{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FMB22R3DOBYXRNZYHNN6QT53CF","short_pith_number":"pith:FMB22R3D","schema_version":"1.0","canonical_sha256":"2b03ad4763707178b7383b5be84fbb115b918270e494adc8e4dc46107ead5f53","source":{"kind":"arxiv","id":"2306.00245","version":2},"attestation_state":"computed","paper":{"title":"From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.HC"],"primary_cat":"cs.LG","authors_text":"Hexiang Hu, James Cohan, Jonathan Berant, Kenton Lee, Kristina Toutanova, Mandar Joshi, Panupong Pasupat, Peter Shaw, Urvashi Khandelwal","submitted_at":"2023-05-31T23:39:18Z","abstract_excerpt":"Much of the previous work towards digital agents for graphical user interfaces (GUIs) has relied on text-based representations (derived from HTML or other structured data sources), which are not always readily available. These input representations have been often coupled with custom, task-specific action spaces. This paper focuses on creating agents that interact with the digital world using the same conceptual interface that humans commonly use -- via pixel-based screenshots and a generic action space corresponding to keyboard and mouse actions. Building upon recent progress in pixel-based p"},"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":"2306.00245","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-05-31T23:39:18Z","cross_cats_sorted":["cs.CL","cs.CV","cs.HC"],"title_canon_sha256":"2f4b0c70cb7549cd5b079e4a1f8ba0795500908c4cf1206c5c02a5505fc25c9d","abstract_canon_sha256":"8b90293a6138e9391dd64fd0523b385a7ef4a0a915e044aaa5cf34922de6b073"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:21:11.200202Z","signature_b64":"FP2ECdptLnRbAAMdIAqndIydCxQ0Vxdt4RHut5nx7oB83WBjvsgFUUhtW24FELqEkDi35nIe804rPQSjMLsdDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b03ad4763707178b7383b5be84fbb115b918270e494adc8e4dc46107ead5f53","last_reissued_at":"2026-07-05T07:21:11.199685Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:21:11.199685Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV","cs.HC"],"primary_cat":"cs.LG","authors_text":"Hexiang Hu, James Cohan, Jonathan Berant, Kenton Lee, Kristina Toutanova, Mandar Joshi, Panupong Pasupat, Peter Shaw, Urvashi Khandelwal","submitted_at":"2023-05-31T23:39:18Z","abstract_excerpt":"Much of the previous work towards digital agents for graphical user interfaces (GUIs) has relied on text-based representations (derived from HTML or other structured data sources), which are not always readily available. These input representations have been often coupled with custom, task-specific action spaces. This paper focuses on creating agents that interact with the digital world using the same conceptual interface that humans commonly use -- via pixel-based screenshots and a generic action space corresponding to keyboard and mouse actions. Building upon recent progress in pixel-based p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.00245","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/2306.00245/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":"2306.00245","created_at":"2026-07-05T07:21:11.199750+00:00"},{"alias_kind":"arxiv_version","alias_value":"2306.00245v2","created_at":"2026-07-05T07:21:11.199750+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.00245","created_at":"2026-07-05T07:21:11.199750+00:00"},{"alias_kind":"pith_short_12","alias_value":"FMB22R3DOBYX","created_at":"2026-07-05T07:21:11.199750+00:00"},{"alias_kind":"pith_short_16","alias_value":"FMB22R3DOBYXRNZY","created_at":"2026-07-05T07:21:11.199750+00:00"},{"alias_kind":"pith_short_8","alias_value":"FMB22R3D","created_at":"2026-07-05T07:21:11.199750+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29537","citing_title":"OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16565","citing_title":"Skim: Speculative Execution for Fast and Efficient Web Agents","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2401.10935","citing_title":"SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2401.01614","citing_title":"GPT-4V(ision) is a Generalist Web Agent, if Grounded","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2404.07972","citing_title":"OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF","json":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF.json","graph_json":"https://pith.science/api/pith-number/FMB22R3DOBYXRNZYHNN6QT53CF/graph.json","events_json":"https://pith.science/api/pith-number/FMB22R3DOBYXRNZYHNN6QT53CF/events.json","paper":"https://pith.science/paper/FMB22R3D"},"agent_actions":{"view_html":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF","download_json":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF.json","view_paper":"https://pith.science/paper/FMB22R3D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2306.00245&json=true","fetch_graph":"https://pith.science/api/pith-number/FMB22R3DOBYXRNZYHNN6QT53CF/graph.json","fetch_events":"https://pith.science/api/pith-number/FMB22R3DOBYXRNZYHNN6QT53CF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF/action/storage_attestation","attest_author":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF/action/author_attestation","sign_citation":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF/action/citation_signature","submit_replication":"https://pith.science/pith/FMB22R3DOBYXRNZYHNN6QT53CF/action/replication_record"}},"created_at":"2026-07-05T07:21:11.199750+00:00","updated_at":"2026-07-05T07:21:11.199750+00:00"}