pith. sign in

arxiv: 2605.18766 · v1 · pith:R5FAQUOSnew · submitted 2026-04-12 · 💻 cs.IR · cs.AI· cs.CL

Retrieve Only Relevant Tables Whether Few or Many: Adaptive Table Retrieval Method

Taehee Kim , Seungbin Yang , Jihwan Kim , Jaegul Choo This is my paper

Pith reviewed 2026-05-21 01:04 UTC · model grok-4.3

classification 💻 cs.IR cs.AIcs.CL
keywords table retrievaltext-to-SQLadaptive retrievalinformation retrievaldatabase queryingquery processingretrieval augmentation
0
0 comments X

The pith

An adaptive thresholding method selects the right number of tables per query instead of using a fixed top-k.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that fixed top-k table retrieval for natural language questions over databases often either misses necessary tables or pulls in too many irrelevant ones because the ideal count changes from query to query. It introduces an adaptive approach that decides on the fly how many tables to keep by applying a learned threshold and a sliding-window reranker that efficiently scans large collections. Experiments on the Spider, BIRD, and Spider 2.0 benchmarks show gains in both table-retrieval accuracy and the quality of the final text-to-SQL outputs that use the retrieved tables. The central claim is therefore that letting the retrieval process adjust its output size to the evidence needs of each individual query removes a systematic source of error that fixed-k methods cannot avoid.

Core claim

The authors present an adaptive table retrieval method that employs an adaptive thresholding mechanism to select tables whose similarity to the query exceeds a dynamically determined cutoff, combined with a sliding-window reranking algorithm that processes large table corpora without exhaustive scoring. This replaces the conventional top-k strategy, which enforces a single predetermined number of tables for every query regardless of how many are actually required.

What carries the argument

Adaptive thresholding mechanism that sets a per-query cutoff on table similarity scores, paired with sliding-window reranking to handle large corpora efficiently.

If this is right

  • Retrieval recall improves because queries that need more than k tables are no longer truncated.
  • Downstream text-to-SQL accuracy rises on Spider, BIRD, and Spider 2.0 because the input to the SQL generator contains fewer irrelevant tables.
  • The same adaptive selector can be applied to any retrieval task in which the optimal result cardinality is query-dependent.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The approach could be tested on retrieval over knowledge graphs or document collections where the number of relevant items also varies widely.
  • If the threshold can be predicted from query features alone, the method might run faster by avoiding full similarity computation for clearly irrelevant tables.
  • Integrating the adaptive selector into an end-to-end differentiable pipeline could allow the SQL model itself to influence how many tables are retrieved.

Load-bearing premise

The number of tables actually needed to answer a query varies from one query to the next and cannot be known ahead of time, so a threshold-based selector can reliably recover the right variable-sized set.

What would settle it

On a held-out set of queries where the minimal sufficient table set is known in advance, measure whether the adaptive method's chosen count matches or exceeds the accuracy of the best fixed-k baseline for the same queries.

Figures

Figures reproduced from arXiv: 2605.18766 by Jaegul Choo, Jihwan Kim, Seungbin Yang, Taehee Kim.

Figure 1
Figure 1. Figure 1: Rather than rely on a rigid fixed k retrieval strategy, ATR retrieves only relevant tables. Gray indi￾cates tables required by the query but not retrieved, red denotes irrelevant tables, and blue highlights retrieved relevant tables. bles (Lewis et al., 2020; Pan et al., 2022; Kothyari et al., 2023; Kang et al., 2024; Kong et al., 2024). Existing table retrieval methods compute query￾table similarity and s… view at source ↗
Figure 2
Figure 2. Figure 2: Retrieving irrelevant tables introduces noise, [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overview of the ATR framework. (A) Inference: ATR takes a query and candidate tables as input to [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of execution accuracy and average token length for the text-to-SQL task across different [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Execution accuracy on the Spider 2.0 dataset [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Analysis of training loss hyper-parameters. Irr. indicates the number of retrieved tables irrelevant to the [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: An illustrative example of the sliding window reranking process in ATR with four input tables, window [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Prompt template for Spider and BIRD datasets. [PITH_FULL_IMAGE:figures/full_fig_p021_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Prompt template for Spider 2.0 (BigQuery dialect) [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Prompt template for Spider 2.0 (Snowflake dialect) [PITH_FULL_IMAGE:figures/full_fig_p023_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Prompt template for Spider 2.0 (SQLite dialect) [PITH_FULL_IMAGE:figures/full_fig_p024_11.png] view at source ↗
read the original abstract

Retrieving relevant tables from extensive databases for a given natural language query is essential for accurately answering questions in tasks such as text-to-SQL. Existing table retrieval approaches select a pre-determined set of k tables with the highest similarity to the query. However, the number of required tables varies across queries and cannot be known in advance. Enforcing a fixed number of retrieved tables regardless of the query may either retrieve an undersized set, failing to obtain all necessary evidence, or retrieve an oversized pool, including irrelevant tables. To address this issue, we propose an adaptive table retrieval method that adjusts the number of tables retrieved according to the requirements of each query. Specifically, we utilize an adaptive thresholding mechanism to selectively retrieve tables and integrate a sliding-window reranking algorithm to efficiently process a large table corpus. Extensive experiments on Spider, BIRD, and Spider 2.0 demonstrate that our method effectively addresses the limitations of the top-k retrieval strategy, improving performance in retrieval and downstream tasks. Our code and data are available at https://github.com/sbY99/Adaptive-Table-Retrieval.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper claims that fixed top-k table retrieval is suboptimal for text-to-SQL because the number of relevant tables varies per query. It proposes an adaptive retrieval method that uses an adaptive thresholding mechanism to select a variable number of tables per query, combined with a sliding-window reranking step to handle large corpora. Experiments on Spider, BIRD, and Spider 2.0 are said to show gains in both retrieval metrics and downstream task performance over standard top-k baselines.

Significance. If the adaptive thresholding rule can be shown to infer the correct variable cardinality directly from each query's similarity distribution without dataset-tuned cutoffs or training-split fitting, the method would address a genuine limitation of fixed-k retrieval in database question answering. Reproducible code is provided, which strengthens the potential impact if the core mechanism proves robust across query distributions.

major comments (3)
  1. Method section (adaptive thresholding description): the paper must specify the exact rule used to set the per-query threshold (e.g., similarity percentile, gap statistic, or learned parameter). If the threshold is determined by any quantity fitted on the training split or held constant across datasets, the 'adaptive' claim reduces to an indirect selection of effective k and does not solve the stated problem for queries whose required table count lies outside the observed training range.
  2. Experiments section (results on Spider/BIRD/Spider 2.0): the abstract and any reported tables must include concrete retrieval metrics (e.g., recall@variable-k, precision, or F1) with error bars or statistical significance tests. Without these numbers, the central claim that the method 'effectively addresses the limitations of the top-k retrieval strategy' lacks load-bearing quantitative support.
  3. Ablation or analysis subsection: an explicit test is needed showing that performance gains persist when the thresholding rule is frozen to a single global value (or when the rule is applied to a held-out dataset with different table-count distribution). Absence of such a control leaves open the possibility that gains arise from dataset-specific tuning rather than query-driven adaptation.
minor comments (2)
  1. Abstract: replace the qualitative statement 'improving performance in retrieval and downstream tasks' with at least one concrete metric (e.g., 'improves retrieval recall by X% and execution accuracy by Y%').
  2. Notation: define the similarity function and the sliding-window reranking procedure with explicit equations or pseudocode so that the adaptive threshold can be reproduced from the text alone.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments. We have addressed each major point below and revised the manuscript to improve clarity, specificity, and empirical support for our claims.

read point-by-point responses

  1. Referee: Method section (adaptive thresholding description): the paper must specify the exact rule used to set the per-query threshold (e.g., similarity percentile, gap statistic, or learned parameter). If the threshold is determined by any quantity fitted on the training split or held constant across datasets, the 'adaptive' claim reduces to an indirect selection of effective k and does not solve the stated problem for queries whose required table count lies outside the observed training range.

    Authors: We agree that an exact specification of the thresholding rule is required to substantiate the adaptive claim. The original manuscript describes the mechanism as selecting tables whose similarity exceeds a per-query threshold derived directly from that query's similarity distribution. In the revision we have added the precise rule: for each query we compute the threshold as the mean of its top-20 similarity scores plus one standard deviation of those scores. This computation uses only the current query's scores and involves no parameters fitted on the training split or held constant across datasets. We have inserted the corresponding equation and pseudocode into Section 3.2. revision: yes

  2. Referee: Experiments section (results on Spider/BIRD/Spider 2.0): the abstract and any reported tables must include concrete retrieval metrics (e.g., recall@variable-k, precision, or F1) with error bars or statistical significance tests. Without these numbers, the central claim that the method 'effectively addresses the limitations of the top-k retrieval strategy' lacks load-bearing quantitative support.

    Authors: We accept that the original presentation relied too heavily on downstream task gains and omitted explicit retrieval metrics. The revised manuscript now reports recall@variable-k, precision, and F1 for the retrieval stage on all three datasets. Each metric is accompanied by standard deviation across five random seeds and paired t-test p-values against the strongest fixed-k baseline. The abstract has been updated to reference these improvements, and the new numbers appear in Tables 2 and 3. revision: yes

  3. Referee: Ablation or analysis subsection: an explicit test is needed showing that performance gains persist when the thresholding rule is frozen to a single global value (or when the rule is applied to a held-out dataset with different table-count distribution). Absence of such a control leaves open the possibility that gains arise from dataset-specific tuning rather than query-driven adaptation.

    Authors: We acknowledge the need for this control. We have added a new ablation (Section 5.4) that freezes the threshold to a single global value obtained by averaging the per-query thresholds on the Spider training split and then evaluates the frozen rule on BIRD and Spider 2.0. The adaptive per-query version continues to outperform the frozen variant on both datasets, with the largest margin on Spider 2.0 whose table-count distribution differs most from Spider. These results are reported with the same retrieval and downstream metrics used in the main experiments. revision: yes

Circularity Check

0 steps flagged

No significant circularity; adaptive thresholding presented as independent mechanism

full rationale

The paper introduces an adaptive thresholding mechanism and sliding-window reranking to handle variable numbers of relevant tables per query, without any quoted equations or self-citations that reduce the core claim to a fitted parameter or self-referential definition. The method is described as directly inferring cardinality from query-specific similarity distributions, and the provided abstract and context show no load-bearing reduction to training-tuned cutoffs or prior author results by construction. This qualifies as a self-contained proposal against external benchmarks like Spider and BIRD.

Axiom & Free-Parameter Ledger

1 free parameters · 0 axioms · 0 invented entities

The adaptive thresholding mechanism likely depends on one or more threshold parameters whose exact fitting or selection process is not detailed in the abstract; no invented entities or additional axioms are explicitly introduced.

free parameters (1)
  • adaptive threshold value
    The mechanism that decides inclusion of tables based on similarity likely requires a tunable or fitted threshold parameter to determine the variable retrieval count.

pith-pipeline@v0.9.0 · 5728 in / 1282 out tokens · 64120 ms · 2026-05-21T01:04:21.155591+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

102 extracted references · 102 canonical work pages · 9 internal anchors

  1. [1]

    Publications Manual , year = "1983", publisher =

    work page 1983

  2. [2]

    Chandra and Dexter C

    Ashok K. Chandra and Dexter C. Kozen and Larry J. Stockmeyer , year = "1981", title =. doi:10.1145/322234.322243

    work page doi:10.1145/322234.322243 1981

  3. [3]

    doi: 10.18653/v1/2023.emnlp-main.495

    Jiang, Zhengbao and Xu, Frank and Gao, Luyu and Sun, Zhiqing and Liu, Qian and Dwivedi-Yu, Jane and Yang, Yiming and Callan, Jamie and Neubig, Graham. Active Retrieval Augmented Generation. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.495

    work page doi:10.18653/v1/2023.emnlp-main.495 2023

  4. [4]

    Findings of the Association for Computational Linguistics: EMNLP 2023 , pages=

    Enhancing Retrieval-Augmented Large Language Models with Iterative Retrieval-Generation Synergy , author=. Findings of the Association for Computational Linguistics: EMNLP 2023 , pages=

    work page 2023

  5. [5]

    Scalable training of

    Andrew, Galen and Gao, Jianfeng , booktitle=. Scalable training of

    work page

  6. [6]

    Dan Gusfield , title =. 1997

    work page 1997

  7. [7]

    Tetreault , title =

    Mohammad Sadegh Rasooli and Joel R. Tetreault , title =. Computing Research Repository , volume =. 2015 , url =

    work page 2015

  8. [8]

    A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =

    Ando, Rie Kubota and Zhang, Tong , Issn =. A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =. Journal of Machine Learning Research , Month = dec, Numpages =

    work page

  9. [9]

    Zero-Shot Cross-Lingual Reranking with Large Language Models for Low-Resource Languages

    Adeyemi, Mofetoluwa and Oladipo, Akintunde and Pradeep, Ronak and Lin, Jimmy. Zero-Shot Cross-Lingual Reranking with Large Language Models for Low-Resource Languages. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) , year =. doi:10.18653/v1/2024.acl-short.59

    work page doi:10.18653/v1/2024.acl-short.59 2024

  10. [10]

    arXiv preprint arXiv:2402.14361 , year=

    Opentab: Advancing large language models as open-domain table reasoners , author=. arXiv preprint arXiv:2402.14361 , year=

    work page arXiv

  11. [11]

    BEAVER: An Enterprise Benchmark for Text-to-SQL

    BEAVER: an enterprise benchmark for text-to-sql , author=. arXiv preprint arXiv:2409.02038 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  12. [12]

    Large Language Models are few(1)-shot Table Reasoners

    Chen, Wenhu. Large Language Models are few(1)-shot Table Reasoners. Findings of the Association for Computational Linguistics: EACL 2023. 2023. doi:10.18653/v1/2023.findings-eacl.83

    work page doi:10.18653/v1/2023.findings-eacl.83 2023

  13. [13]

    The death of schema linking? text-to-sql in the age of well-reasoned language models,

    The death of schema linking? text-to-sql in the age of well-reasoned language models , author=. arXiv preprint arXiv:2408.07702 , year=

    work page arXiv

  14. [14]

    Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval

    Chen, Peter Baile and Zhang, Yi and Roth, Dan. Is Table Retrieval a Solved Problem? Exploring Join-Aware Multi-Table Retrieval. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024. doi:10.18653/v1/2024.acl-long.148

    work page doi:10.18653/v1/2024.acl-long.148 2024

  15. [15]

    Si-An Chen and Lesly Miculicich and Julian Martin Eisenschlos and Zifeng Wang and Zilong Wang and Yanfei Chen and Yasuhisa Fujii and Hsuan-Tien Lin and Chen-Yu Lee and Tomas Pfister , booktitle=. Table. 2024 , url=

    work page 2024

  16. [16]

    FIRST : Faster Improved Listwise Reranking with Single Token Decoding

    Gangi Reddy, Revanth and Doo, JaeHyeok and Xu, Yifei and Sultan, Md Arafat and Swain, Deevya and Sil, Avirup and Ji, Heng. FIRST : Faster Improved Listwise Reranking with Single Token Decoding. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024. doi:10.18653/v1/2024.emnlp-main.491

    work page doi:10.18653/v1/2024.emnlp-main.491 2024

  17. [17]

    2024 , publisher =

    Gao, Dawei and Wang, Haibin and Li, Yaliang and Sun, Xiuyu and Qian, Yichen and Ding, Bolin and Zhou, Jingren , title =. 2024 , publisher =. doi:10.14778/3641204.3641221 , journal =

    work page doi:10.14778/3641204.3641221 2024

  18. [18]

    T a P as: Weakly Supervised Table Parsing via Pre-training

    Herzig, Jonathan and Nowak, Pawel Krzysztof and M. T a P as: Weakly Supervised Table Parsing via Pre-training. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020. doi:10.18653/v1/2020.acl-main.398

    work page doi:10.18653/v1/2020.acl-main.398 2020

  19. [19]

    Tables as Semi-structured Knowledge for Question Answering

    Jauhar, Sujay Kumar and Turney, Peter and Hovy, Eduard. Tables as Semi-structured Knowledge for Question Answering. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2016. doi:10.18653/v1/P16-1045

    work page doi:10.18653/v1/p16-1045 2016

  20. [20]

    ACM Transactions on Intelligent Systems and Technology (TIST) , volume=

    Web table extraction, retrieval, and augmentation: A survey , author=. ACM Transactions on Intelligent Systems and Technology (TIST) , volume=. 2020 , publisher=

    work page 2020

  21. [21]

    A comprehensive evaluation of chatgpt’s zero-shot text-to-sql capability,

    A comprehensive evaluation of ChatGPT's zero-shot Text-to-SQL capability , author=. arXiv preprint arXiv:2303.13547 , year=

    work page arXiv

  22. [22]

    arXiv preprint arXiv:2204.08941 , year=

    Codexdb: Generating code for processing sql queries using gpt-3 codex , author=. arXiv preprint arXiv:2204.08941 , year=

    work page arXiv

  23. [23]

    Evaluating the

    Evaluating the text-to-sql capabilities of large language models , author=. arXiv preprint arXiv:2204.00498 , year=

    work page arXiv

  24. [24]

    Yingqi Gao, Yifu Liu, Xiaoxia Li, Xiaorong Shi, Yin Zhu, Yiming Wang, Shiqi Li, Wei Li, Yun- tao Hong, Zhiling Luo, Jinyang Gao, Liyu Mou, and Yu Li

    Text-to-sql empowered by large language models: A benchmark evaluation , author=. arXiv preprint arXiv:2308.15363 , year=

    work page arXiv

  25. [25]

    Advances in Neural Information Processing Systems , volume=

    Din-sql: Decomposed in-context learning of text-to-sql with self-correction , author=. Advances in Neural Information Processing Systems , volume=

    work page

  26. [26]

    Findings of the Association for Computational Linguistics ACL 2024 , pages=

    Decomposition for Enhancing Attention: Improving LLM-based Text-to-SQL through Workflow Paradigm , author=. Findings of the Association for Computational Linguistics ACL 2024 , pages=

    work page 2024

  27. [27]

    Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    Synthesizing Text-to-SQL Data from Weak and Strong LLMs , author=. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    work page

  28. [28]

    arXiv preprint arXiv:2403.17611 , year=

    Denoising Table-Text Retrieval for Open-Domain Question Answering , author=. arXiv preprint arXiv:2403.17611 , year=

    work page arXiv

  29. [29]

    Open Domain Question Answering over Tables via Dense Retrieval

    Herzig, Jonathan and M. Open Domain Question Answering over Tables via Dense Retrieval. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021. doi:10.18653/v1/2021.naacl-main.43

    work page doi:10.18653/v1/2021.naacl-main.43 2021

  30. [30]

    TABBIE : Pretrained Representations of Tabular Data

    Iida, Hiroshi and Thai, Dung and Manjunatha, Varun and Iyyer, Mohit. TABBIE : Pretrained Representations of Tabular Data. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021. doi:10.18653/v1/2021.naacl-main.270

    work page doi:10.18653/v1/2021.naacl-main.270 2021

  31. [31]

    The Llama 3 Herd of Models

    The llama 3 herd of models , author=. arXiv preprint arXiv:2407.21783 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  32. [32]

    Qwen2.5-Coder Technical Report

    Qwen2. 5-coder technical report , author=. arXiv preprint arXiv:2409.12186 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  33. [33]

    MBA - RAG : a Bandit Approach for Adaptive Retrieval-Augmented Generation through Question Complexity

    Tang, Xiaqiang and Gao, Qiang and Li, Jian and Du, Nan and Li, Qi and Xie, Sihong. MBA - RAG : a Bandit Approach for Adaptive Retrieval-Augmented Generation through Question Complexity. Proceedings of the 31st International Conference on Computational Linguistics. 2025

    work page 2025

  34. [34]

    Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering

    Izacard, Gautier and Grave, Edouard. Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering. Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. 2021. doi:10.18653/v1/2021.eacl-main.74

    work page doi:10.18653/v1/2021.eacl-main.74 2021

  35. [35]

    Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions

    Trivedi, Harsh and Balasubramanian, Niranjan and Khot, Tushar and Sabharwal, Ashish. Interleaving Retrieval with Chain-of-Thought Reasoning for Knowledge-Intensive Multi-Step Questions. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023. doi:10.18653/v1/2023.acl-long.557

    work page doi:10.18653/v1/2023.acl-long.557 2023

  36. [36]

    Advances in neural information processing systems , volume=

    Retrieval-augmented generation for knowledge-intensive nlp tasks , author=. Advances in neural information processing systems , volume=

    work page

  37. [37]

    The Thirteenth International Conference on Learning Representations , year=

    Multi-Field Adaptive Retrieval , author=. The Thirteenth International Conference on Learning Representations , year=

    work page

  38. [38]

    Denoising Table-Text Retrieval for Open-Domain Question Answering

    Kang, Deokhyung and Jung, Baikjin and Kim, Yunsu and Lee, Gary Geunbae. Denoising Table-Text Retrieval for Open-Domain Question Answering. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

    work page 2024

  39. [39]

    2025 , eprint=

    H-STAR: LLM-driven Hybrid SQL-Text Adaptive Reasoning on Tables , author=. 2025 , eprint=

    work page 2025

  40. [40]

    R e2 G : Retrieve, Rerank, Generate

    Glass, Michael and Rossiello, Gaetano and Chowdhury, Md Faisal Mahbub and Naik, Ankita and Cai, Pengshan and Gliozzo, Alfio. R e2 G : Retrieve, Rerank, Generate. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022. doi:10.18653/v1/2022.naacl-main.194

    work page doi:10.18653/v1/2022.naacl-main.194 2022

  41. [41]

    Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) , pages=

    Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexity , author=. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) , pages=

    work page 2024

  42. [42]

    The Twelfth International Conference on Learning Representations , year=

    Self-rag: Learning to retrieve, generate, and critique through self-reflection , author=. The Twelfth International Conference on Learning Representations , year=

    work page

  43. [43]

    1989 , issn =

    Analysis of variance (ANOVA) , journal =. 1989 , issn =. doi:https://doi.org/10.1016/0169-7439(89)80095-4 , url =

    work page doi:10.1016/0169-7439(89)80095-4 1989

  44. [44]

    When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories

    Mallen, Alex and Asai, Akari and Zhong, Victor and Das, Rajarshi and Khashabi, Daniel and Hajishirzi, Hannaneh. When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023. doi:10.18653/v1/2023...

    work page doi:10.18653/v1/2023.acl-long.546 2023

  45. [45]

    Gemma 3 Technical Report

    Gemma 3 technical report , author=. arXiv preprint arXiv:2503.19786 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  46. [46]

    A o E : Angle-optimized Embeddings for Semantic Textual Similarity

    Li, Xianming and Li, Jing. A o E : Angle-optimized Embeddings for Semantic Textual Similarity. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024. doi:10.18653/v1/2024.acl-long.101

    work page doi:10.18653/v1/2024.acl-long.101 2024

  47. [47]

    Unsupervised Dense Information Retrieval with Contrastive Learning

    Unsupervised dense information retrieval with contrastive learning , author=. arXiv preprint arXiv:2112.09118 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  48. [48]

    Transactions on Machine Learning Research , issn=

    Unsupervised Dense Information Retrieval with Contrastive Learning , author=. Transactions on Machine Learning Research , issn=. 2022 , url=

    work page 2022

  49. [49]

    2024 , url=

    Xingyu Ji and Aditya Parameswaran and Madelon Hulsebos , booktitle=. 2024 , url=

    work page 2024

  50. [50]

    CRUSH 4 SQL : Collective Retrieval Using Schema Hallucination For T ext2 SQL

    Kothyari, Mayank and Dhingra, Dhruva and Sarawagi, Sunita and Chakrabarti, Soumen. CRUSH 4 SQL : Collective Retrieval Using Schema Hallucination For T ext2 SQL. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.868

    work page doi:10.18653/v1/2023.emnlp-main.868 2023

  51. [51]

    Adaptive Document Retrieval for Deep Question Answering

    Kratzwald, Bernhard and Feuerriegel, Stefan. Adaptive Document Retrieval for Deep Question Answering. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018. doi:10.18653/v1/D18-1055

    work page doi:10.18653/v1/d18-1055 2018

  52. [52]

    Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-

    Fangyu Lei and Jixuan Chen and Yuxiao Ye and Ruisheng Cao and Dongchan Shin and Hongjin SU and ZHAOQING SUO and Hongcheng Gao and Wenjing Hu and Pengcheng Yin and Victor Zhong and Caiming Xiong and Ruoxi Sun and Qian Liu and Sida Wang and Tao Yu , booktitle=. Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-. 2025 , url=

    work page 2025

  53. [53]

    Jinyang Li and Binyuan Hui and GE QU and Jiaxi Yang and Binhua Li and Bowen Li and Bailin Wang and Bowen Qin and Ruiying Geng and Nan Huo and Xuanhe Zhou and Chenhao Ma and Guoliang Li and Kevin Chang and Fei Huang and Reynold Cheng and Yongbin Li , booktitle=. Can. 2023 , url=

    work page 2023

  54. [54]

    The Death of Schema Linking? Text-to-

    Karime Maamari and Fadhil Abubaker and Daniel Jaroslawicz and Amine Mhedhbi , booktitle=. The Death of Schema Linking? Text-to-. 2024 , url=

    work page 2024

  55. [55]

    arXiv preprint arXiv:2202.08904 , year=

    SGPT: GPT Sentence Embeddings for Semantic Search , author=. arXiv preprint arXiv:2202.08904 , year=

    work page arXiv

  56. [56]

    ISBN 9781713871088

    Pal, Vaishali and Yates, Andrew and Kanoulas, Evangelos and de Rijke, Maarten. M ulti T ab QA : Generating Tabular Answers for Multi-Table Question Answering. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023. doi:10.18653/v1/2023.acl-long.348

    work page doi:10.18653/v1/2023.acl-long.348 2023

  57. [57]

    CHESS: Contextual Harnessing for Efficient SQL Synthesis

    Chess: Contextual harnessing for efficient sql synthesis , author=. arXiv preprint arXiv:2405.16755 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  58. [58]

    2023 , url=

    Mohammadreza Pourreza and Davood Rafiei , booktitle=. 2023 , url=

    work page 2023

  59. [59]

    2025 , url=

    Mohammadreza Pourreza and Hailong Li and Ruoxi Sun and Yeounoh Chung and Shayan Talaei and Gaurav Tarlok Kakkar and Yu Gan and Amin Saberi and Fatma Ozcan and Sercan O Arik , booktitle=. 2025 , url=

    work page 2025

  60. [60]

    doi: 10.18653/v1/2024.findings-naacl.97

    Qin, Zhen and Jagerman, Rolf and Hui, Kai and Zhuang, Honglei and Wu, Junru and Yan, Le and Shen, Jiaming and Liu, Tianqi and Liu, Jialu and Metzler, Donald and Wang, Xuanhui and Bendersky, Michael. Large Language Models are Effective Text Rankers with Pairwise Ranking Prompting. Findings of the Association for Computational Linguistics: NAACL 2024. 2024....

    work page doi:10.18653/v1/2024.findings-naacl.97 2024

  61. [61]

    ArXiv , year=

    TQA-Bench: Evaluating LLMs for Multi-Table Question Answering with Scalable Context and Symbolic Extension , author=. ArXiv , year=

    work page

  62. [62]

    2024 , url=

    Dongyu Ru and Lin Qiu and Xiangkun Hu and Tianhang Zhang and Peng Shi and Shuaichen Chang and Cheng Jiayang and Cunxiang Wang and Shichao Sun and Huanyu Li and Zizhao Zhang and Binjie Wang and Jiarong Jiang and Tong He and Zhiguo Wang and Pengfei Liu and Yue Zhang and Zheng Zhang , booktitle=. 2024 , url=

    work page 2024

  63. [63]

    Improving Passage Retrieval with Zero-Shot Question Generation

    Sachan, Devendra and Lewis, Mike and Joshi, Mandar and Aghajanyan, Armen and Yih, Wen-tau and Pineau, Joelle and Zettlemoyer, Luke , booktitle =. Improving Passage Retrieval with Zero-Shot Question Generation. 2022. doi:10.18653/v1/2022.emnlp-main.249

    work page doi:10.18653/v1/2022.emnlp-main.249 2022

  64. [64]

    arXiv preprint arXiv:2203.16714 , year=

    End-to-end table question answering via retrieval-augmented generation , author=. arXiv preprint arXiv:2203.16714 , year=

    work page arXiv

  65. [65]

    Is C hat GPT Good at Search? Investigating Large Language Models as Re-Ranking Agents

    Sun, Weiwei and Yan, Lingyong and Ma, Xinyu and Wang, Shuaiqiang and Ren, Pengjie and Chen, Zhumin and Yin, Dawei and Ren, Zhaochun. Is C hat GPT Good at Search? Investigating Large Language Models as Re-Ranking Agents. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.923

    work page doi:10.18653/v1/2023.emnlp-main.923 2023

  66. [66]

    and Chopra, S

    Hadsell, R. and Chopra, S. and LeCun, Y. , booktitle=. Dimensionality Reduction by Learning an Invariant Mapping , year=

    work page

  67. [67]

    Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    The power of noise: Redefining retrieval for rag systems , author=. Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    work page

  68. [68]

    ArXiv , year=

    CHESS: Contextual Harnessing for Efficient SQL Synthesis , author=. ArXiv , year=

    work page

  69. [69]

    Table Retrieval May Not Necessitate Table-specific Model Design

    Wang, Zhiruo and Jiang, Zhengbao and Nyberg, Eric and Neubig, Graham. Table Retrieval May Not Necessitate Table-specific Model Design. Proceedings of the Workshop on Structured and Unstructured Knowledge Integration (SUKI). 2022. doi:10.18653/v1/2022.suki-1.5

    work page doi:10.18653/v1/2022.suki-1.5 2022

  70. [70]

    2025 , url=

    Jian Wu and Linyi Yang and Dongyuan Li and Yuliang Ji and Manabu Okumura and Yue Zhang , booktitle=. 2025 , url=

    work page 2025

  71. [71]

    Advances in Neural Information Processing Systems , volume=

    Tablerag: Million-token table understanding with language models , author=. Advances in Neural Information Processing Systems , volume=

    work page

  72. [72]

    Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning

    Seq2sql: Generating structured queries from natural language using reinforcement learning , author=. arXiv preprint arXiv:1709.00103 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  73. [73]

    Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    Retrieving complex tables with multi-granular graph representation learning , author=. Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    work page

  74. [74]

    NO INSIGHT

    Yu, Tao and Zhang, Rui and Yang, Kai and Yasunaga, Michihiro and Wang, Dongxu and Li, Zifan and Ma, James and Li, Irene and Yao, Qingning and Roman, Shanelle and Zhang, Zilin and Radev, Dragomir. S pider: A Large-Scale Human-Labeled Dataset for Complex and Cross-Domain Semantic Parsing and Text-to- SQL Task. Proceedings of the 2018 Conference on Empirical...

    work page doi:10.18653/v1/d18-1425 2018

  75. [75]

    MURRE : Multi-Hop Table Retrieval with Removal for Open-Domain Text-to- SQL

    Zhang, Xuanliang and Wang, Dingzirui and Dou, Longxu and Zhu, Qingfu and Che, Wanxiang. MURRE : Multi-Hop Table Retrieval with Removal for Open-Domain Text-to- SQL. Proceedings of the 31st International Conference on Computational Linguistics. 2025

    work page 2025

  76. [76]

    2024 , eprint=

    Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference , author=. 2024 , eprint=

    work page 2024

  77. [77]

    2019 , isbn =

    Zhang, Li and Zhang, Shuo and Balog, Krisztian , title =. 2019 , isbn =. doi:10.1145/3331184.3331333 , booktitle =

    work page doi:10.1145/3331184.3331333 2019

  78. [78]

    Proceedings of the AAAI conference on artificial intelligence , volume=

    Document-level relation extraction with adaptive thresholding and localized context pooling , author=. Proceedings of the AAAI conference on artificial intelligence , volume=

    work page

  79. [79]

    Beyond Yes and No: Improving Zero-Shot LLM Rankers via Scoring Fine-Grained Relevance Labels

    Zhuang, Honglei and Qin, Zhen and Hui, Kai and Wu, Junru and Yan, Le and Wang, Xuanhui and Bendersky, Michael. Beyond Yes and No: Improving Zero-Shot LLM Rankers via Scoring Fine-Grained Relevance Labels. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2...

    work page doi:10.18653/v1/2024.naacl-short.31 2024

  80. [80]

    2024 , journal=

    Xueguang Ma and Xinyu Zhang and Ronak Pradeep and Jimmy Lin , title =. 2024 , journal=

    work page 2024

Showing first 80 references.