{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:HWHET3WCY2BQM3KCTMBHVUVFH6","short_pith_number":"pith:HWHET3WC","schema_version":"1.0","canonical_sha256":"3d8e49eec2c683066d429b027ad2a53f95b4f4cb206031937fa08d54e05b1518","source":{"kind":"arxiv","id":"2109.00301","version":3},"attestation_state":"computed","paper":{"title":"$\\infty$-former: Infinite Memory Transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'e F. T. Martins, Pedro Henrique Martins, Zita Marinho","submitted_at":"2021-09-01T10:51:58Z","abstract_excerpt":"Transformers are unable to model long-term memories effectively, since the amount of computation they need to perform grows with the context length. While variations of efficient transformers have been proposed, they all have a finite memory capacity and are forced to drop old information. In this paper, we propose the $\\infty$-former, which extends the vanilla transformer with an unbounded long-term memory. By making use of a continuous-space attention mechanism to attend over the long-term memory, the $\\infty$-former's attention complexity becomes independent of the context length, trading o"},"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":"2109.00301","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-01T10:51:58Z","cross_cats_sorted":[],"title_canon_sha256":"cad74018e8442cebf8fe58d194c276a1e371c69b237a9732499f2975b460d2e9","abstract_canon_sha256":"d72afb608f6c2b9b8b3666e32ea649093542632072d9be5654d236cc924e0abf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:08:25.337748Z","signature_b64":"+WgDCB445bQ2uUOC4/A19Bg1nyf0jsbG8DwMPrmlgEKsVBHfq00fDd25iEEmf/CyJEUqkfRhxevl2YtPUx7XAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3d8e49eec2c683066d429b027ad2a53f95b4f4cb206031937fa08d54e05b1518","last_reissued_at":"2026-07-05T04:08:25.337244Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:08:25.337244Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"$\\infty$-former: Infinite Memory Transformer","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'e F. T. Martins, Pedro Henrique Martins, Zita Marinho","submitted_at":"2021-09-01T10:51:58Z","abstract_excerpt":"Transformers are unable to model long-term memories effectively, since the amount of computation they need to perform grows with the context length. While variations of efficient transformers have been proposed, they all have a finite memory capacity and are forced to drop old information. In this paper, we propose the $\\infty$-former, which extends the vanilla transformer with an unbounded long-term memory. By making use of a continuous-space attention mechanism to attend over the long-term memory, the $\\infty$-former's attention complexity becomes independent of the context length, trading o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.00301","kind":"arxiv","version":3},"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/2109.00301/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":"2109.00301","created_at":"2026-07-05T04:08:25.337306+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.00301v3","created_at":"2026-07-05T04:08:25.337306+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.00301","created_at":"2026-07-05T04:08:25.337306+00:00"},{"alias_kind":"pith_short_12","alias_value":"HWHET3WCY2BQ","created_at":"2026-07-05T04:08:25.337306+00:00"},{"alias_kind":"pith_short_16","alias_value":"HWHET3WCY2BQM3KC","created_at":"2026-07-05T04:08:25.337306+00:00"},{"alias_kind":"pith_short_8","alias_value":"HWHET3WC","created_at":"2026-07-05T04:08:25.337306+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.02259","citing_title":"MemAgent: Reshaping Long-Context LLM with Multi-Conv RL-based Memory Agent","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6","json":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6.json","graph_json":"https://pith.science/api/pith-number/HWHET3WCY2BQM3KCTMBHVUVFH6/graph.json","events_json":"https://pith.science/api/pith-number/HWHET3WCY2BQM3KCTMBHVUVFH6/events.json","paper":"https://pith.science/paper/HWHET3WC"},"agent_actions":{"view_html":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6","download_json":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6.json","view_paper":"https://pith.science/paper/HWHET3WC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.00301&json=true","fetch_graph":"https://pith.science/api/pith-number/HWHET3WCY2BQM3KCTMBHVUVFH6/graph.json","fetch_events":"https://pith.science/api/pith-number/HWHET3WCY2BQM3KCTMBHVUVFH6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6/action/storage_attestation","attest_author":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6/action/author_attestation","sign_citation":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6/action/citation_signature","submit_replication":"https://pith.science/pith/HWHET3WCY2BQM3KCTMBHVUVFH6/action/replication_record"}},"created_at":"2026-07-05T04:08:25.337306+00:00","updated_at":"2026-07-05T04:08:25.337306+00:00"}