9 papers
Building AI Research Teams for Autonomous Discovery of Linear Attention Architectures
Yujian Liu, Li An, Zelin He +5
Autonomous agents can increasingly propose ideas, modify code, and run experiments, creating an opportunity to accelerate scientific discovery using teams of collaborative AI resea…
Budgeted KV-Cache Repair for Evolving Contexts
Han Yu, Wenhui Zhu, Jincheng Cao +15
Language-model agents revise their working context as information becomes obsolete, but deleting text does not erase its influence on cached computation. Later key-value (KV) state…
Composing What Each Teacher Learned: Delta Distillation for Multi-Teacher On-Policy Learning
Hejian Sang, Yuanda Xu, Zhengze Zhou +4
Multi-teacher on-policy distillation (MOPD) is used in two settings. In common-domain composition, several teachers score each student rollout from one prompt domain and their sign…
SHARPO: Segment-Level Credit Assignment for Agentic Reinforcement Learning
Xinchen Du, Zhengze Zhou, Wenhui Zhu +4
Agentic reinforcement learning (RL) trains a large language model (LLM) to act over long, multi-step interactions. However, a single localized error can cause task failure, while t…
Smaller Models, Better Rejects: Preference Distillation Scaling
Rui Cai, Wenhui Zhu, Xiwen Chen +15
Preference distillation typically treats a teacher response as preferred and the student's own response as rejected. This assumes that self-generated failures are the most informat…
Practical Online KV Cache Compaction for LLM Agents: An Empirical Study
Yujian Liu, Jiabao Ji, Li An +4
LLM agents accumulate long trajectories of reasoning steps, tool calls, and environment feedback, making the KV cache a major inference bottleneck. KV cache compaction can reduce t…
Speculate While You Reason: Teaching Agents to Predict Their Next Tool Call via Joint Agent-Speculator RL
Jiabao Ji, Yujian Liu, Li An +4
Large language model agents often spend substantial wall-clock time waiting for tool call results. Tool-call speculation can hide this latency by predicting and pre-executing an ag…
Agentic Reinforcement Learning for Real-World Code Repair
Siyu Zhu, Anastasiya Karpovich, Albert Chen +6
We tackle the challenge of training reliable code-fixing agents in real repositories, where complex builds and shifting dependencies make evaluation unstable. We developed a verifi…
Planner-R1: Reward Shaping Enables Efficient Agentic RL with Smaller LLMs
Siyu Zhu, Yanbin Jiang, Hejian Sang +6
We investigated Agentic RL with large language models on the \textsc{TravelPlanner} benchmark. Our approach, \textsc{Planner-R1}, achieved a \textbf{56.9\%} final-pass rate with on…