{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:3X2OFTIFROYUP7BQVYRU3VENAF","short_pith_number":"pith:3X2OFTIF","schema_version":"1.0","canonical_sha256":"ddf4e2cd058bb147fc30ae234dd48d0170ada8eab0a3812e4106fa49f8a7a882","source":{"kind":"arxiv","id":"2305.19847","version":1},"attestation_state":"computed","paper":{"title":"How Does Pretraining Improve Discourse-Aware Translation?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Derek F. Wong, Longyue Wang, Siyou Liu, Zhihong Huang","submitted_at":"2023-05-31T13:36:51Z","abstract_excerpt":"Pretrained language models (PLMs) have produced substantial improvements in discourse-aware neural machine translation (NMT), for example, improved coherence in spoken language translation. However, the underlying reasons for their strong performance have not been well explained. To bridge this gap, we introduce a probing task to interpret the ability of PLMs to capture discourse relation knowledge. We validate three state-of-the-art PLMs across encoder-, decoder-, and encoder-decoder-based models. The analysis shows that (1) the ability of PLMs on discourse modelling varies from architecture "},"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":"2305.19847","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-31T13:36:51Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"046c78911d090de513981e3476a8eeb779da3630281bce457e386bf09e095a78","abstract_canon_sha256":"1aa93a6ba693a7499a765ce51cd80b73a499438cec12d008f9144748e4166183"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:16:03.940706Z","signature_b64":"mqxmhRY8JVU8pbBXIHMTLYH4cJfNRF+9lY8OuGEBwTINP8O0Cquk6gUK/CQM4fpX7iVskeFaTVXmJJ+YBCysDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ddf4e2cd058bb147fc30ae234dd48d0170ada8eab0a3812e4106fa49f8a7a882","last_reissued_at":"2026-07-05T06:16:03.940208Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:16:03.940208Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Does Pretraining Improve Discourse-Aware Translation?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Derek F. Wong, Longyue Wang, Siyou Liu, Zhihong Huang","submitted_at":"2023-05-31T13:36:51Z","abstract_excerpt":"Pretrained language models (PLMs) have produced substantial improvements in discourse-aware neural machine translation (NMT), for example, improved coherence in spoken language translation. However, the underlying reasons for their strong performance have not been well explained. To bridge this gap, we introduce a probing task to interpret the ability of PLMs to capture discourse relation knowledge. We validate three state-of-the-art PLMs across encoder-, decoder-, and encoder-decoder-based models. The analysis shows that (1) the ability of PLMs on discourse modelling varies from architecture "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.19847","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2305.19847/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":"2305.19847","created_at":"2026-07-05T06:16:03.940280+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.19847v1","created_at":"2026-07-05T06:16:03.940280+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.19847","created_at":"2026-07-05T06:16:03.940280+00:00"},{"alias_kind":"pith_short_12","alias_value":"3X2OFTIFROYU","created_at":"2026-07-05T06:16:03.940280+00:00"},{"alias_kind":"pith_short_16","alias_value":"3X2OFTIFROYUP7BQ","created_at":"2026-07-05T06:16:03.940280+00:00"},{"alias_kind":"pith_short_8","alias_value":"3X2OFTIF","created_at":"2026-07-05T06:16:03.940280+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/3X2OFTIFROYUP7BQVYRU3VENAF","json":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF.json","graph_json":"https://pith.science/api/pith-number/3X2OFTIFROYUP7BQVYRU3VENAF/graph.json","events_json":"https://pith.science/api/pith-number/3X2OFTIFROYUP7BQVYRU3VENAF/events.json","paper":"https://pith.science/paper/3X2OFTIF"},"agent_actions":{"view_html":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF","download_json":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF.json","view_paper":"https://pith.science/paper/3X2OFTIF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.19847&json=true","fetch_graph":"https://pith.science/api/pith-number/3X2OFTIFROYUP7BQVYRU3VENAF/graph.json","fetch_events":"https://pith.science/api/pith-number/3X2OFTIFROYUP7BQVYRU3VENAF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF/action/storage_attestation","attest_author":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF/action/author_attestation","sign_citation":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF/action/citation_signature","submit_replication":"https://pith.science/pith/3X2OFTIFROYUP7BQVYRU3VENAF/action/replication_record"}},"created_at":"2026-07-05T06:16:03.940280+00:00","updated_at":"2026-07-05T06:16:03.940280+00:00"}