{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BSD73E6C54RXWSJQ56JMW63TNQ","short_pith_number":"pith:BSD73E6C","schema_version":"1.0","canonical_sha256":"0c87fd93c2ef237b4930ef92cb7b736c3de1530cdd3f91a2d6a5149bd111338f","source":{"kind":"arxiv","id":"2605.15342","version":1},"attestation_state":"computed","paper":{"title":"Minerva-Ego: Spatiotemporal Hints for Egocentric Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Providing models with hints on where and when to look improves performance on complex egocentric video reasoning tasks.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Arsha Nagrani, Bo Hu, Cordelia Schmid, David A Ross, Jasper Uijilings, Ramin Mehran, Shyamal Buch, Sudheendra Vijayanarasimhan, Tobias Weyand","submitted_at":"2026-05-14T19:12:20Z","abstract_excerpt":"Video reasoning models are a core component of egocentric and embodied agents. However, standard benchmarks for assessing models provide only evaluation of the output (e.g. the answer to a question), without evaluation of intermediate reasoning steps, and most provide answers only in the text domain. We introduce Minerva-Ego, a benchmark for evaluating complex egocentric visual reasoning. We extend recent high-quality video data sources recorded from egocentric / embodied settings with a set of challenging, multi-step multimodal questions and spatiotemporally-dense human-annotated reasoning tr"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.15342","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-14T19:12:20Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"58930be17e6972768f6b19177c4427f1535b66aae4840693dc3cee6c2099e079","abstract_canon_sha256":"52b816aa18674263f1c32c9aefb0af114f5f3efe9a12b3f76aaa98495ba5e9d6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:53.469798Z","signature_b64":"NE86/463HjMANsVJY/ygtnSz7FmLflfd50ENVHZickK88RgDz7EJKuvAvTReBn4BzB/ZksPpFDKzO8iH0KLvBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0c87fd93c2ef237b4930ef92cb7b736c3de1530cdd3f91a2d6a5149bd111338f","last_reissued_at":"2026-05-20T00:00:53.468886Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:53.468886Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Minerva-Ego: Spatiotemporal Hints for Egocentric Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Providing models with hints on where and when to look improves performance on complex egocentric video reasoning tasks.","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Arsha Nagrani, Bo Hu, Cordelia Schmid, David A Ross, Jasper Uijilings, Ramin Mehran, Shyamal Buch, Sudheendra Vijayanarasimhan, Tobias Weyand","submitted_at":"2026-05-14T19:12:20Z","abstract_excerpt":"Video reasoning models are a core component of egocentric and embodied agents. However, standard benchmarks for assessing models provide only evaluation of the output (e.g. the answer to a question), without evaluation of intermediate reasoning steps, and most provide answers only in the text domain. We introduce Minerva-Ego, a benchmark for evaluating complex egocentric visual reasoning. We extend recent high-quality video data sources recorded from egocentric / embodied settings with a set of challenging, multi-step multimodal questions and spatiotemporally-dense human-annotated reasoning tr"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Prompting frontier models with hints of 'where' and 'when' to look yields substantial improvements in performance.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The spatiotemporally-dense human-annotated reasoning traces and object masks accurately capture the intermediate steps and visual elements required to solve the multi-step questions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Providing models with hints on where and when to look improves performance on complex egocentric video reasoning tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"94ed0b7c10ce8ca4c9cc30dd5c8e9f698c9fc83c044ab463e3f05c971fde9369"},"source":{"id":"2605.15342","kind":"arxiv","version":1},"verdict":{"id":"5dc1b6de-7966-422a-8776-926c0337aa7a","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T15:57:52.079820Z","strongest_claim":"Prompting frontier models with hints of 'where' and 'when' to look yields substantial improvements in performance.","one_line_summary":"Minerva-Ego is a new benchmark for egocentric visual reasoning with dense human-annotated traces and masks, showing that spatiotemporal hints substantially improve frontier model performance.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The spatiotemporally-dense human-annotated reasoning traces and object masks accurately capture the intermediate steps and visual elements required to solve the multi-step questions.","pith_extraction_headline":"Providing models with hints on where and when to look improves performance on complex egocentric video reasoning tasks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15342/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T16:31:18.280913Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T16:04:47.756146Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T14:21:54.209955Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.755355Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"dda780d69beba49a28c4987322e879fc0ad26903567dad2ce0745c2bd31cc3d3"},"references":{"count":66,"sample":[{"doi":"","year":2023,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2025,"title":"https://openai.com/index/learning- to-reason-with-llms , 2025","work_id":"6141519e-dd8d-4028-ab49-6d63eaa797ea","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"System Card: Claude Opus 4 & Claude Sonnet 4","work_id":"53291f48-fd2d-4d0e-8c77-8dc39d7130de","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Claude 3.5 sonnet v2","work_id":"1025c3c2-15f4-4575-b01b-531d67405eb2","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"In- finiBench: A comprehensive benchmark for large multimodal models in very long video understanding.arXiv preprint arXiv:2406.19875, 2024","work_id":"8aed6ea0-787d-4a47-9cdc-963d83f6de2a","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":66,"snapshot_sha256":"ea42c2c1b0fe92fe0d6e84d62e370e04213601f9903e198cfe277c8adafa7f8f","internal_anchors":10},"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":"2605.15342","created_at":"2026-05-20T00:00:53.469036+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15342v1","created_at":"2026-05-20T00:00:53.469036+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15342","created_at":"2026-05-20T00:00:53.469036+00:00"},{"alias_kind":"pith_short_12","alias_value":"BSD73E6C54RX","created_at":"2026-05-20T00:00:53.469036+00:00"},{"alias_kind":"pith_short_16","alias_value":"BSD73E6C54RXWSJQ","created_at":"2026-05-20T00:00:53.469036+00:00"},{"alias_kind":"pith_short_8","alias_value":"BSD73E6C","created_at":"2026-05-20T00:00:53.469036+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ","json":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ.json","graph_json":"https://pith.science/api/pith-number/BSD73E6C54RXWSJQ56JMW63TNQ/graph.json","events_json":"https://pith.science/api/pith-number/BSD73E6C54RXWSJQ56JMW63TNQ/events.json","paper":"https://pith.science/paper/BSD73E6C"},"agent_actions":{"view_html":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ","download_json":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ.json","view_paper":"https://pith.science/paper/BSD73E6C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15342&json=true","fetch_graph":"https://pith.science/api/pith-number/BSD73E6C54RXWSJQ56JMW63TNQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BSD73E6C54RXWSJQ56JMW63TNQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ/action/storage_attestation","attest_author":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ/action/author_attestation","sign_citation":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ/action/citation_signature","submit_replication":"https://pith.science/pith/BSD73E6C54RXWSJQ56JMW63TNQ/action/replication_record"}},"created_at":"2026-05-20T00:00:53.469036+00:00","updated_at":"2026-05-20T00:00:53.469036+00:00"}