{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:OGQWIBWMF4TLRQTVSZNQ5MP7KS","short_pith_number":"pith:OGQWIBWM","schema_version":"1.0","canonical_sha256":"71a16406cc2f26b8c275965b0eb1ff54999aec9f1b72c9fcc91e48f665e0bd10","source":{"kind":"arxiv","id":"2506.03097","version":1},"attestation_state":"computed","paper":{"title":"EgoVLM: Policy Optimization for Egocentric Video Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Aditya Vavre, Ashwin Vinod, Linshen Liu, Shrey Pandit","submitted_at":"2025-06-03T17:28:00Z","abstract_excerpt":"Emerging embodied AI applications, such as wearable cameras and autonomous agents, have underscored the need for robust reasoning from first person video streams. We introduce EgoVLM, a vision-language model specifically designed to integrate visual comprehension and spatial-temporal reasoning within egocentric video contexts. EgoVLM is fine-tuned via Group Relative Policy Optimization (GRPO), a reinforcement learning method adapted to align model outputs with human-like reasoning steps. Following DeepSeek R1-Zero's approach, we directly tune using RL without any supervised fine-tuning phase o"},"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":"2506.03097","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-03T17:28:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b06ae509808cf75faac24868d9257d6373e7c0c921f5579cb592a2c44d8e401f","abstract_canon_sha256":"717b348c7baa77d13599f9ee0d5cc73ad9ad3ac227fe06d9e815ce4ca9f739c7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:15:14.828094Z","signature_b64":"CfcL+1hgd/XLtKvAytl3uvxejAaaIhU1Ja7hUG1dWI1XPPHy+sUFF0AQjADidALum56WFNZTR9p5MUnxe5TeAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"71a16406cc2f26b8c275965b0eb1ff54999aec9f1b72c9fcc91e48f665e0bd10","last_reissued_at":"2026-07-05T11:15:14.827590Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:15:14.827590Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EgoVLM: Policy Optimization for Egocentric Video Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Aditya Vavre, Ashwin Vinod, Linshen Liu, Shrey Pandit","submitted_at":"2025-06-03T17:28:00Z","abstract_excerpt":"Emerging embodied AI applications, such as wearable cameras and autonomous agents, have underscored the need for robust reasoning from first person video streams. We introduce EgoVLM, a vision-language model specifically designed to integrate visual comprehension and spatial-temporal reasoning within egocentric video contexts. EgoVLM is fine-tuned via Group Relative Policy Optimization (GRPO), a reinforcement learning method adapted to align model outputs with human-like reasoning steps. Following DeepSeek R1-Zero's approach, we directly tune using RL without any supervised fine-tuning phase o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.03097","kind":"arxiv","version":1},"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/2506.03097/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":"2506.03097","created_at":"2026-07-05T11:15:14.827648+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.03097v1","created_at":"2026-07-05T11:15:14.827648+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.03097","created_at":"2026-07-05T11:15:14.827648+00:00"},{"alias_kind":"pith_short_12","alias_value":"OGQWIBWMF4TL","created_at":"2026-07-05T11:15:14.827648+00:00"},{"alias_kind":"pith_short_16","alias_value":"OGQWIBWMF4TLRQTV","created_at":"2026-07-05T11:15:14.827648+00:00"},{"alias_kind":"pith_short_8","alias_value":"OGQWIBWM","created_at":"2026-07-05T11:15:14.827648+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.27621","citing_title":"Robot Learning from Human Videos: A Survey","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04227","citing_title":"Pro$^2$Assist: Continuous Step-aware Proactive Assistance with Multi-modal Egocentric Perception for Long-horizon Procedural Tasks","ref_index":65,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS","json":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS.json","graph_json":"https://pith.science/api/pith-number/OGQWIBWMF4TLRQTVSZNQ5MP7KS/graph.json","events_json":"https://pith.science/api/pith-number/OGQWIBWMF4TLRQTVSZNQ5MP7KS/events.json","paper":"https://pith.science/paper/OGQWIBWM"},"agent_actions":{"view_html":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS","download_json":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS.json","view_paper":"https://pith.science/paper/OGQWIBWM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.03097&json=true","fetch_graph":"https://pith.science/api/pith-number/OGQWIBWMF4TLRQTVSZNQ5MP7KS/graph.json","fetch_events":"https://pith.science/api/pith-number/OGQWIBWMF4TLRQTVSZNQ5MP7KS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS/action/storage_attestation","attest_author":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS/action/author_attestation","sign_citation":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS/action/citation_signature","submit_replication":"https://pith.science/pith/OGQWIBWMF4TLRQTVSZNQ5MP7KS/action/replication_record"}},"created_at":"2026-07-05T11:15:14.827648+00:00","updated_at":"2026-07-05T11:15:14.827648+00:00"}