15 papers
Work Agents: Systems, Capabilities, and Optimization
Zishan Xu, Enjun Du, Mingqi Gao +13
As foundation models evolve from information generation to digital work execution, agents are increasingly capable of completing complex tasks across tools, applications, and workf…
Better Supervision Is Nearby: Neighborhood On-Policy Self-Distillation
Xincheng Wei, Yifan Ding, Yoshua Li +6
On-policy self-distillation (OPSD) trains mathematical reasoning models using a privileged teacher that sees a reference solution and supervises student-sampled prefixes. Standard…
EnvACE: Internalizing Environment Dynamics via World Rehearsal for Agentic Reinforcement Learning
Zishan Xu, Zhiyuan Yao, Yuxin Chen +9
Training large language model agents for long-horizon tool use typically relies on interactions with real or synthesized executable environments, whose construction and verificatio…
SAF-OPD: Stable Advantage Fusion for On-Policy Distillation
Yifan Ding, Xincheng Wei, Yoshua Y. Li +7
Reinforcement learning with verifiable rewards (RLVR) broadcasts a single response-level reward to every token, while on-policy distillation (OPD) scores each token against a stron…
SkillRise: Agentic Reinforcement Learning for Cross-Task Skill Evolution
Zhiyuan Yao, Yuxin Chen, Zhengxi Lu +13
Large language model agents often encounter related yet distinct tasks that share reusable solution patterns. Yet standard agentic reinforcement learning treats tasks as independen…
SIGMA: Skill-Incidence Graphs for Compositional Multi-Agent Design
Kun Zeng, Yu Huo, Siyu Zhang +5
Existing graph-based multi-agent system (MAS) designers mainly improve collaboration by optimizing communication topologies over predefined agents, roles, or groups. However, becau…
From Brewing to Resolution: Tracing the Internal Lifecycle of Code Reasoning in LLMs
Siyue Chen, Yifu Guo, Yuquan Lu +9
Standard accuracy metrics cannot explain why LLMs handle variable tracking but fail on semantically equivalent loops. We study an internal lifecycle of code reasoning in which mode…
Group of Skills: Group-Structured Skill Retrieval for Agent Skill Libraries
Kun Zeng, Yu Huo, Siyu Zhang +6
Skill-augmented agents increasingly rely on large reusable skill libraries, but retrieving relevant skills is not the same as presenting usable context. Existing methods typically…
Zero-Forgetting CISS via Dual-Phase Cognitive Cascades
Yuquan Lu, Yifu Guo, Zishan Xu +6
Continual semantic segmentation (CSS) is a cornerstone task in computer vision that enables a large number of downstream applications, but faces the catastrophic forgetting challen…
EDIS: Diagnosing LLM Reasoning via Entropy Dynamics
Chenghua Zhu, Siyan Wu, Xiangkang Zeng +6
Entropy-based confidence signals are increasingly leveraged to improve reasoning in large language models (LLMs), yet existing approaches treat confidence as a static quantity -- t…
Shape of Thought: Progressive Object Assembly via Visual Chain-of-Thought
Yu Huo, Siyu Zhang, Kun Zeng +7
Multimodal models for text-to-image generation have achieved strong visual fidelity, yet they remain brittle under compositional structural constraints, notably generative numeracy…
RepoShapley: Shapley-Enhanced Context Filtering for Repository-Level Code Completion
Yu Huo, Kun Zeng, Siyu Zhang +4
Repository-level code completion benefits from retrieval-augmented generation (RAG). However, controlling cross-file evidence is difficult because chunk utility is often interactio…
Octopus: Agentic Multimodal Reasoning with Six-Capability Orchestration
Yifu Guo, Zishan Xu, Zhiyuan Yao +6
Existing multimodal reasoning models and frameworks suffer from fundamental architectural limitations: most lack the human-like ability to autonomously explore diverse reasoning pa…
VideoSeg-R1:Reasoning Video Object Segmentation via Reinforcement Learning
Zishan Xu, Yifu Guo, Yuquan Lu +2
Traditional video reasoning segmentation methods rely on supervised fine-tuning, which limits generalization to out-of-distribution scenarios and lacks explicit reasoning. To addre…
Decoupling Continual Semantic Segmentation
Yifu Guo, Yuquan Lu, Wentao Zhang +5
Continual Semantic Segmentation (CSS) requires learning new classes without forgetting previously acquired knowledge, addressing the fundamental challenge of catastrophic forgettin…