{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4KGTWBJGTI3ERLYPRIZON6CAAO","short_pith_number":"pith:4KGTWBJG","schema_version":"1.0","canonical_sha256":"e28d3b05269a3648af0f8a32e6f84003afc82fcd063bfac70ec5ccd87f7ebfb3","source":{"kind":"arxiv","id":"2410.01949","version":2},"attestation_state":"computed","paper":{"title":"Discrete Copula Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anji Liu, Guy Van den Broeck, Mathias Niepert, Oliver Broadrick","submitted_at":"2024-10-02T18:51:38Z","abstract_excerpt":"Discrete diffusion models have recently shown significant progress in modeling complex data, such as natural languages and DNA sequences. However, unlike diffusion models for continuous data, which can generate high-quality samples in just a few denoising steps, modern discrete diffusion models still require hundreds or even thousands of denoising steps to perform well. In this paper, we identify a fundamental limitation that prevents discrete diffusion models from achieving strong performance with fewer steps -- they fail to capture dependencies between output variables at each denoising step"},"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":"2410.01949","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-02T18:51:38Z","cross_cats_sorted":[],"title_canon_sha256":"15f6644a06ff470a1ab93fb781891e114d72f908ff0d7b45081e4acbb8029b1d","abstract_canon_sha256":"cea915438e1786130e0428d78c80ac30a871380dfbbfc0dca7d3cc6397a87c5e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:34:10.696246Z","signature_b64":"SUEq/Sx+HU6g2NUoIKOcwd6geoCG0c3PygIGcIAnsXNKbHYuUgCO2a6PAk45tYuQ9nJd5yS0heLwJn2Q8ERiAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e28d3b05269a3648af0f8a32e6f84003afc82fcd063bfac70ec5ccd87f7ebfb3","last_reissued_at":"2026-07-05T10:34:10.695734Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:34:10.695734Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Discrete Copula Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anji Liu, Guy Van den Broeck, Mathias Niepert, Oliver Broadrick","submitted_at":"2024-10-02T18:51:38Z","abstract_excerpt":"Discrete diffusion models have recently shown significant progress in modeling complex data, such as natural languages and DNA sequences. However, unlike diffusion models for continuous data, which can generate high-quality samples in just a few denoising steps, modern discrete diffusion models still require hundreds or even thousands of denoising steps to perform well. In this paper, we identify a fundamental limitation that prevents discrete diffusion models from achieving strong performance with fewer steps -- they fail to capture dependencies between output variables at each denoising step"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.01949","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/2410.01949/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":"2410.01949","created_at":"2026-07-05T10:34:10.695798+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.01949v2","created_at":"2026-07-05T10:34:10.695798+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.01949","created_at":"2026-07-05T10:34:10.695798+00:00"},{"alias_kind":"pith_short_12","alias_value":"4KGTWBJGTI3E","created_at":"2026-07-05T10:34:10.695798+00:00"},{"alias_kind":"pith_short_16","alias_value":"4KGTWBJGTI3ERLYP","created_at":"2026-07-05T10:34:10.695798+00:00"},{"alias_kind":"pith_short_8","alias_value":"4KGTWBJG","created_at":"2026-07-05T10:34:10.695798+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09159","citing_title":"Unified Energy for Invariant and Independent Decoding in Diffusion Language Models","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02955","citing_title":"Fast-dLLM++: Fr\\'{e}chet Profile Decoding for Faster Diffusion LLM Inference","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2509.19707","citing_title":"Diffusion and Flow-based Copulas: Forgetting and Remembering Dependencies","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2505.17384","citing_title":"Variational Autoencoding Discrete Diffusion with Enhanced Dimensional Correlations Modeling","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2505.22618","citing_title":"Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO","json":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO.json","graph_json":"https://pith.science/api/pith-number/4KGTWBJGTI3ERLYPRIZON6CAAO/graph.json","events_json":"https://pith.science/api/pith-number/4KGTWBJGTI3ERLYPRIZON6CAAO/events.json","paper":"https://pith.science/paper/4KGTWBJG"},"agent_actions":{"view_html":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO","download_json":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO.json","view_paper":"https://pith.science/paper/4KGTWBJG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.01949&json=true","fetch_graph":"https://pith.science/api/pith-number/4KGTWBJGTI3ERLYPRIZON6CAAO/graph.json","fetch_events":"https://pith.science/api/pith-number/4KGTWBJGTI3ERLYPRIZON6CAAO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO/action/storage_attestation","attest_author":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO/action/author_attestation","sign_citation":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO/action/citation_signature","submit_replication":"https://pith.science/pith/4KGTWBJGTI3ERLYPRIZON6CAAO/action/replication_record"}},"created_at":"2026-07-05T10:34:10.695798+00:00","updated_at":"2026-07-05T10:34:10.695798+00:00"}