{"id":"dc627bcd-bebb-4b74-8050-d20c6f4c88d3","arxiv_id":"2605.29670","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"EviLink combines multi-hypothesis schema grounding with uncertainty-guided evidence acquisition, reporting 90.15% field-level recall and 123.30K average tokens on Spider2-Snow while improving downstream SQL generation.","lead":"EviLink reframes schema linking for Text-to-SQL as uncertainty-aware inference over multiple plausible SQL paths, acquiring evidence only for uncertain schema items. This targets better balance of completeness, relevance, and token cost on large ambiguous databases.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's UNVERDICTED status stems directly from abstract-only access; the argument's logic is internally consistent at the level of description given, so the skeptic pass does not alter the verdict. The weakest_assumption is the correct place to focus future verification but does not yet constitute a demonstrated flaw.","tokens_in":1659,"tokens_out":278,"duration_ms":21511,"concrete_test":"Re-run the Spider2-Snow evaluation using the exact path-sampling procedure and uncertainty metric described in the methods; if recall drops below 85% or token cost rises above 150K while holding the generator fixed, the uncertainty-guided acquisition fails to deliver the claimed balance.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract presents an empirical claim of 90.15% field-level strict recall, 123.30K tokens, and downstream improvement on Spider2-Snow via multi-path uncertainty-guided schema linking. The reader's weakest assumption correctly identifies the key precondition (reliable multi-path generation plus calibrated uncertainty to avoid omission), but the provided text contains no internal contradiction, unstated assumption violation, or inconsistent derivation that would falsify the central reframing. Without access to the actual methods or results sections, no load-bearing flaw can be isolated from the argument as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper reframes schema linking for large-scale Text-to-SQL as uncertainty-aware inference over multiple plausible SQL paths rather than deterministic selection around a single path. It introduces EviLink, which performs multi-hypothesis schema grounding combined with uncertainty-guided evidence acquisition to distinguish required schema items from path-dependent ones. Experiments on BIRD-Dev and Spider2-Snow are reported to show improved balance among schema completeness, relevance, and token cost; on Spider2-Snow the method achieves 90.15% field-level strict recall while using 123.30K average tokens and yields better downstream SQL generation under a fixed generator.","tokens_in":1754,"tokens_out":347,"duration_ms":24868,"significance":"If the empirical claims hold under rigorous validation, the work offers a meaningful advance for Text-to-SQL on large ambiguous databases by explicitly modeling query ambiguity through multi-path uncertainty. The reported recall/token trade-off on Spider2-Snow is practically relevant, and the reframing could influence future systems that must avoid both under- and over-retrieval of schema context. No machine-checked proofs or parameter-free derivations are claimed, but the empirical focus on falsifiable downstream improvement is a strength.","major_comments":[],"minor_comments":[{"comment":"Abstract: the description of how multi-hypothesis paths are generated and how uncertainty is quantified would benefit from one additional sentence to allow readers to assess the weakest assumption identified in the review process.","section":null},{"comment":"The manuscript should include a brief error analysis or ablation on cases where uncertainty estimates fail to surface critical schema items, even if only in the appendix.","section":null}],"recommendation":"uncertain","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the positive assessment of the work's significance and for recognizing the practical relevance of the recall/token trade-off on Spider2-Snow. The report does not enumerate any specific major comments, so we have no point-by-point rebuttals to provide at this stage.","responses":[],"tokens_in":1236,"tokens_out":76,"duration_ms":11454,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper's central move is to stop treating schema linking as picking items for one SQL path and instead run inference across multiple plausible paths, then use uncertainty to acquire evidence only where the paths disagree on what is needed. This is the main novelty relative to prior deterministic methods.\n\nThey implement it as EviLink with multi-hypothesis schema grounding and uncertainty-guided acquisition. The results on Spider2-Snow are 90.15% field-level strict recall at 123.30K average tokens, plus better SQL generation with a fixed generator. They also test on BIRD-Dev and claim a better balance of completeness, relevance, and cost.\n\nThis is a practical step because real questions can have several valid SQL realizations that require different schema subsets. Distinguishing required from path-dependent items makes sense for controlling token use in large databases.\n\nThe soft spot is that the abstract and high-level description do not include the method details, baseline comparisons, or error analysis needed to fully verify the claims. The assumption that multiple paths can be generated reliably and that uncertainty distinguishes the right items needs checking in the full paper. No load-bearing flaw is visible from what is presented, but the evidence is still thin without those sections.\n\nThis work is for people building or evaluating Text-to-SQL systems on big schemas. It is worth sending out for peer review because the problem is important in the subfield and the proposed reframing is distinct enough to merit referee scrutiny.","headline":"EviLink reframes schema linking as multi-path uncertainty-aware inference and reports concrete gains on Spider2-Snow, but the abstract leaves the mechanics thin.","tokens_in":2249,"tokens_out":371,"would_cite":false,"duration_ms":24398,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Schema linking for Text-to-SQL improves when multiple plausible paths guide uncertainty-based evidence selection.","keywords":["schema linking","text-to-sql","multi-path inference","uncertainty estimation","evidence acquisition","large-scale databases","sql generation"],"falsifier":"A dataset of ambiguous questions where EviLink's uncertainty scores systematically omit a critical schema item that appears in every valid path, producing lower recall than single-path baselines.","tokens_in":2576,"feed_emoji":"","tokens_out":636,"duration_ms":20357,"temperature":0.7,"pith_summary":"The paper establishes that schema linking should not treat selection as a deterministic choice around one SQL realization. Complex questions often admit several valid SQL paths, each requiring different schema items. By generating multiple hypotheses and using uncertainty estimates to separate items that are required regardless of path from those that vary, the system acquires evidence only for the uncertain cases. This reframing produces a better trade-off between completeness, relevance, and token usage. Experiments on BIRD-Dev and Spider2-Snow confirm higher field-level recall at lower average token cost while also lifting the performance of a fixed downstream SQL generator.","feed_headline":"Multiple SQL paths raise schema recall to 90% at lower token cost","feed_subtitle":"Uncertainty over plausible realizations lets the system request evidence only for items whose need varies across paths.","key_machinery":"Multi-hypothesis schema grounding paired with uncertainty-guided evidence acquisition, which identifies items needed across all paths while limiting evidence collection to uncertain positions.","core_discovery":"EviLink reframes schema linking as uncertainty-aware schema-need inference over multiple plausible SQL paths. It combines multi-hypothesis schema grounding with uncertainty-guided evidence acquisition to distinguish required schema items from path-dependent uncertain ones and acquires evidence only where needed, improving the balance among schema completeness, schema relevance, and token cost.","pith_inferences":["The same uncertainty signal could be reused to prune evidence in other retrieval-augmented generation settings that face large, ambiguous contexts.","If path sampling is cheap, the method suggests a general template for any task where one input maps to several valid outputs with differing context needs.","An extension could test whether the same multi-path uncertainty logic transfers to schema linking in other query languages or in visual question answering over large tables."],"forward_implications":["On Spider2-Snow the method reaches 90.15 percent field-level strict recall while using 123.30K average tokens.","The same procedure improves downstream SQL generation accuracy when the generator is held fixed.","The approach yields a measurable improvement in the three-way balance of completeness, relevance, and token cost on both BIRD-Dev and Spider2-Snow.","Evidence acquisition occurs selectively rather than uniformly across the schema."],"fun_headline_variants":["Multi-path SQL realizations inform schema needs with uncertainty","Uncertainty-guided acquisition only for needed schema evidence","Reframed schema linking distinguishes required and uncertain items","Balances completeness relevance and token cost in schema linking"],"cache_read_input_tokens":64,"weakest_assumption_plain":"Multiple plausible SQL paths can be generated reliably and uncertainty estimates over those paths accurately mark which schema items are required rather than path-specific.","fun_headline_variants_meta":{"raw":{"variants":["Multi-path SQL realizations inform schema needs with uncertainty","Uncertainty-guided acquisition only for needed schema evidence","Reframed schema linking distinguishes required and uncertain items","Balances completeness relevance and token cost in schema linking"]},"model":"grok-4.3","cost_usd":0.00684,"raw_usage":{"total_tokens":3146,"prompt_tokens":605,"num_sources_used":0,"completion_tokens":58,"cost_in_usd_ticks":68399500,"prompt_tokens_details":{"text_tokens":605,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2483,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":605,"tokens_out":58,"duration_ms":19917,"temperature":1.0,"reasoning_tokens":2483,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-29T07:50:20.263714+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A dataset of ambiguous questions where EviLink's uncertainty scores systematically omit a critical schema item that appears in every valid path, producing lower recall than single-path baselines.","supporting_citations":[],"review_version":1}