrated remarkable capability in representation learning for skeleton-based action recognition. Existing methods mainly focus on applying global data augmentation to generate different views of the skeleton sequence for contrastive learning. However, due to the rich action clues in the skeleton sequences, existing methods may only take a global perspective to learn to discriminate different skeletons without thoroughly leveraging the local relationship between different skeleton joints and video frames, which is essential for real-world applications. In this work, we propose a Partial Spatio-Temporal Learning (PSTL) framework to exploit the local relationship from a partial skeleton sequences built by a unique spatio-temporal masking strategy. Specifically, we construct a negative-sample-free triplet steam structure that is composed of an anchor stream without any masking, a spatial masking stream with Central Spatial Masking (CSM), and a temporal masking stream with Motion Attention Temporal Masking (MATM). The feature cross-correlation matrix is measured between the anchor stream and the other two masking streams, respectively.(1) Central Spatial Masking discards selected joints from the feature calculation process, where the joints with a higher degree of centrality have a higher possibility of being selected.(2) Motion Attention Temporal Masking leverages the motion of action and remove frames that move faster with a higher possibility. Our method achieves state-of-the-art performance on NTURGB+ D 60, NTURGB+ D 120 and PKU-MMD under various downstream tasks. Furthermore, to simulate …

Paper Code
93 citations

SkeleTR: Towards Skeleton-based Action Recognition in the Wild

Skeleton RecognitionAction RecognitionHuman Pose

Haodong Duan, Mingze Xu, Bing Shuai, Davide Modolo, Zhuowen Tu, Joseph Tighe, Alessandro Bergamo

IEEE/CVF International Conference on Computer Vision (ICCV), 2023

Full metadata
Publication date
2023
Conference
ICCV, 2023

Description

We present SkeleTR, a new framework for skeleton-based action recognition. In contrast to prior work, which focuses mainly on controlled environments, we target more general scenarios that typically involve a variable number of people and various forms of interaction between people. SkeleTR works with a two-stage paradigm. It first models the intra-person skeleton dynamics for each skeleton sequence with graph convolutions, and then uses stacked Transformer encoders to capture person interactions that are important for action recognition in general scenarios. To mitigate the negative impact of inaccurate skeleton associations, SkeleTR takes relative short skeleton sequences as input and increases the number of sequences. As a unified solution, SkeleTR can be directly applied to multiple skeleton-based action tasks, including video-level action classification, instance-level action detection, and group-level …

Paper
83 citations

2022

Revisiting Skeleton-based Action Recognition

Skeleton RecognitionAction RecognitionHuman Pose

Haodong Duan, Yue Zhao, Kai Chen, Dahua Lin, Bo Dai

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

Full metadata
Publication date
2022
Conference
CVPR Oral, 2022

Description

Human skeleton, as a compact representation of human action, has received increasing attention in recent years. Many skeleton-based action recognition methods adopt GCNs to extract features on top of human skeletons. Despite the positive results shown in these attempts, GCN-based methods are subject to limitations in robustness, interoperability, and scalability. In this work, we propose PoseConv3D, a new approach to skeleton-based action recognition. PoseConv3D relies on a 3D heatmap volume instead of a graph sequence as the base representation of human skeletons. Compared to GCN-based methods, PoseConv3D is more effective in learning spatiotemporal features, more robust against pose estimation noises, and generalizes better in cross-dataset settings. Also, PoseConv3D can handle multiple-person scenarios without additional computation costs. The hierarchical features can be easily …

Paper Code
1349 citations

PYSKL: Towards Good Practices for Skeleton Action Recognition

Evaluation InfrastructureSkeleton RecognitionAction Recognition

Haodong Duan, Jiaqi Wang, Kai Chen, Dahua Lin

ACM International Conference on Multimedia (MM), 2022

Full metadata
Publication date
2022/5/19
Journal
MMOpen, 2022

Description

We present PYSKL: an open-source toolbox for skeleton-based action recognition based on PyTorch. The toolbox supports a wide variety of skeleton action recognition algorithms, including approaches based on GCN and CNN. In contrast to existing open-source skeleton action recognition projects that include only one or two algorithms, PYSKL implements six different algorithms under a unified framework with both the latest and original good practices to ease the comparison of efficacy and efficiency. We also provide an original GCN-based skeleton action recognition model named ST-GCN++, which achieves competitive recognition performance without any complicated attention schemes, serving as a strong baseline. Meanwhile, PYSKL supports the training and testing of nine skeleton-based action recognition benchmarks and achieves state-of-the-art recognition performance on eight of them. To facilitate …

Paper Code
325 citations

Dg-stgcn: Dynamic spatial-temporal modeling for skeleton-based action recognition

Skeleton RecognitionAction RecognitionHuman Pose

Haodong Duan, Jiaqi Wang, Kai Chen, Dahua Lin

arXiv preprint arXiv:2210.05895, 2022

Full metadata
Publication date
2022/10/12
Journal
arXiv preprint arXiv:2210.05895

Description

Graph convolution networks (GCN) have been widely used in skeleton-based action recognition. We note that existing GCN-based approaches primarily rely on prescribed graphical structures (ie., a manually defined topology of skeleton joints), which limits their flexibility to capture complicated correlations between joints. To move beyond this limitation, we propose a new framework for skeleton-based action recognition, namely Dynamic Group Spatio-Temporal GCN (DG-STGCN). It consists of two modules, DG-GCN and DG-TCN, respectively, for spatial and temporal modeling. In particular, DG-GCN uses learned affinity matrices to capture dynamic graphical structures instead of relying on a prescribed one, while DG-TCN performs group-wise temporal convolutions with varying receptive fields and incorporates a dynamic joint-skeleton fusion module for adaptive multi-level temporal modeling. On a wide range of benchmarks, including NTURGB+D, Kinetics-Skeleton, BABEL, and Toyota SmartHome, DG-STGCN consistently outperforms state-of-the-art methods, often by a notable margin.

Paper Code
135 citations

OCSampler: Compressing Videos to One Clip with Single-step Sampling

Video UnderstandingModel Efficiency

Jintao Lin, Haodong Duan, Kai Chen, Dahua Lin, Limin Wang

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

Full metadata
Publication date
2022/1/12
Journal
CVPR, 2022

Description

Videos incorporate rich semantics as well as redundant information. Seeking a compact yet effective video representation, e.g., sample informative frames from the entire video, is critical to efficient video recognition. There have been works that formulate frame sampling as a sequential decision task by selecting frames one by one according to their importance. In this paper, we present a more efficient framework named OCSampler, which explores such a representation with one short clip. OCSampler designs a new paradigm of learning instance-specific video condensation policies to select frames only in a single step. Rather than picking up frames sequentially like previous methods, we simply process a whole sequence at once. Accordingly, these policies are derived from a light-weighted skim network together with a simple yet effective policy network. Moreover, we extend the proposed method with a frame …

Paper
43 citations

TransRank: Self-supervised Video Representation Learning via Ranking-based Transformation Recognition

Self-Supervised LearningVideo UnderstandingAction Recognition

Haodong Duan, Nanxuan Zhao, Kai Chen, Dahua Lin

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

Full metadata
Publication date
2022/5/4
Journal
CVPR Oral, 2022

Description

Recognizing transformation types applied to a video clip (RecogTrans) is a long-established paradigm for selfsupervised video representation learning, which achieves much inferior performance compared to instance discrimination approaches (InstDisc) in recent works. However, based on a thorough comparison of representative Recog-Trans and InstDisc methods, we observe the great potential of RecogTrans on both semantic-related and temporalrelated downstream tasks. Based on hard-label classification, existing RecogTrans approaches suffer from noisy supervision signals in pre-training. To mitigate this problem, we developed TransRank, a unified framework for recognizing Transformations in a Ranking formulation. TransRank provides accurate supervision signals by recognizing transformations relatively, consistently outperforming the classification-based formulation. Meanwhile, the unified framework …

Paper Code
42 citations

Mitigating Representation Bias in Action Recognition: Algorithms and Benchmarks

Action RecognitionVideo UnderstandingBenchmark Design

Haodong Duan, Yue Zhao, Kai Chen, Yuanjun Xiong, Dahua Lin

ECCV Workshop, 2022

Full metadata
Publication date
2022/9/20
Journal
ECCV Workshop, 2022

Description

Deep learning models have achieved excellent recognition results on large-scale video benchmarks. However, they perform poorly when applied to videos with rare scenes or objects, primarily due to the bias of existing video datasets. We tackle this problem from two different angles: algorithm and dataset. From the perspective of algorithms, we propose Spatial-aware Multi-Aspect Debiasing (SMAD), which incorporates both explicit debiasing with multi-aspect adversarial training and implicit debiasing with the spatial actionness reweighting module, to learn a more generic representation invariant to non-action aspects. To neutralize the intrinsic dataset bias, we propose OmniDebias to leverage web data for joint training selectively, which can achieve higher performance with far fewer web data. To verify the effectiveness, we establish evaluation protocols and perform extensive experiments on both re-distributed …

Paper
14 citations

2020

MMAction2: OpenMMLab's Video Understanding Toolbox and Benchmark

Action RecognitionVideo UnderstandingEvaluation Infrastructure

MMAction2 Contributors

OpenMMLab Project, https://github.com/open-mmlab/mmaction2, 2020

Full metadata
Publication date
2020
Source
OpenMMLab Project, https://github.com/open-mmlab/mmaction2
389 citations

Omni-sourced Webly-supervised Learning for Video Recognition

Data-Centric LearningVideo UnderstandingAction Recognition

Haodong Duan, Yue Zhao, Yuanjun Xiong, Wentao Liu, Dahua Lin

European Conference on Computer Vision (ECCV), 2020

Full metadata
Publication date
2020/8/23
Conference
ECCV, 2020
Publisher
Springer, Cham

Description

We introduce OmniSource, a novel framework for leveraging web data to train video recognition models. OmniSource overcomes the barriers between data formats, such as images, short videos, and long untrimmed videos for webly-supervised learning. First, data samples with multiple formats, curated by task-specific data collection and automatically filtered by a teacher model, are transformed into a unified form. Then a joint-training strategy is proposed to deal with the domain gaps between multiple data sources and formats in webly-supervised learning. Several good practices, including data balancing, resampling, and cross-dataset mixup are adopted in joint training. Experiments show that by utilizing data from multiple sources and formats, OmniSource is more data-efficient in training. With only 3.5M images and 800K min videos crawled from the internet without human labeling (less than of prior works …

Paper Code
139 citations

2019

TRB: A Novel Triplet Representation for Understanding 2D Human Body

Human PoseImage Understanding

Haodong Duan, Kwan-Yee Lin, Sheng Jin, Wentao Liu, Chen Qian, Wanli Ouyang

IEEE/CVF International Conference on Computer Vision (ICCV), 2019

Full metadata
Publication date
2019
Conference
ICCV Oral, 2019

Description

Human pose and shape are two important components of 2D human body. However, how to efficiently represent both of them in images is still an open question. In this paper, we propose the Triplet Representation for Body (TRB) --- a compact 2D human body representation, with skeleton keypoints capturing human pose information and contour keypoints containing human shape information. TRB not only preserves the flexibility of skeleton keypoint representation, but also contains rich pose and human shape information. Therefore, it promises broader application areas, such as human shape editing and conditional image generation. We further introduce the challenging problem of TRB estimation, where joint learning of human pose and shape is required. We construct several large-scale TRB estimation datasets, based on the popular 2D pose datasets LSP, MPII and COCO. To effectively solve TRB estimation …

Paper
21 citations

2017

SRPGAN: perceptual generative adversarial network for single image super resolution

Multimodal GenerationImage Understanding

Bingzhe Wu, Haodong Duan, Zhichao Liu, Guangyu Sun

arXiv preprint arXiv:1712.05927, 2017

Full metadata
Publication date
2017/12/16
Journal
arXiv preprint arXiv:1712.05927

Description

Single image super resolution (SISR) is to reconstruct a high resolution image from a single low resolution image. The SISR task has been a very attractive research topic over the last two decades. In recent years, convolutional neural network (CNN) based models have achieved great performance on SISR task. Despite the breakthroughs achieved by using CNN models, there are still some problems remaining unsolved, such as how to recover high frequency details of high resolution images. Previous CNN based models always use a pixel wise loss, such as l2 loss. Although the high resolution images constructed by these models have high peak signal-to-noise ratio (PSNR), they often tend to be blurry and lack high-frequency details, especially at a large scaling factor. In this paper, we build a super resolution perceptual generative adversarial network (SRPGAN) framework for SISR tasks. In the framework, we propose a robust perceptual loss based on the discriminator of the built SRPGAN model. We use the Charbonnier loss function to build the content loss and combine it with the proposed perceptual loss and the adversarial loss. Compared with other state-of-the-art methods, our method has demonstrated great ability to construct images with sharp edges and rich details. We also evaluate our method on different benchmarks and compare it with previous CNN based methods. The results show that our method can achieve much higher structural similarity index (SSIM) scores on most of the benchmarks than the previous state-of-art methods.

Paper
91 citations