{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YEZ6FCCBMOYUPQQ6PEFAPNPNSC","short_pith_number":"pith:YEZ6FCCB","schema_version":"1.0","canonical_sha256":"c133e2884163b147c21e790a07b5ed90b73c76a5a7f0b521f380e3e0d655feb7","source":{"kind":"arxiv","id":"2508.04700","version":2},"attestation_state":"computed","paper":{"title":"SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.LG","cs.MA","cs.MM"],"primary_cat":"cs.AI","authors_text":"Dahua Lin, Jiaqi Wang, Tong Wu, Xiaoyi Dong, Yuhang Cao, Yuhang Zang, Zeyi Sun, Ziyu Liu","submitted_at":"2025-08-06T17:58:46Z","abstract_excerpt":"Repurposing large vision-language models (LVLMs) as computer use agents (CUAs) has led to substantial breakthroughs, primarily driven by human-labeled data. However, these models often struggle with novel and specialized software, particularly in scenarios lacking human annotations. To address this challenge, we propose SEAgent, an agentic self-evolving framework enabling CUAs to autonomously evolve through interactions with unfamiliar software. Specifically, SEAgent empowers computer-use agents to autonomously master novel software environments via experiential learning, where agents explore "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2508.04700","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-08-06T17:58:46Z","cross_cats_sorted":["cs.CL","cs.CV","cs.LG","cs.MA","cs.MM"],"title_canon_sha256":"78a1e975ce5c4d69ca4eb7b0a852632f6b6d0768aad87f55448de3d41e91f48c","abstract_canon_sha256":"68e218b2311cc078b205f6b729e718d36ae3ea0a186ae423a5b85139031125c5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:23.247136Z","signature_b64":"dszW9ZNyd90QnyNcmAsqKYywrxTNSBYldVUQbgBkl98NHjmjiKT1mGuCs0SjlQFxiywjI0hRxDMrzUSZ/LmGBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c133e2884163b147c21e790a07b5ed90b73c76a5a7f0b521f380e3e0d655feb7","last_reissued_at":"2026-07-05T11:52:23.246612Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:23.246612Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.LG","cs.MA","cs.MM"],"primary_cat":"cs.AI","authors_text":"Dahua Lin, Jiaqi Wang, Tong Wu, Xiaoyi Dong, Yuhang Cao, Yuhang Zang, Zeyi Sun, Ziyu Liu","submitted_at":"2025-08-06T17:58:46Z","abstract_excerpt":"Repurposing large vision-language models (LVLMs) as computer use agents (CUAs) has led to substantial breakthroughs, primarily driven by human-labeled data. However, these models often struggle with novel and specialized software, particularly in scenarios lacking human annotations. To address this challenge, we propose SEAgent, an agentic self-evolving framework enabling CUAs to autonomously evolve through interactions with unfamiliar software. Specifically, SEAgent empowers computer-use agents to autonomously master novel software environments via experiential learning, where agents explore "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.04700","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.04700/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2508.04700","created_at":"2026-07-05T11:52:23.246679+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.04700v2","created_at":"2026-07-05T11:52:23.246679+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.04700","created_at":"2026-07-05T11:52:23.246679+00:00"},{"alias_kind":"pith_short_12","alias_value":"YEZ6FCCBMOYU","created_at":"2026-07-05T11:52:23.246679+00:00"},{"alias_kind":"pith_short_16","alias_value":"YEZ6FCCBMOYUPQQ6","created_at":"2026-07-05T11:52:23.246679+00:00"},{"alias_kind":"pith_short_8","alias_value":"YEZ6FCCB","created_at":"2026-07-05T11:52:23.246679+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":22,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19930","citing_title":"MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.19338","citing_title":"Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09447","citing_title":"AliyunConsoleAgent: Training Web Agents in Real-World Cloud Environments via Distillation and Reinforcement Learning","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07367","citing_title":"Self-evolving LLM agents with in-distribution Optimization","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01936","citing_title":"What to Format and How: A Benchmark and Workflow Approach for Document Formatting","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00384","citing_title":"VESTA: Visual Exploration with Statistical Tool Agents","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26200","citing_title":"Workflow Closure Is Not Scientific Closure in Auto-Research Systems","ref_index":99,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31270","citing_title":"Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24426","citing_title":"SEAL: Synergistic Co-Evolution of Agents and Learning Environments","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29705","citing_title":"GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25624","citing_title":"CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28775","citing_title":"Learn from Weaknesses: Automated Domain Specialization for Small Computer-Use Agents","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26614","citing_title":"HiLSVA: Design and Evaluation of a Human-in-the-Loop Agentic System for Scientific Visualization","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18652","citing_title":"MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2602.06663","citing_title":"PlanViz: Evaluating Planning-Oriented Image Generation and Editing for Computer-Use Tasks","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27996","citing_title":"Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2602.12430","citing_title":"Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11169","citing_title":"OLIVIA: Online Learning via Inference-time Action Adaptation for Decision Making in LLM ReAct Agents","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11039","citing_title":"The Granularity Mismatch in Agent Security: Argument-Level Provenance Solves Enforcement and Isolates the LLM Reasoning Bottleneck","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27996","citing_title":"Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09359","citing_title":"Skill-R1: Agent Skill Evolution via Reinforcement Learning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10064","citing_title":"MAGE: Multi-Agent Self-Evolution with Co-Evolutionary Knowledge Graphs","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC","json":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC.json","graph_json":"https://pith.science/api/pith-number/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/graph.json","events_json":"https://pith.science/api/pith-number/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/events.json","paper":"https://pith.science/paper/YEZ6FCCB"},"agent_actions":{"view_html":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC","download_json":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC.json","view_paper":"https://pith.science/paper/YEZ6FCCB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.04700&json=true","fetch_graph":"https://pith.science/api/pith-number/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/graph.json","fetch_events":"https://pith.science/api/pith-number/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/action/storage_attestation","attest_author":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/action/author_attestation","sign_citation":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/action/citation_signature","submit_replication":"https://pith.science/pith/YEZ6FCCBMOYUPQQ6PEFAPNPNSC/action/replication_record"}},"created_at":"2026-07-05T11:52:23.246679+00:00","updated_at":"2026-07-05T11:52:23.246679+00:00"}