{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:QIQWH6BYM5BQKFXEQ4BE4XQO2H","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"02578bfa6f47c678462ed2526ac1a460eaaee929b86309a493c3c5ee9dd69693","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-11T17:21:00Z","title_canon_sha256":"2334de0b5de562250b0177cc3d3d15e4cf31b9170e0a5fd37f50f018fb48121f"},"schema_version":"1.0","source":{"id":"2409.07429","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.07429","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"arxiv_version","alias_value":"2409.07429v1","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.07429","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"pith_short_12","alias_value":"QIQWH6BYM5BQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"QIQWH6BYM5BQKFXE","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"QIQWH6BY","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:775d70d3eb91d2007f6e359815e53e313d17a2dabd66a1c2cdd91d1c966170c1","target":"graph","created_at":"2026-05-17T23:39:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"AWM substantially improves the baseline results by 24.6% and 51.1% relative success rate on Mind2Web and WebArena while reducing the number of steps taken to solve WebArena tasks successfully. Furthermore, online AWM robustly generalizes in cross-task, website, and domain evaluations."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That workflows induced from past examples can be reliably identified as reusable and selectively provided without introducing noise or incorrect guidance that harms the agent's generation process on new queries."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"AWM induces reusable workflows from agent experiences and provides them selectively to improve success rates by 24.6% on Mind2Web and 51.1% on WebArena while reducing steps taken."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Agent Workflow Memory extracts reusable task routines from past examples to guide language model agents on new web navigation tasks."}],"snapshot_sha256":"cca5dd8afd234ed69fd894ee7a4d864c18221d17f95799ae9b71b615660d378d"},"formal_canon":{"evidence_count":3,"snapshot_sha256":"d84e2aa5161256fdc62e9f25446c3f98b4d28470786b022fe6fff3eb219b4a4f"},"paper":{"abstract_excerpt":"Despite the potential of language model-based agents to solve real-world tasks such as web navigation, current methods still struggle with long-horizon tasks with complex action trajectories. In contrast, humans can flexibly solve complex tasks by learning reusable task workflows from past experiences and using them to guide future actions. To build agents that can similarly benefit from this process, we introduce Agent Workflow Memory (AWM), a method for inducing commonly reused routines, i.e., workflows, and selectively providing workflows to the agent to guide subsequent generations. AWM fl","authors_text":"Daniel Fried, Graham Neubig, Jiayuan Mao, Zora Zhiruo Wang","cross_cats":[],"headline":"Agent Workflow Memory extracts reusable task routines from past examples to guide language model agents on new web navigation tasks.","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-11T17:21:00Z","title":"Agent Workflow Memory"},"references":{"count":53,"internal_anchors":2,"resolved_work":53,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Proceedings of the 34th International Conference on Machine Learning , pages=","work_id":"be6e663e-22ef-435c-9083-e7561430a01e","year":2017},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"International Conference on Learning Representations , year=","work_id":"16a4c773-aff5-4990-915d-355b14c5ea2a","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents , url=","work_id":"3329f59b-7615-4c1c-913d-22fcecac99cd","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"The Twelfth International Conference on Learning Representations , year=","work_id":"440637a6-6d56-4bae-8884-02313ea3f325","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track , year=","work_id":"c27ea3b4-c03e-46f6-b46b-bb8af9f71751","year":null}],"snapshot_sha256":"fafa0b2278afbb69259f1ced109a8c08f7bf5d4e4f6bc7abc58bfdb3efd4b699"},"source":{"id":"2409.07429","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T00:47:34.019674Z","id":"b6644fdd-093a-4cb6-8809-4d059483a2fa","model_set":{"reader":"grok-4.3"},"one_line_summary":"AWM induces reusable workflows from agent experiences and provides them selectively to improve success rates by 24.6% on Mind2Web and 51.1% on WebArena while reducing steps taken.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Agent Workflow Memory extracts reusable task routines from past examples to guide language model agents on new web navigation tasks.","strongest_claim":"AWM substantially improves the baseline results by 24.6% and 51.1% relative success rate on Mind2Web and WebArena while reducing the number of steps taken to solve WebArena tasks successfully. Furthermore, online AWM robustly generalizes in cross-task, website, and domain evaluations.","weakest_assumption":"That workflows induced from past examples can be reliably identified as reusable and selectively provided without introducing noise or incorrect guidance that harms the agent's generation process on new queries."}},"verdict_id":"b6644fdd-093a-4cb6-8809-4d059483a2fa"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:91840005ebce2cbcd278f51a302fa50764a307ca425e91118bea0f3ef43280b3","target":"record","created_at":"2026-05-17T23:39:05Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"02578bfa6f47c678462ed2526ac1a460eaaee929b86309a493c3c5ee9dd69693","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-11T17:21:00Z","title_canon_sha256":"2334de0b5de562250b0177cc3d3d15e4cf31b9170e0a5fd37f50f018fb48121f"},"schema_version":"1.0","source":{"id":"2409.07429","kind":"arxiv","version":1}},"canonical_sha256":"822163f83867430516e487024e5e0ed1c0adc8371db017172de12240604399e5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"822163f83867430516e487024e5e0ed1c0adc8371db017172de12240604399e5","first_computed_at":"2026-05-17T23:39:05.201050Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:05.201050Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hY/wKmBhhLGAPtkp8PnwCt8tVfVHw7AzfcTFqX3Znne3NYWL57TudTP6m+1wIqRNHqT2Zw6iOMAfAsEbum31Bw==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:05.201768Z","signed_message":"canonical_sha256_bytes"},"source_id":"2409.07429","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:91840005ebce2cbcd278f51a302fa50764a307ca425e91118bea0f3ef43280b3","sha256:775d70d3eb91d2007f6e359815e53e313d17a2dabd66a1c2cdd91d1c966170c1"],"state_sha256":"3304298acdcb55b675999695b7a24696242defd1ea5af952e4b71bd61de18e56"}