{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:76UZAUQCBNUD3HF7D33JL6RM65","short_pith_number":"pith:76UZAUQC","schema_version":"1.0","canonical_sha256":"ffa99052020b683d9cbf1ef695fa2cf74c9ecfc2275569ad5ca05c94c818be1e","source":{"kind":"arxiv","id":"1905.07799","version":2},"attestation_state":"computed","paper":{"title":"Adaptive Attention Span in Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Armand Joulin, Edouard Grave, Piotr Bojanowski, Sainbayar Sukhbaatar","submitted_at":"2019-05-19T19:43:14Z","abstract_excerpt":"We propose a novel self-attention mechanism that can learn its optimal attention span. This allows us to extend significantly the maximum context size used in Transformer, while maintaining control over their memory footprint and computational time. We show the effectiveness of our approach on the task of character level language modeling, where we achieve state-of-the-art performances on text8 and enwiki8 by using a maximum context of 8k characters."},"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":"1905.07799","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-05-19T19:43:14Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"c0ffe795dae5c5a2b9f75d6c2d0f99c597494742dd19c3eeb6368247cc29ea7b","abstract_canon_sha256":"6bc2b9cdc8ff4eb80f866026780b247ae667b5daf6c6d1cb65159bedd2891745"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-04T23:52:21.409621Z","signature_b64":"raAguYEeWK8PQsjtggm7ds+toruF6b4THMeENZQ+Vaom7mzmUWu9BDIAY0mWAlKuZGB0uEPfXWS8Dl2ORzWKDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ffa99052020b683d9cbf1ef695fa2cf74c9ecfc2275569ad5ca05c94c818be1e","last_reissued_at":"2026-07-04T23:52:21.409200Z","signature_status":"signed_v1","first_computed_at":"2026-07-04T23:52:21.409200Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive Attention Span in Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Armand Joulin, Edouard Grave, Piotr Bojanowski, Sainbayar Sukhbaatar","submitted_at":"2019-05-19T19:43:14Z","abstract_excerpt":"We propose a novel self-attention mechanism that can learn its optimal attention span. This allows us to extend significantly the maximum context size used in Transformer, while maintaining control over their memory footprint and computational time. We show the effectiveness of our approach on the task of character level language modeling, where we achieve state-of-the-art performances on text8 and enwiki8 by using a maximum context of 8k characters."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.07799","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/1905.07799/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":"1905.07799","created_at":"2026-07-04T23:52:21.409253+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.07799v2","created_at":"2026-07-04T23:52:21.409253+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.07799","created_at":"2026-07-04T23:52:21.409253+00:00"},{"alias_kind":"pith_short_12","alias_value":"76UZAUQCBNUD","created_at":"2026-07-04T23:52:21.409253+00:00"},{"alias_kind":"pith_short_16","alias_value":"76UZAUQCBNUD3HF7","created_at":"2026-07-04T23:52:21.409253+00:00"},{"alias_kind":"pith_short_8","alias_value":"76UZAUQC","created_at":"2026-07-04T23:52:21.409253+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"1911.05507","citing_title":"Compressive Transformers for Long-Range Sequence Modelling","ref_index":121,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03263","citing_title":"LPC-SM: Local Predictive Coding and Sparse Memory for Long-Context Language Modeling","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2101.03961","citing_title":"Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2202.08906","citing_title":"ST-MoE: Designing Stable and Transferable Sparse Expert Models","ref_index":96,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10875","citing_title":"Compute Where it Counts: Self Optimizing Language Models","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65","json":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65.json","graph_json":"https://pith.science/api/pith-number/76UZAUQCBNUD3HF7D33JL6RM65/graph.json","events_json":"https://pith.science/api/pith-number/76UZAUQCBNUD3HF7D33JL6RM65/events.json","paper":"https://pith.science/paper/76UZAUQC"},"agent_actions":{"view_html":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65","download_json":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65.json","view_paper":"https://pith.science/paper/76UZAUQC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.07799&json=true","fetch_graph":"https://pith.science/api/pith-number/76UZAUQCBNUD3HF7D33JL6RM65/graph.json","fetch_events":"https://pith.science/api/pith-number/76UZAUQCBNUD3HF7D33JL6RM65/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65/action/timestamp_anchor","attest_storage":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65/action/storage_attestation","attest_author":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65/action/author_attestation","sign_citation":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65/action/citation_signature","submit_replication":"https://pith.science/pith/76UZAUQCBNUD3HF7D33JL6RM65/action/replication_record"}},"created_at":"2026-07-04T23:52:21.409253+00:00","updated_at":"2026-07-04T23:52:21.409253+00:00"}