paper-with-me

홈 › Papers

Towards Feedback-to-Plan Decisions for Self-Evolving LLM Agents in CUDA Kernel Generation

2026-05-26 · Yee Hin Chong, Jiaming Wu, Youhui Zhang, Peng Qu arxiv

Large language models (LLMs) have shown strong empirical gains as self-evolving agents for CUDA kernel generation, driven by feedback-conditioned planning across generations. However, how planning decisions attribute and combine heterogeneous feedback signals remains opaque. Standard end-to-end ablations fail to resolve this question, as iterative planning amplifies early perturbations and conflates feedback effects with trajectory-dependent drift. We introduce \texttt{CUDAnalyst}, a unified analysis layer for controlled, generation-level attribution of planning decisions to feedback components via trajectory freezing and selective feedback injection. \texttt{CUDAnalyst} enables stable generation-level evaluation and principled coalitional-style attribution of feedback effects and interactions. Our results show that explicit planning is beneficial only when feedback is aligned, that effective planning emerges from structured multi-feedback interactions, and that high-level plans from stronger reasoning models can partially transfer to weaker ones. These trends hold across reference backbones, representative workloads, and reference induction regimes, indicating that the identified feedback-to-plan structure is robust within the controlled axes studied.

📄 PDF Abstract BibTeX arXiv:2605.26720

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Toward Vibe Medicine: A Self-Evolving Multi-Agent Framework for Clinical Decision Support

2026-04-01 · Qianxue Zhang, Yiming Ren, Shihuan Qin, Xiao Zhang 외 arxiv

In recent years, the advances of large language models and autonomous agents have revolutionized the healthcare field, facilitating diagnosis and improving treatment results. However, most existing AI systems rely on pre…

Self-Evolving Coding Agents

2026-08-04 · Hao Zhou, Haichuan Hu, Ye Shang, Quanjun Zhang hf

Large language models are increasingly embedded in software engineering workflows as coding agents that can inspect repositories, invoke tools, execute tests, debug failures, and generate patches. Yet most existing agent…

VASO: Formally Verifiable Self-Evolving Skills for Physical AI Agents

2026-06-03 · Yunhao Yang, Neel P. Bhatt, Kevin Wang, Samuel Tetteh 외 arxiv

Reusable robot skills are becoming the basic units through which embodied agents turn open-ended instructions into long-horizon physical behavior. We argue that, while foundation models have collapsed the cost of creatin…

MineEvolve: Self-Evolution with Accumulated Knowledge for Long-Horizon Embodied Minecraft Agents

2026-03-13 · Zhengwei Xie, Zhisheng Chen, Ziyan Weng, Jinhan Li 외 arxiv

Long-horizon embodied intelligence requires agents to improve through interaction, not merely to execute plans generated from static goals. A central challenge is therefore to transform past executions into knowledge tha…

ProPlay: Procedural World Models for Self-Evolving LLM Agents

2026-06-11 · Yijun Ma, Zehong Wang, Yiyang Li, Ziming Li 외 arxiv

Self-evolving agents are expected to improve through interaction without external supervision, but this remains difficult in partially observable environments where agents must explore actively, learn from limited feedba…