{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:FHXUMIEKFCKMMRYQ5VB3H26VSE","merge_version":"pith-open-graph-merge-v1","event_count":5,"valid_event_count":5,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9e7f88313850a5eaee3ab810d62ea492776f67bf233a5c15e27493754773f730","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T07:00:26Z","title_canon_sha256":"d621ba3c98dd9262c7ecb24095d95a674207be560406ebfead348bb859db2e4a"},"schema_version":"1.0","source":{"id":"2605.14465","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.14465","created_at":"2026-05-17T23:39:06Z"},{"alias_kind":"arxiv_version","alias_value":"2605.14465v1","created_at":"2026-05-17T23:39:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14465","created_at":"2026-05-17T23:39:06Z"},{"alias_kind":"pith_short_12","alias_value":"FHXUMIEKFCKM","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"FHXUMIEKFCKMMRYQ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"FHXUMIEK","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:921ae89b24640efde4b60a762f0b1a0185c1fc0ebc9f8deb37d49ca3ac5e4043","target":"graph","created_at":"2026-05-17T23:39:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Across eight benchmarks covering table question answering and fact verification, TABALIGN improves average accuracy by 15.76 percentage points over the strongest open-source baseline at comparable 8B-class scale, with a matched-backbone ablation attributing 2.87 percentage points of this gain to the DLM planner over an AR planner on a fixed reasoner."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The 1,600 human-verified attention standards used to train TABATTN are representative and stable across different table layouts, domains, and model scales; if they are not, the verifier's scoring may not reliably enforce the cell-grounding contract."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"TABALIGN pairs a diffusion language model planner emitting binary cell masks with a trained attention verifier, raising average accuracy 15.76 points over strong baselines on eight table benchmarks while speeding execution 44.64%."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"TABALIGN pairs a diffusion language model planner that emits binary cell masks with an attention verifier trained on human standards to enforce cell grounding in table reasoning."}],"snapshot_sha256":"c3522d648c2513534933f5c268480c369bc6378ce1d392ac19867ded80cb822d"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Multi-step LLM reasoning over structured tables fails because planning and execution share no explicit cell-grounding contract. Existing methods constrain the planner to a left-to-right factorization at odds with table permutation invariance, and score intermediate states by generated content alone, overlooking cell grounding. We conduct a pilot study showing that diffusion language models (DLMs) produce more human-aligned and permutation-stable cell attention on tables than autoregressive models, with a 40.2% median reduction in attention-AUROC variability under row reordering. Motivated by t","authors_text":"Chunhe Wang, Guang Cheng, Hanwei Wu, Lei Ding, Tung Sum Thomas Kwok, Xiaofeng Lin, Xinyu Wang, Zeyong Zhang, Zhijiang Guo","cross_cats":[],"headline":"TABALIGN pairs a diffusion language model planner that emits binary cell masks with an attention verifier trained on human standards to enforce cell grounding in table reasoning.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T07:00:26Z","title":"From Table to Cell: Attention for Better Reasoning with TABALIGN"},"references":{"count":87,"internal_anchors":15,"resolved_work":87,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Adapting autoregressive vision language models for parallel diffusion decoding, 2025","work_id":"7ffb24b6-5db0-46fd-a761-f5611015ec40","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Johnson, Jonathan Ho, Daniel Tarlow, and Rianne van den Berg","work_id":"368943f8-485f-4c02-aa7b-601f5fc8ce36","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"URLhttps://openreview.net/forum?id=h7-XixPCAL","work_id":"54befcc6-ed61-456b-bfa5-58d9d4a3238e","year":null},{"cited_arxiv_id":"2511.21631","doi":"","is_internal_anchor":true,"ref_index":4,"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","year":2025},{"cited_arxiv_id":"2502.13923","doi":"","is_internal_anchor":true,"ref_index":5,"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","year":2025}],"snapshot_sha256":"bff895ffc51f7c669858c88e4a91ae9781cccab955420665c020172f61480284"},"source":{"id":"2605.14465","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T01:25:47.957637Z","id":"650f87fb-fe46-407e-b4bd-8ef7b50b9cb6","model_set":{"reader":"grok-4.3"},"one_line_summary":"TABALIGN pairs a diffusion language model planner emitting binary cell masks with a trained attention verifier, raising average accuracy 15.76 points over strong baselines on eight table benchmarks while speeding execution 44.64%.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"TABALIGN pairs a diffusion language model planner that emits binary cell masks with an attention verifier trained on human standards to enforce cell grounding in table reasoning.","strongest_claim":"Across eight benchmarks covering table question answering and fact verification, TABALIGN improves average accuracy by 15.76 percentage points over the strongest open-source baseline at comparable 8B-class scale, with a matched-backbone ablation attributing 2.87 percentage points of this gain to the DLM planner over an AR planner on a fixed reasoner.","weakest_assumption":"The 1,600 human-verified attention standards used to train TABATTN are representative and stable across different table layouts, domains, and model scales; if they are not, the verifier's scoring may not reliably enforce the cell-grounding contract."}},"verdict_id":"650f87fb-fe46-407e-b4bd-8ef7b50b9cb6"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:16858fe3c72b8a497ff8895476027b85c71d64e2fb495e3ac848b5f0d0ef0af2","target":"record","created_at":"2026-05-17T23:39:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9e7f88313850a5eaee3ab810d62ea492776f67bf233a5c15e27493754773f730","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T07:00:26Z","title_canon_sha256":"d621ba3c98dd9262c7ecb24095d95a674207be560406ebfead348bb859db2e4a"},"schema_version":"1.0","source":{"id":"2605.14465","kind":"arxiv","version":1}},"canonical_sha256":"29ef46208a2894c64710ed43b3ebd5913f1406d4c3ba7d4c75de612c9832ec84","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"29ef46208a2894c64710ed43b3ebd5913f1406d4c3ba7d4c75de612c9832ec84","first_computed_at":"2026-05-17T23:39:06.731364Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:06.731364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"wn95A7Lx2xKQjVZ+/0yEctJjO53uXChKXofCBbWjdI2g3kFG5TqEins6akMw7f7HbQ1lXGp+gdeMqdqxddkaAQ==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:06.732158Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.14465","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:1e75e00b0df1ba2735017ab27b267906e3ebcb827ae4c20102e100c1d7519633","sha256:551ff17555c7ff2af5adb81e28297dd9cd32d5be767fe9dd303b1c3b6ee752c4","sha256:fcd2c994dd1c758f221b516bbef9aa5eab415291b184febb64169c7de9038c07"]}],"invalid_events":[],"applied_event_ids":["sha256:16858fe3c72b8a497ff8895476027b85c71d64e2fb495e3ac848b5f0d0ef0af2","sha256:921ae89b24640efde4b60a762f0b1a0185c1fc0ebc9f8deb37d49ca3ac5e4043"],"state_sha256":"d5f9c64eee135b3ea75a6cd43379dd1f316852ba0f4f1e3cceeeccf51d40e6b0"}