{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:KKUJ66P5TYD7PMEDBHFMBG6B7N","short_pith_number":"pith:KKUJ66P5","schema_version":"1.0","canonical_sha256":"52a89f79fd9e07f7b08309cac09bc1fb4febe2f25eb0b8d3f9d7a4a05dd01210","source":{"kind":"arxiv","id":"2012.15045","version":2},"attestation_state":"computed","paper":{"title":"Reservoir Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexei Baevski, Ari S. Morcos, Douwe Kiela, Kurt Keutzer, Michael Auli, Sheng Shen","submitted_at":"2020-12-30T05:20:16Z","abstract_excerpt":"We demonstrate that transformers obtain impressive performance even when some of the layers are randomly initialized and never updated. Inspired by old and well-established ideas in machine learning, we explore a variety of non-linear \"reservoir\" layers interspersed with regular transformer layers, and show improvements in wall-clock compute time until convergence, as well as overall performance, on various machine translation and (masked) language modelling tasks."},"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":"2012.15045","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-12-30T05:20:16Z","cross_cats_sorted":[],"title_canon_sha256":"3565e6fd5b7722abecc950cc878cc664d4eb729a21402f0104af6845c90c3278","abstract_canon_sha256":"2553c51047b54f6232fcfcec819ebe196a6259127689e59b9999813efa12876d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:45:35.451817Z","signature_b64":"198R0JOEax1g236X6nhPoZSp8THhs+KAqyfUpVoSuNMEqAYMoEWX1sI9TgnSyK1VUQBPU7qqUHxNNmxmBRgcAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"52a89f79fd9e07f7b08309cac09bc1fb4febe2f25eb0b8d3f9d7a4a05dd01210","last_reissued_at":"2026-07-05T02:45:35.451368Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:45:35.451368Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reservoir Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexei Baevski, Ari S. Morcos, Douwe Kiela, Kurt Keutzer, Michael Auli, Sheng Shen","submitted_at":"2020-12-30T05:20:16Z","abstract_excerpt":"We demonstrate that transformers obtain impressive performance even when some of the layers are randomly initialized and never updated. Inspired by old and well-established ideas in machine learning, we explore a variety of non-linear \"reservoir\" layers interspersed with regular transformer layers, and show improvements in wall-clock compute time until convergence, as well as overall performance, on various machine translation and (masked) language modelling tasks."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.15045","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/2012.15045/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":"2012.15045","created_at":"2026-07-05T02:45:35.451425+00:00"},{"alias_kind":"arxiv_version","alias_value":"2012.15045v2","created_at":"2026-07-05T02:45:35.451425+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.15045","created_at":"2026-07-05T02:45:35.451425+00:00"},{"alias_kind":"pith_short_12","alias_value":"KKUJ66P5TYD7","created_at":"2026-07-05T02:45:35.451425+00:00"},{"alias_kind":"pith_short_16","alias_value":"KKUJ66P5TYD7PMED","created_at":"2026-07-05T02:45:35.451425+00:00"},{"alias_kind":"pith_short_8","alias_value":"KKUJ66P5","created_at":"2026-07-05T02:45:35.451425+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.18552","citing_title":"Quantum Reservoir Computing: Recent Advances and Future Directions","ref_index":2020,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N","json":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N.json","graph_json":"https://pith.science/api/pith-number/KKUJ66P5TYD7PMEDBHFMBG6B7N/graph.json","events_json":"https://pith.science/api/pith-number/KKUJ66P5TYD7PMEDBHFMBG6B7N/events.json","paper":"https://pith.science/paper/KKUJ66P5"},"agent_actions":{"view_html":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N","download_json":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N.json","view_paper":"https://pith.science/paper/KKUJ66P5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2012.15045&json=true","fetch_graph":"https://pith.science/api/pith-number/KKUJ66P5TYD7PMEDBHFMBG6B7N/graph.json","fetch_events":"https://pith.science/api/pith-number/KKUJ66P5TYD7PMEDBHFMBG6B7N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N/action/storage_attestation","attest_author":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N/action/author_attestation","sign_citation":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N/action/citation_signature","submit_replication":"https://pith.science/pith/KKUJ66P5TYD7PMEDBHFMBG6B7N/action/replication_record"}},"created_at":"2026-07-05T02:45:35.451425+00:00","updated_at":"2026-07-05T02:45:35.451425+00:00"}