{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:O3GZJ236CJL4B5JPPXZCSG2UKF","short_pith_number":"pith:O3GZJ236","schema_version":"1.0","canonical_sha256":"76cd94eb7e1257c0f52f7df2291b54514d4efb5ebaa32fbdf6742ea6a99db470","source":{"kind":"arxiv","id":"2512.16615","version":2},"attestation_state":"computed","paper":{"title":"Trainable Log-linear Sparse Attention for Efficient Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Shuai Yang, Tianyi Wei, Xingang Pan, Yifan Zhou, Zeqi Xiao","submitted_at":"2025-12-18T14:53:12Z","abstract_excerpt":"Diffusion Transformers (DiTs) set the state of the art in visual generation, yet their quadratic self-attention cost fundamentally limits scaling to long token sequences. Recent Top-K sparse attention approaches reduce the computation of DiTs by compressing tokens into block-wise representation and selecting a small set of relevant key blocks, but still suffer from (i) quadratic selection cost on compressed tokens and (ii) increasing K required to maintain model quality as sequences grow. We identify that their inefficiency is due to the single-level design, as a single coarse level is insuffi"},"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":"2512.16615","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-12-18T14:53:12Z","cross_cats_sorted":[],"title_canon_sha256":"84674cdd158f774eeda7e95be64af1a42b90ab391ddcc30b13dcd05809eea708","abstract_canon_sha256":"70bbe8e55113e71d6f0fb990d88f06512574f66fc6ff76fc4ab930da7ab105b1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T01:24:05.806937Z","signature_b64":"URYmGuanCoG29ah42cfYrmeAinkq53Us404msg3AaZFaz4s85K1PfCxqEFif9KrMDBOzigNx/zN4HZACmV6BDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"76cd94eb7e1257c0f52f7df2291b54514d4efb5ebaa32fbdf6742ea6a99db470","last_reissued_at":"2026-07-24T01:24:05.805939Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T01:24:05.805939Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Trainable Log-linear Sparse Attention for Efficient Diffusion Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Shuai Yang, Tianyi Wei, Xingang Pan, Yifan Zhou, Zeqi Xiao","submitted_at":"2025-12-18T14:53:12Z","abstract_excerpt":"Diffusion Transformers (DiTs) set the state of the art in visual generation, yet their quadratic self-attention cost fundamentally limits scaling to long token sequences. Recent Top-K sparse attention approaches reduce the computation of DiTs by compressing tokens into block-wise representation and selecting a small set of relevant key blocks, but still suffer from (i) quadratic selection cost on compressed tokens and (ii) increasing K required to maintain model quality as sequences grow. We identify that their inefficiency is due to the single-level design, as a single coarse level is insuffi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.16615","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/2512.16615/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":"2512.16615","created_at":"2026-07-24T01:24:05.806420+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.16615v2","created_at":"2026-07-24T01:24:05.806420+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.16615","created_at":"2026-07-24T01:24:05.806420+00:00"},{"alias_kind":"pith_short_12","alias_value":"O3GZJ236CJL4","created_at":"2026-07-24T01:24:05.806420+00:00"},{"alias_kind":"pith_short_16","alias_value":"O3GZJ236CJL4B5JP","created_at":"2026-07-24T01:24:05.806420+00:00"},{"alias_kind":"pith_short_8","alias_value":"O3GZJ236","created_at":"2026-07-24T01:24:05.806420+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/O3GZJ236CJL4B5JPPXZCSG2UKF","json":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF.json","graph_json":"https://pith.science/api/pith-number/O3GZJ236CJL4B5JPPXZCSG2UKF/graph.json","events_json":"https://pith.science/api/pith-number/O3GZJ236CJL4B5JPPXZCSG2UKF/events.json","paper":"https://pith.science/paper/O3GZJ236"},"agent_actions":{"view_html":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF","download_json":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF.json","view_paper":"https://pith.science/paper/O3GZJ236","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.16615&json=true","fetch_graph":"https://pith.science/api/pith-number/O3GZJ236CJL4B5JPPXZCSG2UKF/graph.json","fetch_events":"https://pith.science/api/pith-number/O3GZJ236CJL4B5JPPXZCSG2UKF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF/action/storage_attestation","attest_author":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF/action/author_attestation","sign_citation":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF/action/citation_signature","submit_replication":"https://pith.science/pith/O3GZJ236CJL4B5JPPXZCSG2UKF/action/replication_record"}},"created_at":"2026-07-24T01:24:05.806420+00:00","updated_at":"2026-07-24T01:24:05.806420+00:00"}