{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:UYAWYWEUTIYHP5SE6XUXZZLUQP","short_pith_number":"pith:UYAWYWEU","schema_version":"1.0","canonical_sha256":"a6016c58949a3077f644f5e97ce57483d64d9e2509edb5f4f97a318f25ed76b8","source":{"kind":"arxiv","id":"2205.14219","version":2},"attestation_state":"computed","paper":{"title":"Controllable Text Generation with Neurally-Decomposed Oracle","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kai-Wei Chang, Nanyun Peng, Sidi Lu, Tao Meng","submitted_at":"2022-05-27T20:17:53Z","abstract_excerpt":"We propose a general and efficient framework to control auto-regressive generation models with NeurAlly-Decomposed Oracle (NADO). Given a pre-trained base language model and a sequence-level boolean oracle function, we propose to decompose the oracle function into token-level guidance to steer the base model in text generation. Specifically, the token-level guidance is approximated by a neural model trained with examples sampled from the base model, demanding no additional auxiliary labeled data. Based on posterior regularization, we present the closed-form optimal solution to incorporate the "},"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":"2205.14219","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-27T20:17:53Z","cross_cats_sorted":[],"title_canon_sha256":"61a0c872f546cc7b5ace12e9fcf02374d8b850773480153d06670acc3c7e05b9","abstract_canon_sha256":"3ada9de59fae073c6210bb243c695df413ab9ca1f8b639ca16ef84aa83966899"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:08:52.186673Z","signature_b64":"C7lfMuNZLVnSsEIxYQfDSJ8R1/QS/z0KMAYzL8xSXiARb7tn0MV4S+jIY3Di6Wr/qTSkoYwc1iO8LZwbs/IcDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a6016c58949a3077f644f5e97ce57483d64d9e2509edb5f4f97a318f25ed76b8","last_reissued_at":"2026-07-05T05:08:52.186270Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:08:52.186270Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Controllable Text Generation with Neurally-Decomposed Oracle","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kai-Wei Chang, Nanyun Peng, Sidi Lu, Tao Meng","submitted_at":"2022-05-27T20:17:53Z","abstract_excerpt":"We propose a general and efficient framework to control auto-regressive generation models with NeurAlly-Decomposed Oracle (NADO). Given a pre-trained base language model and a sequence-level boolean oracle function, we propose to decompose the oracle function into token-level guidance to steer the base model in text generation. Specifically, the token-level guidance is approximated by a neural model trained with examples sampled from the base model, demanding no additional auxiliary labeled data. Based on posterior regularization, we present the closed-form optimal solution to incorporate the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.14219","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/2205.14219/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":"2205.14219","created_at":"2026-07-05T05:08:52.186325+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.14219v2","created_at":"2026-07-05T05:08:52.186325+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.14219","created_at":"2026-07-05T05:08:52.186325+00:00"},{"alias_kind":"pith_short_12","alias_value":"UYAWYWEUTIYH","created_at":"2026-07-05T05:08:52.186325+00:00"},{"alias_kind":"pith_short_16","alias_value":"UYAWYWEUTIYHP5SE","created_at":"2026-07-05T05:08:52.186325+00:00"},{"alias_kind":"pith_short_8","alias_value":"UYAWYWEU","created_at":"2026-07-05T05:08:52.186325+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.06930","citing_title":"DiscoSum: Discourse-aware News Summarization","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP","json":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP.json","graph_json":"https://pith.science/api/pith-number/UYAWYWEUTIYHP5SE6XUXZZLUQP/graph.json","events_json":"https://pith.science/api/pith-number/UYAWYWEUTIYHP5SE6XUXZZLUQP/events.json","paper":"https://pith.science/paper/UYAWYWEU"},"agent_actions":{"view_html":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP","download_json":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP.json","view_paper":"https://pith.science/paper/UYAWYWEU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.14219&json=true","fetch_graph":"https://pith.science/api/pith-number/UYAWYWEUTIYHP5SE6XUXZZLUQP/graph.json","fetch_events":"https://pith.science/api/pith-number/UYAWYWEUTIYHP5SE6XUXZZLUQP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP/action/storage_attestation","attest_author":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP/action/author_attestation","sign_citation":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP/action/citation_signature","submit_replication":"https://pith.science/pith/UYAWYWEUTIYHP5SE6XUXZZLUQP/action/replication_record"}},"created_at":"2026-07-05T05:08:52.186325+00:00","updated_at":"2026-07-05T05:08:52.186325+00:00"}