{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MEPPTSMUZPMAVTVPTBCSHGJLEV","short_pith_number":"pith:MEPPTSMU","schema_version":"1.0","canonical_sha256":"611ef9c994cbd80aceaf984523992b256a091152d2cff49be5c5594f93f190a6","source":{"kind":"arxiv","id":"2402.04239","version":1},"attestation_state":"computed","paper":{"title":"CAST: Clustering Self-Attention using Surrogate Tokens for Efficient Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adjorn van Engelenhoven, Estefan\\'ia Talavera, Nicola Strisciuglio","submitted_at":"2024-02-06T18:47:52Z","abstract_excerpt":"The Transformer architecture has shown to be a powerful tool for a wide range of tasks. It is based on the self-attention mechanism, which is an inherently computationally expensive operation with quadratic computational complexity: memory usage and compute time increase quadratically with the length of the input sequences, thus limiting the application of Transformers. In this work, we propose a novel Clustering self-Attention mechanism using Surrogate Tokens (CAST), to optimize the attention computation and achieve efficient transformers. CAST utilizes learnable surrogate tokens to construct"},"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":"2402.04239","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-06T18:47:52Z","cross_cats_sorted":[],"title_canon_sha256":"cd96f4b69cb15dc6aae42f81728bac692f4187efdeca532e4c925797a73bbf81","abstract_canon_sha256":"5fd2791f24a3f4aace2ab3a2f7c8eb0b90630500c08ce992e0b3869e983d273d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:42:05.287855Z","signature_b64":"hmZXWuJe26+9yoobRztEho2nIGWsKQwx9KHt2KEPLjvrLTRstUATzbUY9mM4mJ6fFVnG5DQs+T1r+l1zp+9uDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"611ef9c994cbd80aceaf984523992b256a091152d2cff49be5c5594f93f190a6","last_reissued_at":"2026-07-05T07:42:05.287360Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:42:05.287360Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CAST: Clustering Self-Attention using Surrogate Tokens for Efficient Transformers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Adjorn van Engelenhoven, Estefan\\'ia Talavera, Nicola Strisciuglio","submitted_at":"2024-02-06T18:47:52Z","abstract_excerpt":"The Transformer architecture has shown to be a powerful tool for a wide range of tasks. It is based on the self-attention mechanism, which is an inherently computationally expensive operation with quadratic computational complexity: memory usage and compute time increase quadratically with the length of the input sequences, thus limiting the application of Transformers. In this work, we propose a novel Clustering self-Attention mechanism using Surrogate Tokens (CAST), to optimize the attention computation and achieve efficient transformers. CAST utilizes learnable surrogate tokens to construct"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.04239","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/2402.04239/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":"2402.04239","created_at":"2026-07-05T07:42:05.287421+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.04239v1","created_at":"2026-07-05T07:42:05.287421+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.04239","created_at":"2026-07-05T07:42:05.287421+00:00"},{"alias_kind":"pith_short_12","alias_value":"MEPPTSMUZPMA","created_at":"2026-07-05T07:42:05.287421+00:00"},{"alias_kind":"pith_short_16","alias_value":"MEPPTSMUZPMAVTVP","created_at":"2026-07-05T07:42:05.287421+00:00"},{"alias_kind":"pith_short_8","alias_value":"MEPPTSMU","created_at":"2026-07-05T07:42:05.287421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.19755","citing_title":"EGA-V1: Unifying Online Advertising with End-to-End Learning","ref_index":24,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV","json":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV.json","graph_json":"https://pith.science/api/pith-number/MEPPTSMUZPMAVTVPTBCSHGJLEV/graph.json","events_json":"https://pith.science/api/pith-number/MEPPTSMUZPMAVTVPTBCSHGJLEV/events.json","paper":"https://pith.science/paper/MEPPTSMU"},"agent_actions":{"view_html":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV","download_json":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV.json","view_paper":"https://pith.science/paper/MEPPTSMU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.04239&json=true","fetch_graph":"https://pith.science/api/pith-number/MEPPTSMUZPMAVTVPTBCSHGJLEV/graph.json","fetch_events":"https://pith.science/api/pith-number/MEPPTSMUZPMAVTVPTBCSHGJLEV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV/action/storage_attestation","attest_author":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV/action/author_attestation","sign_citation":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV/action/citation_signature","submit_replication":"https://pith.science/pith/MEPPTSMUZPMAVTVPTBCSHGJLEV/action/replication_record"}},"created_at":"2026-07-05T07:42:05.287421+00:00","updated_at":"2026-07-05T07:42:05.287421+00:00"}