Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2026-01-16T01:38:37.356Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7056677341461182},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2601.09465","authors":[{"_id":"696867a90ac10a06522f6a82","user":{"_id":"65562edfb7bad186e877c724","avatarUrl":"/avatars/bb91f42b102e113208bbe3238916a015.svg","isPro":false,"fullname":"zhangshuo","user":"mcflurryshuoz","type":"user"},"name":"Shuo Zhang","status":"claimed_verified","statusLastChangedAt":"2026-01-15T09:33:43.605Z","hidden":false},{"_id":"696867a90ac10a06522f6a83","name":"Chaofa Yuan","hidden":false},{"_id":"696867a90ac10a06522f6a84","name":"Ryan Guo","hidden":false},{"_id":"696867a90ac10a06522f6a85","user":{"_id":"64084fa192033c150738e4f2","avatarUrl":"/avatars/dfff2216eb235c635e5abe6fda3084f0.svg","isPro":false,"fullname":"Yu_xm","user":"Yu2020","type":"user"},"name":"Xiaomin Yu","status":"claimed_verified","statusLastChangedAt":"2026-01-15T15:02:52.674Z","hidden":false},{"_id":"696867a90ac10a06522f6a86","name":"Rui Xu","hidden":false},{"_id":"696867a90ac10a06522f6a87","name":"Zhangquan Chen","hidden":false},{"_id":"696867a90ac10a06522f6a88","name":"Zinuo Li","hidden":false},{"_id":"696867a90ac10a06522f6a89","user":{"_id":"64aa645404e7b379feccc490","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64aa645404e7b379feccc490/4m8qcdy2OGK8visR5Jjl5.png","isPro":false,"fullname":"Zhi Yang","user":"yangzhi1","type":"user"},"name":"Zhi Yang","status":"claimed_verified","statusLastChangedAt":"2026-01-15T09:33:45.582Z","hidden":false},{"_id":"696867a90ac10a06522f6a8a","name":"Shuhao Guan","hidden":false},{"_id":"696867a90ac10a06522f6a8b","name":"Zhenheng Tang","hidden":false},{"_id":"696867a90ac10a06522f6a8c","name":"Sen Hu","hidden":false},{"_id":"696867a90ac10a06522f6a8d","user":{"_id":"64d4aa25263aa362d4644737","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d4aa25263aa362d4644737/XB4VybeUyLpz1qVtqXHbp.png","isPro":false,"fullname":"Liwen Zhang","user":"SUFE-AIFLM-Lab","type":"user"},"name":"Liwen Zhang","status":"claimed_verified","statusLastChangedAt":"2026-01-16T15:28:47.062Z","hidden":false},{"_id":"696867a90ac10a06522f6a8e","name":"Ronghao Chen","hidden":false},{"_id":"696867a90ac10a06522f6a8f","name":"Huacan Wang","hidden":false}],"publishedAt":"2026-01-14T13:19:13.000Z","submittedOnDailyAt":"2026-01-15T01:36:08.471Z","title":"EvoFSM: Controllable Self-Evolution for Deep Research with Finite State Machines","submittedOnDailyBy":{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},"summary":"While LLM-based agents have shown promise for deep research, most existing approaches rely on fixed workflows that struggle to adapt to real-world, open-ended queries. Recent work therefore explores self-evolution by allowing agents to rewrite their own code or prompts to improve problem-solving ability, but unconstrained optimization often triggers instability, hallucinations, and instruction drift. We propose EvoFSM, a structured self-evolving framework that achieves both adaptability and control by evolving an explicit Finite State Machine (FSM) instead of relying on free-form rewriting. EvoFSM decouples the optimization space into macroscopic Flow (state-transition logic) and microscopic Skill (state-specific behaviors), enabling targeted improvements under clear behavioral boundaries. Guided by a critic mechanism, EvoFSM refines the FSM through a small set of constrained operations, and further incorporates a self-evolving memory that distills successful trajectories as reusable priors and failure patterns as constraints for future queries. Extensive evaluations on five multi-hop QA benchmarks demonstrate the effectiveness of EvoFSM. In particular, EvoFSM reaches 58.0% accuracy on the DeepSearch benchmark. Additional results on interactive decision-making tasks further validate its generalization.","upvotes":41,"discussionId":"696867a90ac10a06522f6a90","ai_summary":"EvoFSM is a structured self-evolving framework for LLM agents that uses finite state machines to improve adaptability while maintaining control through constrained optimization and memory mechanisms.","ai_keywords":["LLM-based agents","self-evolution","finite state machine","workflow adaptation","constrained optimization","critic mechanism","self-evolving memory","multi-hop QA","DeepSearch benchmark","interactive decision-making"],"organization":{"_id":"68b33ab6a9ed99140481cf44","name":"QuantaAlpha","fullname":"QuantaAlpha","avatar":"https://cdn-uploads.huggingface.co/production/uploads/63f7767fbd28622c9b9915e9/DRN8PvmnpKmn2MSLQ7qhF.jpeg"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"684d57f26e04c265777ead3f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/cuOj-bQqukSZreXgUJlfm.png","isPro":false,"fullname":"Joakim Lee","user":"Reinforcement4All","type":"user"},{"_id":"66a0ab4923e426e19db92773","avatarUrl":"/avatars/19517dd085a3e48e644613ca0b2c3753.svg","isPro":false,"fullname":"ronghaochen","user":"cristiano28","type":"user"},{"_id":"64084fa192033c150738e4f2","avatarUrl":"/avatars/dfff2216eb235c635e5abe6fda3084f0.svg","isPro":false,"fullname":"Yu_xm","user":"Yu2020","type":"user"},{"_id":"67eced92cf3e57ee31806ea9","avatarUrl":"/avatars/88d9026b7505596477bbda5ee9fa0972.svg","isPro":false,"fullname":"Gong Chuanzheng","user":"linkkkkkk","type":"user"},{"_id":"68eb753bd7b8e72eb5a53e10","avatarUrl":"/avatars/3ef63182b4d929f5b670aa248474409c.svg","isPro":false,"fullname":"ben_hit_save","user":"ben-hit-save","type":"user"},{"_id":"65562edfb7bad186e877c724","avatarUrl":"/avatars/bb91f42b102e113208bbe3238916a015.svg","isPro":false,"fullname":"zhangshuo","user":"mcflurryshuoz","type":"user"},{"_id":"648a78280d49ffe259ede456","avatarUrl":"/avatars/190c520e9cb44bdc7908bc5d4415256b.svg","isPro":false,"fullname":"zhaopengshuai","user":"ZPS45683968","type":"user"},{"_id":"692881094c3f4293dfe29e3d","avatarUrl":"/avatars/bddfaae8041a45498d46ef65ba17c920.svg","isPro":false,"fullname":"qihao wang","user":"jimson991","type":"user"},{"_id":"6965ec520ccf0b50b9ab6864","avatarUrl":"/avatars/a7a509479429c64ec6e4b86acf36d5c1.svg","isPro":false,"fullname":"Tinghe Yan","user":"TingheYan","type":"user"},{"_id":"636249a34c9f829ef7f7c710","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/636249a34c9f829ef7f7c710/M7DF88ByGAAXwARjEy29l.jpeg","isPro":false,"fullname":"bbruceyuan","user":"bbruceyuan","type":"user"},{"_id":"687f601b7170fd281b898c0f","avatarUrl":"/avatars/d77ceb1dca5115d50abb1bb300d05209.svg","isPro":false,"fullname":"yifan","user":"dongyifan","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0,"organization":{"_id":"68b33ab6a9ed99140481cf44","name":"QuantaAlpha","fullname":"QuantaAlpha","avatar":"https://cdn-uploads.huggingface.co/production/uploads/63f7767fbd28622c9b9915e9/DRN8PvmnpKmn2MSLQ7qhF.jpeg"}}">

Papers

arxiv:2601.09465

EvoFSM: Controllable Self-Evolution for Deep Research with Finite State Machines

Published on Jan 14

· Submitted by

taesiri on Jan 15

QuantaAlpha

Upvote

Authors:

Shuo Zhang ,

Xiaomin Yu ,

Zhi Yang ,

Liwen Zhang ,

Abstract

EvoFSM is a structured self-evolving framework for LLM agents that uses finite state machines to improve adaptability while maintaining control through constrained optimization and memory mechanisms.

AI-generated summary

While LLM-based agents have shown promise for deep research, most existing approaches rely on fixed workflows that struggle to adapt to real-world, open-ended queries. Recent work therefore explores self-evolution by allowing agents to rewrite their own code or prompts to improve problem-solving ability, but unconstrained optimization often triggers instability, hallucinations, and instruction drift. We propose EvoFSM, a structured self-evolving framework that achieves both adaptability and control by evolving an explicit Finite State Machine (FSM) instead of relying on free-form rewriting. EvoFSM decouples the optimization space into macroscopic Flow (state-transition logic) and microscopic Skill (state-specific behaviors), enabling targeted improvements under clear behavioral boundaries. Guided by a critic mechanism, EvoFSM refines the FSM through a small set of constrained operations, and further incorporates a self-evolving memory that distills successful trajectories as reusable priors and failure patterns as constraints for future queries. Extensive evaluations on five multi-hop QA benchmarks demonstrate the effectiveness of EvoFSM. In particular, EvoFSM reaches 58.0% accuracy on the DeepSearch benchmark. Additional results on interactive decision-making tasks further validate its generalization.

View arXiv page View PDF Add to collection

Community

taesiri

Paper submitter Jan 15

EvoFSM presents a controllable self-evolution framework using a finite state machine to guide adaptive problem-solving, separating macroscopic flow and microscopic skills with critic-guided updates and reusable priors.