{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:A5DAUK3U3OYFWU2QGULP33AY62","short_pith_number":"pith:A5DAUK3U","schema_version":"1.0","canonical_sha256":"07460a2b74dbb05b53503516fdec18f6831da3c4c264590a422d53b0092c54f9","source":{"kind":"arxiv","id":"2506.13654","version":1},"attestation_state":"computed","paper":{"title":"Ego-R1: Chain-of-Tool-Thought for Ultra-Long Egocentric Video Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hao Zhang, Hongming Guo, Hongyuan Zhu, Jingkang Yang, Penghao Wu, Ruiqi Wang, Shulin Tian, Xiuying Wang, Yuhao Dong, Ziwei Liu","submitted_at":"2025-06-16T16:17:08Z","abstract_excerpt":"We introduce Ego-R1, a novel framework for reasoning over ultra-long (i.e., in days and weeks) egocentric videos, which leverages a structured Chain-of-Tool-Thought (CoTT) process, orchestrated by an Ego-R1 Agent trained via reinforcement learning (RL). Inspired by human problem-solving strategies, CoTT decomposes complex reasoning into modular steps, with the RL agent invoking specific tools, one per step, to iteratively and collaboratively answer sub-questions tackling such tasks as temporal retrieval and multi-modal understanding. We design a two-stage training paradigm involving supervised"},"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":"2506.13654","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-16T16:17:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b281807f94000931e2909648a024762656df85791849bc5e9b68836a94ae5412","abstract_canon_sha256":"f519f8791d038ec5792f7b89d138810b6b2bd47dfb80d596ee8bf6010497bc63"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:22:24.373375Z","signature_b64":"4LwT73IgTiFLbg1uaW4yKozZAJiKPbG2wYXZ+uFgJlMZlvCe3Od3hZnDPBL/vFebey6htGWi46wsqA0vNFKHAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"07460a2b74dbb05b53503516fdec18f6831da3c4c264590a422d53b0092c54f9","last_reissued_at":"2026-07-05T11:22:24.372812Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:22:24.372812Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ego-R1: Chain-of-Tool-Thought for Ultra-Long Egocentric Video Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Hao Zhang, Hongming Guo, Hongyuan Zhu, Jingkang Yang, Penghao Wu, Ruiqi Wang, Shulin Tian, Xiuying Wang, Yuhao Dong, Ziwei Liu","submitted_at":"2025-06-16T16:17:08Z","abstract_excerpt":"We introduce Ego-R1, a novel framework for reasoning over ultra-long (i.e., in days and weeks) egocentric videos, which leverages a structured Chain-of-Tool-Thought (CoTT) process, orchestrated by an Ego-R1 Agent trained via reinforcement learning (RL). Inspired by human problem-solving strategies, CoTT decomposes complex reasoning into modular steps, with the RL agent invoking specific tools, one per step, to iteratively and collaboratively answer sub-questions tackling such tasks as temporal retrieval and multi-modal understanding. We design a two-stage training paradigm involving supervised"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.13654","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/2506.13654/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":"2506.13654","created_at":"2026-07-05T11:22:24.372873+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.13654v1","created_at":"2026-07-05T11:22:24.372873+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.13654","created_at":"2026-07-05T11:22:24.372873+00:00"},{"alias_kind":"pith_short_12","alias_value":"A5DAUK3U3OYF","created_at":"2026-07-05T11:22:24.372873+00:00"},{"alias_kind":"pith_short_16","alias_value":"A5DAUK3U3OYFWU2Q","created_at":"2026-07-05T11:22:24.372873+00:00"},{"alias_kind":"pith_short_8","alias_value":"A5DAUK3U","created_at":"2026-07-05T11:22:24.372873+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07433","citing_title":"Watch, Remember, Reason: Human-View Video Understanding with MLLMs","ref_index":240,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02482","citing_title":"X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02482","citing_title":"X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22907","citing_title":"VideoOdyssey: A Benchmark for Ultra-Long-Context and Omni-Modal Video Understanding","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2511.20785","citing_title":"LongVT: Incentivizing \"Thinking with Long Videos\" via Native Tool Calling","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16079","citing_title":"VideoSeeker: Incentivizing Instance-level Video Understanding via Native Agentic Tool Invocation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2512.03438","citing_title":"Multimodal Reinforcement Learning with Adaptive Verifier for AI Agents","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2602.20913","citing_title":"LongVideo-R1: Smart Navigation for Low-cost Long Video Understanding","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13228","citing_title":"ReTool-Video: Recursive Tool-Using Video Agents with Meta-Augmented Tool Grounding","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05015","citing_title":"Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14520","citing_title":"Chain of Modality: From Static Fusion to Dynamic Orchestration in Omni-MLLMs","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62","json":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62.json","graph_json":"https://pith.science/api/pith-number/A5DAUK3U3OYFWU2QGULP33AY62/graph.json","events_json":"https://pith.science/api/pith-number/A5DAUK3U3OYFWU2QGULP33AY62/events.json","paper":"https://pith.science/paper/A5DAUK3U"},"agent_actions":{"view_html":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62","download_json":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62.json","view_paper":"https://pith.science/paper/A5DAUK3U","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.13654&json=true","fetch_graph":"https://pith.science/api/pith-number/A5DAUK3U3OYFWU2QGULP33AY62/graph.json","fetch_events":"https://pith.science/api/pith-number/A5DAUK3U3OYFWU2QGULP33AY62/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62/action/storage_attestation","attest_author":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62/action/author_attestation","sign_citation":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62/action/citation_signature","submit_replication":"https://pith.science/pith/A5DAUK3U3OYFWU2QGULP33AY62/action/replication_record"}},"created_at":"2026-07-05T11:22:24.372873+00:00","updated_at":"2026-07-05T11:22:24.372873+00:00"}