{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:RL2ABPPAXKF7KZBEKKCGILMMEO","short_pith_number":"pith:RL2ABPPA","schema_version":"1.0","canonical_sha256":"8af400bde0ba8bf564245284642d8c23885a62500535df0188edf72514de603d","source":{"kind":"arxiv","id":"2109.10164","version":2},"attestation_state":"computed","paper":{"title":"RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abbas Ghaddar, Md Akmal Haidar, Mehdi Rezagholizadeh, Nithin Anchuri, Pascal Poupart, Philippe Langlais","submitted_at":"2021-09-21T13:21:13Z","abstract_excerpt":"Intermediate layer knowledge distillation (KD) can improve the standard KD technique (which only targets the output of teacher and student models) especially over large pre-trained language models. However, intermediate layer distillation suffers from excessive computational burdens and engineering efforts required for setting up a proper layer mapping. To address these problems, we propose a RAndom Intermediate Layer Knowledge Distillation (RAIL-KD) approach in which, intermediate layers from the teacher model are selected randomly to be distilled into the intermediate layers of the student m"},"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":"2109.10164","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-09-21T13:21:13Z","cross_cats_sorted":[],"title_canon_sha256":"692fa171fc922a3417dbbf27c04af1bb5288a015f36b864db7a0c8506698931f","abstract_canon_sha256":"e0882a2399f35e43eece65e4b71d9ea4ac9b1bba3408ce5296b08d45dae5f69b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:19:31.610394Z","signature_b64":"x2IT+LoXlYhKrkfwb8S4QrQM7oIKk/MjULil6FhMXOutsUaah+GM1eIB8+wvoSwz1aKiklWwxnI+uuv7flpgDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8af400bde0ba8bf564245284642d8c23885a62500535df0188edf72514de603d","last_reissued_at":"2026-07-05T03:19:31.609956Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:19:31.609956Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RAIL-KD: RAndom Intermediate Layer Mapping for Knowledge Distillation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abbas Ghaddar, Md Akmal Haidar, Mehdi Rezagholizadeh, Nithin Anchuri, Pascal Poupart, Philippe Langlais","submitted_at":"2021-09-21T13:21:13Z","abstract_excerpt":"Intermediate layer knowledge distillation (KD) can improve the standard KD technique (which only targets the output of teacher and student models) especially over large pre-trained language models. However, intermediate layer distillation suffers from excessive computational burdens and engineering efforts required for setting up a proper layer mapping. To address these problems, we propose a RAndom Intermediate Layer Knowledge Distillation (RAIL-KD) approach in which, intermediate layers from the teacher model are selected randomly to be distilled into the intermediate layers of the student m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.10164","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/2109.10164/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":"2109.10164","created_at":"2026-07-05T03:19:31.610003+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.10164v2","created_at":"2026-07-05T03:19:31.610003+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.10164","created_at":"2026-07-05T03:19:31.610003+00:00"},{"alias_kind":"pith_short_12","alias_value":"RL2ABPPAXKF7","created_at":"2026-07-05T03:19:31.610003+00:00"},{"alias_kind":"pith_short_16","alias_value":"RL2ABPPAXKF7KZBE","created_at":"2026-07-05T03:19:31.610003+00:00"},{"alias_kind":"pith_short_8","alias_value":"RL2ABPPA","created_at":"2026-07-05T03:19:31.610003+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/RL2ABPPAXKF7KZBEKKCGILMMEO","json":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO.json","graph_json":"https://pith.science/api/pith-number/RL2ABPPAXKF7KZBEKKCGILMMEO/graph.json","events_json":"https://pith.science/api/pith-number/RL2ABPPAXKF7KZBEKKCGILMMEO/events.json","paper":"https://pith.science/paper/RL2ABPPA"},"agent_actions":{"view_html":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO","download_json":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO.json","view_paper":"https://pith.science/paper/RL2ABPPA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.10164&json=true","fetch_graph":"https://pith.science/api/pith-number/RL2ABPPAXKF7KZBEKKCGILMMEO/graph.json","fetch_events":"https://pith.science/api/pith-number/RL2ABPPAXKF7KZBEKKCGILMMEO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO/action/storage_attestation","attest_author":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO/action/author_attestation","sign_citation":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO/action/citation_signature","submit_replication":"https://pith.science/pith/RL2ABPPAXKF7KZBEKKCGILMMEO/action/replication_record"}},"created_at":"2026-07-05T03:19:31.610003+00:00","updated_at":"2026-07-05T03:19:31.610003+00:00"}