5 papers
When to Forget: Risk-Controlled Context Adaptation for Long-Horizon Agents
Chenle Chen, Yangbo Wei, Zhijie Zheng +6
Long-horizon agents learn from failure, but the value of a failure record can change as execution progresses. We study when retaining, compressing, or removing the same history blo…
Serve Now or Improve Later? Scheduling Self-Evolution in Online Agent Systems
Yangbo Wei, Junhong Qian, Zhen Huang +6
Online agents can improve future service by constructing reusable tools, guidance, or model states, but this work competes with current requests for the same GPUs. Exploiting idle…
Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents
Chao Yao, Yangbo Wei, Zhen Huang +5
Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache. Yet t…
Competence, Not Accuracy: A Diagnostic for Reference-Free Judge Gates in Skill Optimization
Chenle Chen, Yangbo Wei, Chao Yao +4
Text-space skill optimization adapts a frozen agent by evolving a natural-language skill document, accepting each candidate through a validation gate. Existing gates rely on verifi…
SkillSmith: Co-Evolving Skills and Tools for Self-Improving Agent Systems
Yangbo Wei, Zhen Huang, Shaoqiang Lu +4
LLM agents increasingly store capabilities in external skill and tool libraries. Libraries can grow, but execution must fit a finite context window and improvement consumes a finit…