16 papers
Noisy Test-Time Reinforcement Learning for Code LLMs
Xikai Yang, Hieu Trung Nguyen, Dunyuan Xu +4
Large language models (LLMs) have demonstrated remarkable performance across various code-related tasks. However, unlike carefully curated datasets that are typically high-quality…
Failure-Informed Image Self-Augmentation for Multimodal Large Language Model Self-Improvement
Chunyang Jiang, Pingping Zhang, Yuzhi Zhao +9
Multimodal large language models (MLLMs) have achieved remarkable performance across vision-language tasks, but their progress depends heavily on large-scale, high-quality multimod…
SoftSkill: Behavioral Compression for Contextual Adaptation
Xijia Tao, Yihua Teng, Xinyu Fu +6
Natural-language skills let agents reuse task knowledge, yet deploying a long Markdown document makes the model interpret that knowledge anew on every call. We ask whether the beha…
Unified Context Evolution for LLM Agents
Zixuan Zhu, Yitong Hu, Yong Dai +4
LLM-based agents can solve multi-step interactive tasks by combining reasoning with environment feedback, yet each episode starts from the same fixed context and any useful strateg…
Self-Induced Outcome Potential: Turn-Level Credit Assignment for Agents without Verifiers
Senkang Hu, Yong Dai, Xudong Han +4
Long-horizon LLM agents depend on intermediate information-gathering turns, yet training feedback is usually observed only at the final answer, because process-level rewards requir…
Skill-Conditioned Gated Self-Distillation for LLM Reasoning
Jiazhen Huang, Xiao Chen, Xiao Luo +3
On-policy self-distillation (SD) improves LLM reasoning by using teacher-side privileged information (PI) to turn sparse verifier outcomes into dense token-level supervision. Exist…
Adaptive Rollout Allocation for Online Reinforcement Learning with Verifiable Rewards
Hieu Trung Nguyen, Bao Nguyen, Wenao Ma +3
Sampling efficiency is a key bottleneck in reinforcement learning with verifiable rewards. Existing group-based policy optimization methods, such as GRPO, allocate a fixed number o…
Optimizing Agentic Reasoning with Retrieval via Synthetic Semantic Information Gain Reward
Senkang Hu, Yong Dai, Yuzhi Zhao +5
Agentic reasoning enables large reasoning models (LRMs) to dynamically acquire external knowledge, but yet optimizing the retrieval process remains challenging due to the lack of d…
VP-Bench: A Comprehensive Benchmark for Visual Prompting in Multimodal Large Language Models
Mingjie Xu, Jinpeng Chen, Yuzhi Zhao +12
Multimodal large language models (MLLMs) have enabled a wide range of advanced vision-language applications, including fine-grained object recognition and contextual understanding.…
From Exploration to Exploitation: A Two-Stage Entropy RLVR Approach for Noise-Tolerant MLLM Training
Donglai Xu, Hongzheng Yang, Yuzhi Zhao +10
Reinforcement Learning with Verifiable Rewards (RLVR) for Multimodal Large Language Models (MLLMs) is highly dependent on high-quality labeled data, which is often scarce and prone…
AgenticMath: Enhancing LLM Reasoning via Agentic-based Math Data Generation
Xianyang Liu, Yilin Liu, Shuai Wang +5
The creation of high-quality datasets to improve Large Language Model (LLM) reasoning remains a significant challenge, as current methods often suffer from generating low-quality/i…
KG-RAG: Enhancing GUI Agent Decision-Making via Knowledge Graph-Driven Retrieval-Augmented Generation
Ziyi Guan, Jason Chun Lok Li, Zhijian Hou +11
Despite recent progress, Graphic User Interface (GUI) agents powered by Large Language Models (LLMs) struggle with complex mobile tasks due to limited app-specific knowledge. While…
Better Reasoning with Less Data: Enhancing VLMs Through Unified Modality Scoring
Mingjie Xu, Andrew Estornell, Hongzheng Yang +4
The application of visual instruction tuning and other post-training techniques has significantly enhanced the capabilities of Large Language Models (LLMs) in visual understanding,…
SEFE: Superficial and Essential Forgetting Eliminator for Multimodal Continual Instruction Tuning
Jinpeng Chen, Runmin Cong, Yuzhi Zhao +4
Multimodal Continual Instruction Tuning (MCIT) aims to enable Multimodal Large Language Models (MLLMs) to incrementally learn new tasks without catastrophic forgetting. In this pap…
ICM-Assistant: Instruction-tuning Multimodal Large Language Models for Rule-based Explainable Image Content Moderation
Mengyang Wu, Yuzhi Zhao, Jialun Cao +7
Controversial contents largely inundate the Internet, infringing various cultural norms and child protection standards. Traditional Image Content Moderation (ICM) models fall short…
LLaVA-SpaceSGG: Visual Instruct Tuning for Open-vocabulary Scene Graph Generation with Enhanced Spatial Relations
Mingjie Xu, Mengyang Wu, Yuzhi Zhao +2
Scene Graph Generation (SGG) converts visual scenes into structured graph representations, providing deeper scene understanding for complex vision tasks. However, existing SGG mode…