{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6FZOTTRRYSMQUT6JXVZ4XZPS3U","short_pith_number":"pith:6FZOTTRR","schema_version":"1.0","canonical_sha256":"f172e9ce31c4990a4fc9bd73cbe5f2dd207eb1d88b8162fb23d4685857f4ab9c","source":{"kind":"arxiv","id":"2605.25820","version":1},"attestation_state":"computed","paper":{"title":"Visual-Redundancy-Controlled Parallel Decoding for Diffusion-Based Multimodal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongshuo Zhao, Xiangming Meng, Yulin Yuan","submitted_at":"2026-05-25T13:16:51Z","abstract_excerpt":"Diffusion-based multimodal large language models (dMLLMs) decode by iteratively predicting tokens at multiple masked positions in parallel. This turns each decoding step into a position-selection problem: the model must choose not only which predictions are reliable in isolation, but also which positions should be committed together as context for later decoding steps. Existing confidence-based decoding ranks masked positions independently and commits the top-K positions, largely ignoring whether the committed tokens provide complementary visual grounding. We identify a step-level limitation o"},"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":"2605.25820","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T13:16:51Z","cross_cats_sorted":[],"title_canon_sha256":"11506ae90fb9209514ca4c0c9cad22273fb0cef48c1be46dec7cb5d59739029b","abstract_canon_sha256":"7f37226635dd6517c1ae7a08a1b84912cb62fc87c8239898bc365b3d0d0bd674"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:05:13.279878Z","signature_b64":"GmCTY7P8czrGBXX9acNaZh55kkKql8wcw5uKiOZhq5yFeWH6DxEcGL1/DUFFkZ01vdgwZbqvLpF7VNRNe4SCDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f172e9ce31c4990a4fc9bd73cbe5f2dd207eb1d88b8162fb23d4685857f4ab9c","last_reissued_at":"2026-05-26T02:05:13.279145Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:05:13.279145Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Visual-Redundancy-Controlled Parallel Decoding for Diffusion-Based Multimodal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongshuo Zhao, Xiangming Meng, Yulin Yuan","submitted_at":"2026-05-25T13:16:51Z","abstract_excerpt":"Diffusion-based multimodal large language models (dMLLMs) decode by iteratively predicting tokens at multiple masked positions in parallel. This turns each decoding step into a position-selection problem: the model must choose not only which predictions are reliable in isolation, but also which positions should be committed together as context for later decoding steps. Existing confidence-based decoding ranks masked positions independently and commits the top-K positions, largely ignoring whether the committed tokens provide complementary visual grounding. We identify a step-level limitation o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25820","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/2605.25820/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":"2605.25820","created_at":"2026-05-26T02:05:13.279253+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.25820v1","created_at":"2026-05-26T02:05:13.279253+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25820","created_at":"2026-05-26T02:05:13.279253+00:00"},{"alias_kind":"pith_short_12","alias_value":"6FZOTTRRYSMQ","created_at":"2026-05-26T02:05:13.279253+00:00"},{"alias_kind":"pith_short_16","alias_value":"6FZOTTRRYSMQUT6J","created_at":"2026-05-26T02:05:13.279253+00:00"},{"alias_kind":"pith_short_8","alias_value":"6FZOTTRR","created_at":"2026-05-26T02:05:13.279253+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/6FZOTTRRYSMQUT6JXVZ4XZPS3U","json":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U.json","graph_json":"https://pith.science/api/pith-number/6FZOTTRRYSMQUT6JXVZ4XZPS3U/graph.json","events_json":"https://pith.science/api/pith-number/6FZOTTRRYSMQUT6JXVZ4XZPS3U/events.json","paper":"https://pith.science/paper/6FZOTTRR"},"agent_actions":{"view_html":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U","download_json":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U.json","view_paper":"https://pith.science/paper/6FZOTTRR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.25820&json=true","fetch_graph":"https://pith.science/api/pith-number/6FZOTTRRYSMQUT6JXVZ4XZPS3U/graph.json","fetch_events":"https://pith.science/api/pith-number/6FZOTTRRYSMQUT6JXVZ4XZPS3U/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U/action/storage_attestation","attest_author":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U/action/author_attestation","sign_citation":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U/action/citation_signature","submit_replication":"https://pith.science/pith/6FZOTTRRYSMQUT6JXVZ4XZPS3U/action/replication_record"}},"created_at":"2026-05-26T02:05:13.279253+00:00","updated_at":"2026-05-26T02:05:13.279253+00:00"}