{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LMFIHSXG7HIS2GI2DLPKA3FXYS","short_pith_number":"pith:LMFIHSXG","schema_version":"1.0","canonical_sha256":"5b0a83cae6f9d12d191a1adea06cb7c49b7eea5e6e194069faee52cc668a5528","source":{"kind":"arxiv","id":"2607.03012","version":1},"attestation_state":"computed","paper":{"title":"HyperVAttention: Efficient Sparse Attention with Spatio-Temporal Clustering for Video Diffusion","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Amir Zandieh, Dongyeun Lee, Insu Han, Junmo Kim, Vahab Mirrokni","submitted_at":"2026-07-03T06:46:35Z","abstract_excerpt":"Video Diffusion Transformers (VDiTs) have demonstrated significant capabilities in high-fidelity video generation. However, their ability to produce long-duration videos is fundamentally constrained by the quadratic complexity of the self-attention mechanism. Recent clustering-based sparse attention methods improve the quality-speed trade-off by grouping semantically similar tokens, but their practical efficiency remains limited by two bottlenecks: substantial clustering overhead and low CTA utilization caused by irregular cluster-induced blocks. We propose HyperVAttention (HVA), a training-fr"},"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":"2607.03012","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-07-03T06:46:35Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"7cd115337e71c1c8afd28728d29f93c97320affa9fb872e5ce9b7e74ce8f1e2b","abstract_canon_sha256":"6be278b2f7203efc6a324f9c3fd861fe7a186af61e04a5244a31e5ee8b1010ad"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T01:16:39.645697Z","signature_b64":"5Ilv/X4uS39EzTVS/ex2dFMMmfwJ2t/ARd3965Wct+SsgyrYwx3BdkoTMsJIrjGVUojAUqAzPTGDMGmUDru9DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5b0a83cae6f9d12d191a1adea06cb7c49b7eea5e6e194069faee52cc668a5528","last_reissued_at":"2026-07-07T01:16:39.645240Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T01:16:39.645240Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HyperVAttention: Efficient Sparse Attention with Spatio-Temporal Clustering for Video Diffusion","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Amir Zandieh, Dongyeun Lee, Insu Han, Junmo Kim, Vahab Mirrokni","submitted_at":"2026-07-03T06:46:35Z","abstract_excerpt":"Video Diffusion Transformers (VDiTs) have demonstrated significant capabilities in high-fidelity video generation. However, their ability to produce long-duration videos is fundamentally constrained by the quadratic complexity of the self-attention mechanism. Recent clustering-based sparse attention methods improve the quality-speed trade-off by grouping semantically similar tokens, but their practical efficiency remains limited by two bottlenecks: substantial clustering overhead and low CTA utilization caused by irregular cluster-induced blocks. We propose HyperVAttention (HVA), a training-fr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.03012","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/2607.03012/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":"2607.03012","created_at":"2026-07-07T01:16:39.645305+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.03012v1","created_at":"2026-07-07T01:16:39.645305+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.03012","created_at":"2026-07-07T01:16:39.645305+00:00"},{"alias_kind":"pith_short_12","alias_value":"LMFIHSXG7HIS","created_at":"2026-07-07T01:16:39.645305+00:00"},{"alias_kind":"pith_short_16","alias_value":"LMFIHSXG7HIS2GI2","created_at":"2026-07-07T01:16:39.645305+00:00"},{"alias_kind":"pith_short_8","alias_value":"LMFIHSXG","created_at":"2026-07-07T01:16:39.645305+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/LMFIHSXG7HIS2GI2DLPKA3FXYS","json":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS.json","graph_json":"https://pith.science/api/pith-number/LMFIHSXG7HIS2GI2DLPKA3FXYS/graph.json","events_json":"https://pith.science/api/pith-number/LMFIHSXG7HIS2GI2DLPKA3FXYS/events.json","paper":"https://pith.science/paper/LMFIHSXG"},"agent_actions":{"view_html":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS","download_json":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS.json","view_paper":"https://pith.science/paper/LMFIHSXG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.03012&json=true","fetch_graph":"https://pith.science/api/pith-number/LMFIHSXG7HIS2GI2DLPKA3FXYS/graph.json","fetch_events":"https://pith.science/api/pith-number/LMFIHSXG7HIS2GI2DLPKA3FXYS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS/action/storage_attestation","attest_author":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS/action/author_attestation","sign_citation":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS/action/citation_signature","submit_replication":"https://pith.science/pith/LMFIHSXG7HIS2GI2DLPKA3FXYS/action/replication_record"}},"created_at":"2026-07-07T01:16:39.645305+00:00","updated_at":"2026-07-07T01:16:39.645305+00:00"}