{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2019:M2LDPC6T2WJ34DGR7WSDJZAHY3","short_pith_number":"pith:M2LDPC6T","canonical_record":{"source":{"id":"1903.09722","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-03-22T22:14:51Z","cross_cats_sorted":[],"title_canon_sha256":"105032e0fb31c84105fe857b10a2a249dcf6e2f5ac75ec688e12329e4d6b89dd","abstract_canon_sha256":"d099c1d35336950ab0d9b27186c5930ce10cc38fe433c931d1f9fce629c1c5aa"},"schema_version":"1.0"},"canonical_sha256":"6696378bd3d593be0cd1fda434e407c6f141d86fbe52c82875ead17559db9443","source":{"kind":"arxiv","id":"1903.09722","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1903.09722","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"arxiv_version","alias_value":"1903.09722v2","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1903.09722","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"pith_short_12","alias_value":"M2LDPC6T2WJ3","created_at":"2026-05-18T12:33:21Z"},{"alias_kind":"pith_short_16","alias_value":"M2LDPC6T2WJ34DGR","created_at":"2026-05-18T12:33:21Z"},{"alias_kind":"pith_short_8","alias_value":"M2LDPC6T","created_at":"2026-05-18T12:33:21Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2019:M2LDPC6T2WJ34DGR7WSDJZAHY3","target":"record","payload":{"canonical_record":{"source":{"id":"1903.09722","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-03-22T22:14:51Z","cross_cats_sorted":[],"title_canon_sha256":"105032e0fb31c84105fe857b10a2a249dcf6e2f5ac75ec688e12329e4d6b89dd","abstract_canon_sha256":"d099c1d35336950ab0d9b27186c5930ce10cc38fe433c931d1f9fce629c1c5aa"},"schema_version":"1.0"},"canonical_sha256":"6696378bd3d593be0cd1fda434e407c6f141d86fbe52c82875ead17559db9443","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:49:44.948165Z","signature_b64":"8z3HOXBedbicFlM2YRbSokDJ9yKR/Md9DunQkmJc0U6C1iKUAhoNJGVbwPkgkVU3w7a78HouWoODJQYkvg9qDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6696378bd3d593be0cd1fda434e407c6f141d86fbe52c82875ead17559db9443","last_reissued_at":"2026-05-17T23:49:44.947702Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:49:44.947702Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1903.09722","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:49:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pMlTfFSYaDFIKUtahqLNWNE48lMfvz/ctAifXZG9D+4EDSxcn953f+Aj8ylPt6jItRG17oRZ5hHC/tNC32cODg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T02:10:51.199137Z"},"content_sha256":"7d86a08606274d69ba58c45d468ccf80ab81206180b5aba24bf0b7cbf20bf88c","schema_version":"1.0","event_id":"sha256:7d86a08606274d69ba58c45d468ccf80ab81206180b5aba24bf0b7cbf20bf88c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2019:M2LDPC6T2WJ34DGR7WSDJZAHY3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Pre-trained Language Model Representations for Language Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alexei Baevski, Michael Auli, Sergey Edunov","submitted_at":"2019-03-22T22:14:51Z","abstract_excerpt":"Pre-trained language model representations have been successful in a wide range of language understanding tasks. In this paper, we examine different strategies to integrate pre-trained representations into sequence to sequence models and apply it to neural machine translation and abstractive summarization. We find that pre-trained representations are most effective when added to the encoder network which slows inference by only 14%. Our experiments in machine translation show gains of up to 5.3 BLEU in a simulated resource-poor setup. While returns diminish with more labeled data, we still obs"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1903.09722","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":""},"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":"2026-05-17T23:49:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z70BIKqZPSio8nAoPFCPzpBinft6t+jrme4PsiRw7f0jRyYgYTrNosPlUqoD7ZWlf/sbGrnKkVxDiwbHjt7oDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T02:10:51.199502Z"},"content_sha256":"bb27378b359a1b3f0cc1cf4a6b01e579cf0594cea31c2c803127d7202a0c50ef","schema_version":"1.0","event_id":"sha256:bb27378b359a1b3f0cc1cf4a6b01e579cf0594cea31c2c803127d7202a0c50ef"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/bundle.json","state_url":"https://pith.science/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/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":"2026-05-27T02:10:51Z","links":{"resolver":"https://pith.science/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3","bundle":"https://pith.science/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/bundle.json","state":"https://pith.science/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/M2LDPC6T2WJ34DGR7WSDJZAHY3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2019:M2LDPC6T2WJ34DGR7WSDJZAHY3","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":"d099c1d35336950ab0d9b27186c5930ce10cc38fe433c931d1f9fce629c1c5aa","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-03-22T22:14:51Z","title_canon_sha256":"105032e0fb31c84105fe857b10a2a249dcf6e2f5ac75ec688e12329e4d6b89dd"},"schema_version":"1.0","source":{"id":"1903.09722","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1903.09722","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"arxiv_version","alias_value":"1903.09722v2","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1903.09722","created_at":"2026-05-17T23:49:44Z"},{"alias_kind":"pith_short_12","alias_value":"M2LDPC6T2WJ3","created_at":"2026-05-18T12:33:21Z"},{"alias_kind":"pith_short_16","alias_value":"M2LDPC6T2WJ34DGR","created_at":"2026-05-18T12:33:21Z"},{"alias_kind":"pith_short_8","alias_value":"M2LDPC6T","created_at":"2026-05-18T12:33:21Z"}],"graph_snapshots":[{"event_id":"sha256:bb27378b359a1b3f0cc1cf4a6b01e579cf0594cea31c2c803127d7202a0c50ef","target":"graph","created_at":"2026-05-17T23:49:44Z","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":"Pre-trained language model representations have been successful in a wide range of language understanding tasks. In this paper, we examine different strategies to integrate pre-trained representations into sequence to sequence models and apply it to neural machine translation and abstractive summarization. We find that pre-trained representations are most effective when added to the encoder network which slows inference by only 14%. Our experiments in machine translation show gains of up to 5.3 BLEU in a simulated resource-poor setup. While returns diminish with more labeled data, we still obs","authors_text":"Alexei Baevski, Michael Auli, Sergey Edunov","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-03-22T22:14:51Z","title":"Pre-trained Language Model Representations for Language Generation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1903.09722","kind":"arxiv","version":2},"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":"sha256:7d86a08606274d69ba58c45d468ccf80ab81206180b5aba24bf0b7cbf20bf88c","target":"record","created_at":"2026-05-17T23:49:44Z","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":"d099c1d35336950ab0d9b27186c5930ce10cc38fe433c931d1f9fce629c1c5aa","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-03-22T22:14:51Z","title_canon_sha256":"105032e0fb31c84105fe857b10a2a249dcf6e2f5ac75ec688e12329e4d6b89dd"},"schema_version":"1.0","source":{"id":"1903.09722","kind":"arxiv","version":2}},"canonical_sha256":"6696378bd3d593be0cd1fda434e407c6f141d86fbe52c82875ead17559db9443","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6696378bd3d593be0cd1fda434e407c6f141d86fbe52c82875ead17559db9443","first_computed_at":"2026-05-17T23:49:44.947702Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:49:44.947702Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8z3HOXBedbicFlM2YRbSokDJ9yKR/Md9DunQkmJc0U6C1iKUAhoNJGVbwPkgkVU3w7a78HouWoODJQYkvg9qDQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:49:44.948165Z","signed_message":"canonical_sha256_bytes"},"source_id":"1903.09722","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7d86a08606274d69ba58c45d468ccf80ab81206180b5aba24bf0b7cbf20bf88c","sha256:bb27378b359a1b3f0cc1cf4a6b01e579cf0594cea31c2c803127d7202a0c50ef"],"state_sha256":"8367e9201c3f4b028aeb711296e674be088435652d8c2296ae08a9d874d80fa9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6WPAblihx6Wpjn9fBNqJ6sSm+jOUS5lmgyMCFzzcElZVhCx7+Jkjc50MCFD+GTr9Qtx+PAty0PONylgPu3vvAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-27T02:10:51.201606Z","bundle_sha256":"b6c86126ad90309fee906740e9c0dc50f7e0a140ea78e70c7a74a2a60f2c7ac1"}}