{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2019:T7IKL3GLAXLBJOYXKZMF3MDWHY","short_pith_number":"pith:T7IKL3GL","canonical_record":{"source":{"id":"1907.04829","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-07-10T17:14:47Z","cross_cats_sorted":[],"title_canon_sha256":"62e94252249c83973ddb886d192787ba0c94db747de22add975161a39a53b7f9","abstract_canon_sha256":"5caf51a454972cc2c521357ef1c8c17c63cac92de4edea6257949f2243781e41"},"schema_version":"1.0"},"canonical_sha256":"9fd0a5eccb05d614bb1756585db0763e3a6c805b3115a552673dc43cafa8d0c7","source":{"kind":"arxiv","id":"1907.04829","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1907.04829","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"arxiv_version","alias_value":"1907.04829v1","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1907.04829","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"pith_short_12","alias_value":"T7IKL3GLAXLB","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_16","alias_value":"T7IKL3GLAXLBJOYX","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_8","alias_value":"T7IKL3GL","created_at":"2026-05-18T12:33:27Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2019:T7IKL3GLAXLBJOYXKZMF3MDWHY","target":"record","payload":{"canonical_record":{"source":{"id":"1907.04829","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-07-10T17:14:47Z","cross_cats_sorted":[],"title_canon_sha256":"62e94252249c83973ddb886d192787ba0c94db747de22add975161a39a53b7f9","abstract_canon_sha256":"5caf51a454972cc2c521357ef1c8c17c63cac92de4edea6257949f2243781e41"},"schema_version":"1.0"},"canonical_sha256":"9fd0a5eccb05d614bb1756585db0763e3a6c805b3115a552673dc43cafa8d0c7","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:40:56.570042Z","signature_b64":"PrPIx31YPJbbX1T7pm7XI4w291IYXpTGrw7lAYngoX/IroRog270UmyYpH0bxY99rM0j2COKTX22m0rlK5x+Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9fd0a5eccb05d614bb1756585db0763e3a6c805b3115a552673dc43cafa8d0c7","last_reissued_at":"2026-05-17T23:40:56.569407Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:40:56.569407Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1907.04829","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:40:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BvbvzU2AyPurrQOzg2snqf2pOhv0ZwbJWsoinUJgi/Bp9/nlzFD/KG/zHJ/IMTEwthkG8zJ6VBBAxr5PoBRCDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T16:30:07.485283Z"},"content_sha256":"2fc58fcf9e83c974fee3df99e18e1bbb36429ff30afabbc2ddbdc7d32af97a2c","schema_version":"1.0","event_id":"sha256:2fc58fcf9e83c974fee3df99e18e1bbb36429ff30afabbc2ddbdc7d32af97a2c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2019:T7IKL3GLAXLBJOYXKZMF3MDWHY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"BAM! Born-Again Multi-Task Networks for Natural Language Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christopher D. Manning, Kevin Clark, Minh-Thang Luong, Quoc V. Le, Urvashi Khandelwal","submitted_at":"2019-07-10T17:14:47Z","abstract_excerpt":"It can be challenging to train multi-task neural networks that outperform or even match their single-task counterparts. To help address this, we propose using knowledge distillation where single-task models teach a multi-task model. We enhance this training with teacher annealing, a novel method that gradually transitions the model from distillation to supervised learning, helping the multi-task model surpass its single-task teachers. We evaluate our approach by multi-task fine-tuning BERT on the GLUE benchmark. Our method consistently improves over standard single-task and multi-task training"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1907.04829","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":""},"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:40:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SXyb0/p8L/PbxfgWHHgW9PtugDK/aw2dZZJNVPViMJObN3GfbiCuAG0iX/iqDEm6NvLAuPgzdi0S7rVlsnyJCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T16:30:07.485648Z"},"content_sha256":"54188db03a6e4bc4727cf9528b527cf2b6f280565591a43dc834c8017c208c9e","schema_version":"1.0","event_id":"sha256:54188db03a6e4bc4727cf9528b527cf2b6f280565591a43dc834c8017c208c9e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/bundle.json","state_url":"https://pith.science/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-30T16:30:07Z","links":{"resolver":"https://pith.science/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY","bundle":"https://pith.science/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/bundle.json","state":"https://pith.science/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/T7IKL3GLAXLBJOYXKZMF3MDWHY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2019:T7IKL3GLAXLBJOYXKZMF3MDWHY","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5caf51a454972cc2c521357ef1c8c17c63cac92de4edea6257949f2243781e41","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-07-10T17:14:47Z","title_canon_sha256":"62e94252249c83973ddb886d192787ba0c94db747de22add975161a39a53b7f9"},"schema_version":"1.0","source":{"id":"1907.04829","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1907.04829","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"arxiv_version","alias_value":"1907.04829v1","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1907.04829","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"pith_short_12","alias_value":"T7IKL3GLAXLB","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_16","alias_value":"T7IKL3GLAXLBJOYX","created_at":"2026-05-18T12:33:27Z"},{"alias_kind":"pith_short_8","alias_value":"T7IKL3GL","created_at":"2026-05-18T12:33:27Z"}],"graph_snapshots":[{"event_id":"sha256:54188db03a6e4bc4727cf9528b527cf2b6f280565591a43dc834c8017c208c9e","target":"graph","created_at":"2026-05-17T23:40:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"It can be challenging to train multi-task neural networks that outperform or even match their single-task counterparts. To help address this, we propose using knowledge distillation where single-task models teach a multi-task model. We enhance this training with teacher annealing, a novel method that gradually transitions the model from distillation to supervised learning, helping the multi-task model surpass its single-task teachers. We evaluate our approach by multi-task fine-tuning BERT on the GLUE benchmark. Our method consistently improves over standard single-task and multi-task training","authors_text":"Christopher D. Manning, Kevin Clark, Minh-Thang Luong, Quoc V. Le, Urvashi Khandelwal","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-07-10T17:14:47Z","title":"BAM! Born-Again Multi-Task Networks for Natural Language Understanding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1907.04829","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:2fc58fcf9e83c974fee3df99e18e1bbb36429ff30afabbc2ddbdc7d32af97a2c","target":"record","created_at":"2026-05-17T23:40:56Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5caf51a454972cc2c521357ef1c8c17c63cac92de4edea6257949f2243781e41","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-07-10T17:14:47Z","title_canon_sha256":"62e94252249c83973ddb886d192787ba0c94db747de22add975161a39a53b7f9"},"schema_version":"1.0","source":{"id":"1907.04829","kind":"arxiv","version":1}},"canonical_sha256":"9fd0a5eccb05d614bb1756585db0763e3a6c805b3115a552673dc43cafa8d0c7","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9fd0a5eccb05d614bb1756585db0763e3a6c805b3115a552673dc43cafa8d0c7","first_computed_at":"2026-05-17T23:40:56.569407Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:40:56.569407Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PrPIx31YPJbbX1T7pm7XI4w291IYXpTGrw7lAYngoX/IroRog270UmyYpH0bxY99rM0j2COKTX22m0rlK5x+Dw==","signature_status":"signed_v1","signed_at":"2026-05-17T23:40:56.570042Z","signed_message":"canonical_sha256_bytes"},"source_id":"1907.04829","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2fc58fcf9e83c974fee3df99e18e1bbb36429ff30afabbc2ddbdc7d32af97a2c","sha256:54188db03a6e4bc4727cf9528b527cf2b6f280565591a43dc834c8017c208c9e"],"state_sha256":"1405416dfb92d7ba260c72123009f7183c1a317bde56f9054dfa9ab88ddde227"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"675jqATxCdePOba+BMYbRN9PW4V6sLRktQykpl84YFw0+up7MVUnh7/ZsZZvG/bq7QUVbPU1JnEjN1275JQaDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-30T16:30:07.487630Z","bundle_sha256":"8ac51131eeae9b4d1300d4a5e4888c5dcae3e5eda08fa7be8397d572b877480d"}}