{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O5FPJH7I2B4HNHMA5SHOEQKEEE","short_pith_number":"pith:O5FPJH7I","schema_version":"1.0","canonical_sha256":"774af49fe8d078769d80ec8ee24144212348484984fc484b38ae6301fddf0def","source":{"kind":"arxiv","id":"2602.22779","version":3},"attestation_state":"computed","paper":{"title":"TrajTok: Learning Trajectory Tokens enables better Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"TrajTok learns trajectory tokens end-to-end through implicit space-time clustering to improve video model accuracy and efficiency.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ashutosh Kumar, Chenhao Zheng, Chun-Liang Li, Jianing Zhang, Jieyu Zhang, Oncel Tuzel, Quan Kong, Ranjay Krishna, Weikai Huang","submitted_at":"2026-02-26T09:15:34Z","abstract_excerpt":"Tokenization in video models, typically through patchification, generates an excessive and redundant number of tokens. This severely limits video efficiency and scalability. While recent trajectory-based tokenizers offer a promising solution by decoupling video duration from token count, they rely on complex external segmentation and tracking pipelines that are slow and task-agnostic. We propose TrajTok, an end-to-end video tokenizer module that is fully integrated and co-trained with video models for a downstream objective, dynamically adapting its token granularity to semantic complexity, in"},"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":true},"canonical_record":{"source":{"id":"2602.22779","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-02-26T09:15:34Z","cross_cats_sorted":[],"title_canon_sha256":"72079029b17a646fb5973a5cdd367d3c100b454490f0e26fa12335cd158e1b76","abstract_canon_sha256":"7c063636c3c98fd927b376d4a70fb97379984d7bda9b537f2c7cdd773ad63a2d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:08:46.330549Z","signature_b64":"Uhmyq2YOnoggC/IXu9YZH70ZoAbj8kMesSSA3z2vgMH5EawmNeFAOnv5/i5o+2zFOpZCkCscXh02ftJ6tQCDCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"774af49fe8d078769d80ec8ee24144212348484984fc484b38ae6301fddf0def","last_reissued_at":"2026-06-04T01:08:46.329945Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:08:46.329945Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TrajTok: Learning Trajectory Tokens enables better Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"TrajTok learns trajectory tokens end-to-end through implicit space-time clustering to improve video model accuracy and efficiency.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ashutosh Kumar, Chenhao Zheng, Chun-Liang Li, Jianing Zhang, Jieyu Zhang, Oncel Tuzel, Quan Kong, Ranjay Krishna, Weikai Huang","submitted_at":"2026-02-26T09:15:34Z","abstract_excerpt":"Tokenization in video models, typically through patchification, generates an excessive and redundant number of tokens. This severely limits video efficiency and scalability. While recent trajectory-based tokenizers offer a promising solution by decoupling video duration from token count, they rely on complex external segmentation and tracking pipelines that are slow and task-agnostic. We propose TrajTok, an end-to-end video tokenizer module that is fully integrated and co-trained with video models for a downstream objective, dynamically adapting its token granularity to semantic complexity, in"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"With TrajTok, we implement a video CLIP model trained from scratch (TrajViT2). It achieves the best accuracy at scale across both classification and retrieval benchmarks, while maintaining efficiency comparable to the best token-merging methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That implicit clustering of pixels in space and time will produce trajectories that are semantically useful for downstream video understanding tasks when the segmenter is co-trained only for adaptability rather than pixel-level fidelity.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TrajTok learns adaptive trajectory tokens for videos through a unified end-to-end segmenter, improving understanding performance and efficiency over patch-based or external-pipeline tokenizers.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"TrajTok learns trajectory tokens end-to-end through implicit space-time clustering to improve video model accuracy and efficiency.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fb3d03326b4bab152b6eb6a052f46efbb6b21779496465b5da1faf5b21816447"},"source":{"id":"2602.22779","kind":"arxiv","version":3},"verdict":{"id":"36578a04-2a7a-4bf2-a3dc-589019cfb825","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T19:09:34.106646Z","strongest_claim":"With TrajTok, we implement a video CLIP model trained from scratch (TrajViT2). It achieves the best accuracy at scale across both classification and retrieval benchmarks, while maintaining efficiency comparable to the best token-merging methods.","one_line_summary":"TrajTok learns adaptive trajectory tokens for videos through a unified end-to-end segmenter, improving understanding performance and efficiency over patch-based or external-pipeline tokenizers.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That implicit clustering of pixels in space and time will produce trajectories that are semantically useful for downstream video understanding tasks when the segmenter is co-trained only for adaptability rather than pixel-level fidelity.","pith_extraction_headline":"TrajTok learns trajectory tokens end-to-end through implicit space-time clustering to improve video model accuracy and efficiency."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2602.22779/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":2,"snapshot_sha256":"4ce750f8c498ee634682461a650164d2639b67f339e7b1fc3a5b36892539e714"},"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":"2602.22779","created_at":"2026-06-04T01:08:46.330011+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.22779v3","created_at":"2026-06-04T01:08:46.330011+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.22779","created_at":"2026-06-04T01:08:46.330011+00:00"},{"alias_kind":"pith_short_12","alias_value":"O5FPJH7I2B4H","created_at":"2026-06-04T01:08:46.330011+00:00"},{"alias_kind":"pith_short_16","alias_value":"O5FPJH7I2B4HNHMA","created_at":"2026-06-04T01:08:46.330011+00:00"},{"alias_kind":"pith_short_8","alias_value":"O5FPJH7I","created_at":"2026-06-04T01:08:46.330011+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE","json":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE.json","graph_json":"https://pith.science/api/pith-number/O5FPJH7I2B4HNHMA5SHOEQKEEE/graph.json","events_json":"https://pith.science/api/pith-number/O5FPJH7I2B4HNHMA5SHOEQKEEE/events.json","paper":"https://pith.science/paper/O5FPJH7I"},"agent_actions":{"view_html":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE","download_json":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE.json","view_paper":"https://pith.science/paper/O5FPJH7I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.22779&json=true","fetch_graph":"https://pith.science/api/pith-number/O5FPJH7I2B4HNHMA5SHOEQKEEE/graph.json","fetch_events":"https://pith.science/api/pith-number/O5FPJH7I2B4HNHMA5SHOEQKEEE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE/action/storage_attestation","attest_author":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE/action/author_attestation","sign_citation":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE/action/citation_signature","submit_replication":"https://pith.science/pith/O5FPJH7I2B4HNHMA5SHOEQKEEE/action/replication_record"}},"created_at":"2026-06-04T01:08:46.330011+00:00","updated_at":"2026-06-04T01:08:46.330011+00:00"}