10 papers
Trajectory-Level Speculative Decoding for Diffusion Language Models
Tianxiang Pan, Baitao Gong, Mo Guang +5
Diffusion-based language models (dLLMs) enable parallel token generation through iterative denoising, but existing decoding strategies collapse to single-token generation under low…
MS-Resampler: Multi-Scope Visual Resampling for Efficient Multimodal LLMs
Zhongyang Li, Yaqian Li, Faming Fang +6
Multimodal large language models (MLLMs) typically employ resampling-based projectors to transform dense visual features into a compact token sequence for language modeling. Most e…
SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs
Zi-Hao Bo, Yaqian Li, Anzhou Hou +6
Mixture-of-Experts (MoE) has become a prevalent backbone for large vision-language models (VLMs), yet how modality-specific signals should guide expert routing remains under-explor…
LearnPruner: Rethinking Attention-based Token Pruning in Vision Language Models
Rinyoichi Takezoe, Yaqian Li, Zihao Bo +3
Vision-Language Models (VLMs) have recently demonstrated remarkable capabilities in visual understanding and reasoning, but they also impose significant computational burdens due t…
QMoP: Query Guided Mixture-of-Projector for Efficient Visual Token Compression
Zhongyang Li, Yaqian Li, Faming Fang +6
Multimodal large language models suffer from severe computational and memory bottlenecks, as the number of visual tokens far exceeds that of textual tokens. While recent methods em…
MARSHAL: Incentivizing Multi-Agent Reasoning via Self-Play with Strategic LLMs
Huining Yuan, Zelai Xu, Zheyue Tan +10
Developing Large Language Models (LLMs) to cooperate and compete effectively within multi-agent systems (MASs) is a critical step towards more advanced intelligence. While reinforc…
SAC Flow: Sample-Efficient Reinforcement Learning of Flow-Based Policies via Velocity-Reparameterized Sequential Modeling
Yixian Zhang, Shu'ang Yu, Tonghe Zhang +6
Training expressive flow-based policies with off-policy reinforcement learning is notoriously unstable due to gradient pathologies in the multi-step action sampling process. We tra…
RE-PO: Robust Enhanced Policy Optimization as a General Framework for LLM Alignment
Xiaoyang Cao, Zelai Xu, Mo Guang +4
Standard human preference-based alignment methods, such as Reinforcement Learning from Human Feedback (RLHF), are a cornerstone for aligning large language models (LLMs) with human…
SATQuest: A Verifier for Logical Reasoning Evaluation and Reinforcement Fine-Tuning of LLMs
Yanxiao Zhao, Yaqian Li, Zihao Bo +6
Large language models (LLMs) exhibit strong general reasoning, yet the community lacks controllable, scalable, and verifiable tools to analyze and improve these abilities. We prese…
VS-Bench: Evaluating VLMs for Strategic Abilities in Multi-Agent Environments
Zelai Xu, Zhexuan Xu, Xiangmin Yi +7
Recent advancements in Vision Language Models (VLMs) have expanded their capabilities to interactive agent tasks, yet existing benchmarks remain limited to single-agent or text-onl…