{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VI6DPXKK7VLBXFNZTCHKVXUDTZ","short_pith_number":"pith:VI6DPXKK","schema_version":"1.0","canonical_sha256":"aa3c37dd4afd561b95b9988eaade839e564a23a9343a79d27c945ba20b7c7eae","source":{"kind":"arxiv","id":"2607.24507","version":1},"attestation_state":"computed","paper":{"title":"UNIFUSION: Adapting Autoregressive Language Models into Discrete Diffusion under a Unified Reverse-Rate Objective","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jingyuan Li, Pipi Hu, Wei Liu, Xiaoyi Jiang, Yixuan Jiang, Yi Zhu, Zuoqiang Shi","submitted_at":"2026-07-27T14:40:26Z","abstract_excerpt":"Existing methods mainly adapt pretrained autoregressive (AR) language models to masked diffusion, whereas we directly adapt them to uniform-noise diffusion, where every token remains editable during sampling. However, adapting AR checkpoints across corruption kernels remains challenging because existing DLMs use different objectives and prediction parameterizations. We establish connections among SEDD, MDLM/GIDD, M2S, and Neural CTMC by expressing their conditional losses as a single generalized Kullback--Leibler objective over model reverse rates. We further derive conversions from clean-toke"},"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":"2607.24507","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-27T14:40:26Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0efef5f41fa1722de38eb3c8efaceaea2e1e8a8a08748d262c950dce81b2a31f","abstract_canon_sha256":"095f64ca9ca78f5bd8cd01b5959c064593268516af93fa33ab15f0880ba199f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-28T02:24:07.142081Z","signature_b64":"cbPw4UHqYhchxnkmMIdlW1NkYiGdKk+AXksLpEZUgQ/AGzU4NmYNFYJTSkzNiOx16cAcXfcS4GSQPwS4zxkNDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aa3c37dd4afd561b95b9988eaade839e564a23a9343a79d27c945ba20b7c7eae","last_reissued_at":"2026-07-28T02:24:07.141265Z","signature_status":"signed_v1","first_computed_at":"2026-07-28T02:24:07.141265Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UNIFUSION: Adapting Autoregressive Language Models into Discrete Diffusion under a Unified Reverse-Rate Objective","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jingyuan Li, Pipi Hu, Wei Liu, Xiaoyi Jiang, Yixuan Jiang, Yi Zhu, Zuoqiang Shi","submitted_at":"2026-07-27T14:40:26Z","abstract_excerpt":"Existing methods mainly adapt pretrained autoregressive (AR) language models to masked diffusion, whereas we directly adapt them to uniform-noise diffusion, where every token remains editable during sampling. However, adapting AR checkpoints across corruption kernels remains challenging because existing DLMs use different objectives and prediction parameterizations. We establish connections among SEDD, MDLM/GIDD, M2S, and Neural CTMC by expressing their conditional losses as a single generalized Kullback--Leibler objective over model reverse rates. We further derive conversions from clean-toke"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.24507","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/2607.24507/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":"2607.24507","created_at":"2026-07-28T02:24:07.141686+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.24507v1","created_at":"2026-07-28T02:24:07.141686+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.24507","created_at":"2026-07-28T02:24:07.141686+00:00"},{"alias_kind":"pith_short_12","alias_value":"VI6DPXKK7VLB","created_at":"2026-07-28T02:24:07.141686+00:00"},{"alias_kind":"pith_short_16","alias_value":"VI6DPXKK7VLBXFNZ","created_at":"2026-07-28T02:24:07.141686+00:00"},{"alias_kind":"pith_short_8","alias_value":"VI6DPXKK","created_at":"2026-07-28T02:24:07.141686+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/VI6DPXKK7VLBXFNZTCHKVXUDTZ","json":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ.json","graph_json":"https://pith.science/api/pith-number/VI6DPXKK7VLBXFNZTCHKVXUDTZ/graph.json","events_json":"https://pith.science/api/pith-number/VI6DPXKK7VLBXFNZTCHKVXUDTZ/events.json","paper":"https://pith.science/paper/VI6DPXKK"},"agent_actions":{"view_html":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ","download_json":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ.json","view_paper":"https://pith.science/paper/VI6DPXKK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.24507&json=true","fetch_graph":"https://pith.science/api/pith-number/VI6DPXKK7VLBXFNZTCHKVXUDTZ/graph.json","fetch_events":"https://pith.science/api/pith-number/VI6DPXKK7VLBXFNZTCHKVXUDTZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ/action/storage_attestation","attest_author":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ/action/author_attestation","sign_citation":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ/action/citation_signature","submit_replication":"https://pith.science/pith/VI6DPXKK7VLBXFNZTCHKVXUDTZ/action/replication_record"}},"created_at":"2026-07-28T02:24:07.141686+00:00","updated_at":"2026-07-28T02:24:07.141686+00:00"}