{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Y72K42YQRAOW6D3SGIMCTGSXZ2","short_pith_number":"pith:Y72K42YQ","schema_version":"1.0","canonical_sha256":"c7f4ae6b10881d6f0f723218299a57ceae138457142f0ddcf1eeb256a1d819cd","source":{"kind":"arxiv","id":"2504.05579","version":2},"attestation_state":"computed","paper":{"title":"TAPNext: Tracking Any Point (TAP) as Next Token Prediction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Artem Zholus, Carl Doersch, Ignacio Rocco, Mehdi S. M. Sajjadi, Ross Goroshin, Sarath Chandar, Skanda Koppula, Viorica Patraucean, Xu Owen He, Yi Yang","submitted_at":"2025-04-08T00:28:42Z","abstract_excerpt":"Tracking Any Point (TAP) in a video is a challenging computer vision problem with many demonstrated applications in robotics, video editing, and 3D reconstruction. Existing methods for TAP rely heavily on complex tracking-specific inductive biases and heuristics, limiting their generality and potential for scaling. To address these challenges, we present TAPNext, a new approach that casts TAP as sequential masked token decoding. Our model is causal, tracks in a purely online fashion, and removes tracking-specific inductive biases. This enables TAPNext to run with minimal latency, and removes t"},"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":"2504.05579","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-08T00:28:42Z","cross_cats_sorted":[],"title_canon_sha256":"13417c4ccd331166ee1b7f9251ccd1966394503156052449c783c6731f15ec99","abstract_canon_sha256":"b71050752ad5733a28828cf19ab19e478556b5bb702ad38144fd1cd3dc206e46"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:48:30.406726Z","signature_b64":"0n2r5z5QlavM9mN0K/loLVL4Z/Ms0Bc1M82WyUMIPH1XWSNJKrxknnlP8IKy8OvYwfCaMATbuKWSiy/fKWaPBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c7f4ae6b10881d6f0f723218299a57ceae138457142f0ddcf1eeb256a1d819cd","last_reissued_at":"2026-07-05T10:48:30.406231Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:48:30.406231Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TAPNext: Tracking Any Point (TAP) as Next Token Prediction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Artem Zholus, Carl Doersch, Ignacio Rocco, Mehdi S. M. Sajjadi, Ross Goroshin, Sarath Chandar, Skanda Koppula, Viorica Patraucean, Xu Owen He, Yi Yang","submitted_at":"2025-04-08T00:28:42Z","abstract_excerpt":"Tracking Any Point (TAP) in a video is a challenging computer vision problem with many demonstrated applications in robotics, video editing, and 3D reconstruction. Existing methods for TAP rely heavily on complex tracking-specific inductive biases and heuristics, limiting their generality and potential for scaling. To address these challenges, we present TAPNext, a new approach that casts TAP as sequential masked token decoding. Our model is causal, tracks in a purely online fashion, and removes tracking-specific inductive biases. This enables TAPNext to run with minimal latency, and removes t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.05579","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/2504.05579/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":"2504.05579","created_at":"2026-07-05T10:48:30.406290+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.05579v2","created_at":"2026-07-05T10:48:30.406290+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.05579","created_at":"2026-07-05T10:48:30.406290+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y72K42YQRAOW","created_at":"2026-07-05T10:48:30.406290+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y72K42YQRAOW6D3S","created_at":"2026-07-05T10:48:30.406290+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y72K42YQ","created_at":"2026-07-05T10:48:30.406290+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.30174","citing_title":"LiveSVG: Zero-Shot SVG Animation via Video Generation","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11737","citing_title":"Learning Long-term Motion Embeddings for Efficient Kinematics Generation","ref_index":54,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2","json":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2.json","graph_json":"https://pith.science/api/pith-number/Y72K42YQRAOW6D3SGIMCTGSXZ2/graph.json","events_json":"https://pith.science/api/pith-number/Y72K42YQRAOW6D3SGIMCTGSXZ2/events.json","paper":"https://pith.science/paper/Y72K42YQ"},"agent_actions":{"view_html":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2","download_json":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2.json","view_paper":"https://pith.science/paper/Y72K42YQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.05579&json=true","fetch_graph":"https://pith.science/api/pith-number/Y72K42YQRAOW6D3SGIMCTGSXZ2/graph.json","fetch_events":"https://pith.science/api/pith-number/Y72K42YQRAOW6D3SGIMCTGSXZ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2/action/storage_attestation","attest_author":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2/action/author_attestation","sign_citation":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2/action/citation_signature","submit_replication":"https://pith.science/pith/Y72K42YQRAOW6D3SGIMCTGSXZ2/action/replication_record"}},"created_at":"2026-07-05T10:48:30.406290+00:00","updated_at":"2026-07-05T10:48:30.406290+00:00"}