pith. machine review for the scientific record. sign in

arxiv: 2604.17653 · v1 · submitted 2026-04-19 · 💻 cs.AI · cs.DB

Recognition: unknown

PV-SQL: Synergizing Database Probing and Rule-based Verification for Text-to-SQL Agents

Yuan Tian , Tianyi Zhang
Authors on Pith no claims yet

Pith reviewed 2026-05-10 05:19 UTC · model grok-4.3

classification 💻 cs.AI cs.DB
keywords text-to-SQLagentic frameworkdatabase probingrule-based verificationSQL generationBIRD benchmarkcontextual understanding
0
0 comments X

The pith

PV-SQL combines iterative database probing with rule-based verification to improve text-to-SQL performance on complex queries.

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

The paper presents PV-SQL as an agentic framework to overcome limitations in text-to-SQL systems when handling queries with subtle or ambiguous requirements. It introduces a Probe component that generates iterative queries to fetch actual database records, clarifying details like value formats, column meanings, and table connections. A Verify component then applies rules to extract conditions and create a checklist that guides refinement of the SQL query to avoid missing constraints. Experiments demonstrate gains over prior methods on standard benchmarks while using fewer resources, suggesting that blending data exploration with structured checking strengthens automated query generation.

Core claim

PV-SQL is an agentic framework consisting of a Probe component that iteratively generates probing queries to retrieve concrete records from the database, resolving ambiguities in value formats, column semantics, and inter-table relationships, and a Verify component that employs a rule-based method to extract verifiable conditions and construct an executable checklist for iterative SQL refinement.

What carries the argument

The Probe and Verify components, where Probe builds richer context by retrieving concrete database records through targeted queries and Verify enforces completeness by turning extracted conditions into a checklist for SQL iteration.

If this is right

  • Outperforms the best text-to-SQL baseline by 5% in execution accuracy on BIRD benchmarks.
  • Improves valid efficiency score by 20.8% while using fewer tokens.
  • Reduces missing constraints in generated SQL through iterative checklist refinement.
  • Handles complex queries with subtle requirements more effectively than generation-only approaches.

Where Pith is reading between the lines

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

  • The approach could extend to other structured generation tasks where LLMs need to ground outputs in external data sources before finalizing results.
  • Hybrid agent designs that separate exploration of the environment from verification of constraints may reduce errors in domains like code synthesis or data extraction.
  • Real-world database interfaces might become more robust if probing and rule checks are integrated to handle incomplete user specifications or evolving schemas.

Load-bearing premise

Iteratively generating probing queries will reliably resolve ambiguities in value formats, column semantics, and inter-table relationships, and the rule-based checklist will consistently reduce missing constraints without introducing new errors or excessive overhead.

What would settle it

Experiments on the BIRD benchmark showing no gain in execution accuracy or valid efficiency score over the best baseline, or cases where probing queries produce misleading records that lead to incorrect SQL.

Figures

Figures reproduced from arXiv: 2604.17653 by Tianyi Zhang, Yuan Tian.

Figure 1
Figure 1. Figure 1: An example of how PV-SQL effectively solves a text-to-SQL task. abbreviated as “CA”? How is “late” represented? Does the database have a “late” flag? Answering these questions requires examining actual database values, yet existing methods typically rely only on schema descriptions, which do not contain such information.2 Our empirical study (Section 3.2) re￾veals that approximately 41% of failed tasks ste… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of PV-SQL. Left: The agent generates probing SQL to discover database content and enriches the context with insights. Top: Rule-based extraction identifies semantic constraints from the question. Right: A rule-based verifier provides feedback for iterative repair until all checks pass. 3. Constraint Check: Unlike syntax and execu￾tion errors that databases catch automatically, se￾mantic constraint… view at source ↗
Figure 4
Figure 4. Figure 4: Error distribution w/ and w/o PV-SQL. 6.5 Efficiency Analysis [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 3
Figure 3. Figure 3: Accuracy vs. token consumption6 on BIRD [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Case study showing how probing discovers [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
read the original abstract

Text-to-SQL systems often struggle with deep contextual understanding, particularly for complex queries with subtle requirements. We present PV-SQL, an agentic framework that addresses these failures through two complementary components: Probe and Verify. The Probe component iteratively generates probing queries to retrieve concrete records from the database, resolving ambiguities in value formats, column semantics, and inter-table relationships to build richer contextual understanding. The Verify component employs a rule-based method to extract verifiable conditions and construct an executable checklist, enabling iterative SQL refinement that effectively reduces missing constraints. Experiments on the BIRD benchmarks show that PV-SQL outperforms the best text-to-SQL baseline by 5% in execution accuracy and 20.8% in valid efficiency score while consuming fewer tokens.

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

2 major / 2 minor

Summary. The paper introduces PV-SQL, an agentic Text-to-SQL framework with two components: Probe, which iteratively generates database probing queries to resolve ambiguities in value formats, column semantics, and inter-table relationships; and Verify, which applies a rule-based method to extract conditions and build an executable checklist for iterative SQL refinement. Experiments on the BIRD benchmark report that PV-SQL outperforms the best baseline by 5% in execution accuracy and 20.8% in valid efficiency score while using fewer tokens.

Significance. If the reported gains hold under rigorous evaluation, the work demonstrates a practical synergy between database probing for richer context and rule-based verification for constraint coverage, addressing common failure modes in complex Text-to-SQL queries. This could inform future agentic systems by showing that lightweight, non-LLM mechanisms can improve accuracy and efficiency without increasing token costs.

major comments (2)
  1. [Experiments / §4] The central performance claims (5% execution accuracy and 20.8% valid efficiency gains) are load-bearing for the paper's contribution, yet the abstract and experimental description provide no details on the exact baselines, their configurations, statistical significance testing, error bars, or full experimental conditions (e.g., number of runs, prompt templates). This makes it difficult to assess whether the improvements are robust or attributable to the Probe+Verify mechanisms.
  2. [Method / §3] The assumption that iterative probing reliably resolves ambiguities without introducing new errors or excessive overhead is central to the Probe component, but the manuscript does not appear to include ablation studies isolating the contribution of probing iterations versus the Verify checklist, nor analysis of failure cases where probing might degrade performance.
minor comments (2)
  1. [§3.2] Clarify the exact rule set used in the Verify component and how the checklist is converted into executable verification steps; this would aid reproducibility.
  2. [Table 1 or §4.2] The BIRD benchmark results would benefit from a table breaking down performance by query complexity or error type to show where the gains are concentrated.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive assessment and recommendation for minor revision. We address each major comment below with specific revisions to improve clarity, reproducibility, and analysis of the proposed components.

read point-by-point responses

  1. Referee: [Experiments / §4] The central performance claims (5% execution accuracy and 20.8% valid efficiency gains) are load-bearing for the paper's contribution, yet the abstract and experimental description provide no details on the exact baselines, their configurations, statistical significance testing, error bars, or full experimental conditions (e.g., number of runs, prompt templates). This makes it difficult to assess whether the improvements are robust or attributable to the Probe+Verify mechanisms.

    Authors: We agree that greater experimental transparency is needed. In the revised manuscript, Section 4 and the appendix have been expanded to specify: all baselines (including their exact model versions, prompting strategies, and hyper-parameters drawn from their original papers); the full prompt templates for Probe and Verify; the number of runs (three independent runs with different random seeds, reporting mean ± std); and statistical significance via paired t-tests (p < 0.05 for both metrics). Error bars are now shown in Table 1. These additions confirm that the reported gains are robust and directly attributable to the synergy of Probe and Verify rather than implementation variance. revision: yes

  2. Referee: [Method / §3] The assumption that iterative probing reliably resolves ambiguities without introducing new errors or excessive overhead is central to the Probe component, but the manuscript does not appear to include ablation studies isolating the contribution of probing iterations versus the Verify checklist, nor analysis of failure cases where probing might degrade performance.

    Authors: We acknowledge the value of component-level analysis. The revised Section 4.3 now includes a dedicated ablation study comparing four variants: full PV-SQL, Probe-only, Verify-only, and the strongest baseline. Results show that Probe contributes the larger share of the accuracy improvement while Verify primarily boosts valid efficiency; their combination yields the best overall score. We have also added a failure-case analysis subsection that quantifies cases where probing introduced minor overhead or transient errors (occurring in <6% of BIRD queries) and demonstrates how the Verify checklist mitigates them. Token overhead remains lower than all baselines even with up to three probing iterations. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper describes an empirical agentic framework (Probe for iterative database probing and Verify for rule-based checklist refinement) evaluated directly on the external BIRD benchmark. Reported gains (5% execution accuracy, 20.8% valid efficiency) are experimental outcomes against baselines, not quantities derived from internal equations, fitted parameters, or self-citation chains that reduce to the inputs by construction. No mathematical derivation, ansatz, or uniqueness theorem is invoked; the work is self-contained against external data.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No mathematical derivations, fitted parameters, or new postulated entities are described; the work is an empirical agentic system evaluated on benchmarks.

pith-pipeline@v0.9.0 · 5414 in / 1141 out tokens · 25554 ms · 2026-05-10T05:19:16.215235+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

65 extracted references · 15 canonical work pages · 1 internal anchor

  1. [1]

    S pider: A Large-Scale Human-Labeled Dataset for Complex and Cross-Domain Semantic Parsing and Text-to- SQL Task

    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...

    2018

  2. [2]

    Can LLM Already Serve as A Database Interface? A B ig Bench for Large-Scale Database Grounded Text-to- SQL s

    Li, Jinyang and Hui, Binyuan and Qu, Ge and Yang, Jiaxi and Li, Binhua and Li, Bowen and Wang, Bailin and Qin, Bowen and Geng, Ruiying and Huo, Nan and Zhou, Xuanhe and Ma, Chenhao and Li, Guoliang and Chang, Kevin Chen-Chuan and Li, Fei and Hui, Bei and Li, Yongbin. Can LLM Already Serve as A Database Interface? A B ig Bench for Large-Scale Database Grou...

    2024

  3. [3]

    2025 , eprint=

    E-SQL: Direct Schema Linking via Question Enrichment in Text-to-SQL , author=. 2025 , eprint=

    2025

  4. [4]

    2022 , eprint=

    Evaluating the Text-to-SQL Capabilities of Large Language Models , author=. 2022 , eprint=

    2022

  5. [5]

    Re-examining the Role of Schema Linking in Text-to- SQL

    Lei, Wenqiang and Wang, Weixin and Ma, Zhixin and Gan, Tian and Lu, Wei and Kan, Min-Yen and Chua, Tat-Seng. Re-examining the Role of Schema Linking in Text-to- SQL. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). 2020. doi:10.18653/v1/2020.emnlp-main.564

    work page doi:10.18653/v1/2020.emnlp-main.564 2020

  6. [6]

    Li, Haoyang and Zhang, Jing and Li, Cuiping and Chen, Hong , title =. Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence and Thirteenth Symposium on Educational Advances in Artificial Intelligence , articleno =. 2023 , isbn =. doi:10.1609/aaai.v37i1...

    work page doi:10.1609/aaai.v37i11.26535 2023

  7. [7]

    Bridging Textual and Tabular Data for Cross-Domain Text-to- SQL Semantic Parsing

    Lin, Xi Victoria and Socher, Richard and Xiong, Caiming. Bridging Textual and Tabular Data for Cross-Domain Text-to- SQL Semantic Parsing. Findings of the Association for Computational Linguistics: EMNLP 2020. 2020. doi:10.18653/v1/2020.findings-emnlp.438

    work page doi:10.18653/v1/2020.findings-emnlp.438 2020

  8. [8]

    Self-Consistency Improves Chain of Thought Reasoning in Language Models

    Self-Consistency Improves Chain of Thought Reasoning in Language Models , author=. arXiv preprint arXiv:2203.11171 , year=

    work page Pith review arXiv

  9. [9]

    2023 , eprint=

    Self-Refine: Iterative Refinement with Self-Feedback , author=. 2023 , eprint=

    2023

  10. [10]

    2023 , eprint=

    Teaching Large Language Models to Self-Debug , author=. 2023 , eprint=

    2023

  11. [11]

    and Lin, Xi Victoria , title =

    Ni, Ansong and Iyer, Srini and Radev, Dragomir and Stoyanov, Ves and Yih, Wen-tau and Wang, Sida I. and Lin, Xi Victoria , title =. Proceedings of the 40th International Conference on Machine Learning , articleno =. 2023 , publisher =

    2023

  12. [12]

    2024 , eprint=

    Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters , author=. 2024 , eprint=

    2024

  13. [13]

    and Le, Quoc V

    Wei, Jason and Wang, Xuezhi and Schuurmans, Dale and Bosma, Maarten and Ichter, Brian and Xia, Fei and Chi, Ed H. and Le, Quoc V. and Zhou, Denny , title =. Proceedings of the 36th International Conference on Neural Information Processing Systems , articleno =. 2022 , isbn =

    2022

  14. [14]

    and Cao, Yuan and Narasimhan, Karthik , title =

    Yao, Shunyu and Yu, Dian and Zhao, Jeffrey and Shafran, Izhak and Griffiths, Thomas L. and Cao, Yuan and Narasimhan, Karthik , title =. Proceedings of the 37th International Conference on Neural Information Processing Systems , articleno =. 2023 , publisher =

    2023

  15. [15]

    DAIL - SQL : Optimized LLM Prompt for Text-to- SQL

    Gao, Dawei and Wang, Haibin and Li, Yaliang and Sun, Xiuyu and Qian, Yichen and Ding, Bolin and Zhou, Jingren. DAIL - SQL : Optimized LLM Prompt for Text-to- SQL. Proceedings of the VLDB Endowment. 2024

    2024

  16. [16]

    2023 , eprint=

    DIN-SQL: Decomposed In-Context Learning of Text-to-SQL with Self-Correction , author=. 2023 , eprint=

    2023

  17. [17]

    2024 , eprint=

    MAC-SQL: A Multi-Agent Collaborative Framework for Text-to-SQL , author=. 2024 , eprint=

    2024

  18. [18]

    2025 , eprint=

    Chain-of-Query: Unleashing the Power of LLMs in SQL-Aided Table Understanding via Multi-Agent Collaboration , author=. 2025 , eprint=

    2025

  19. [19]

    2024 , eprint=

    A Survey of Text-to-SQL Methods Enhanced by Large Language Models , author=. 2024 , eprint=

    2024

  20. [20]

    2024 , eprint=

    Evaluating SQL Understanding in Large Language Models , author=. 2024 , eprint=

    2024

  21. [21]

    2024 , eprint=

    RSL-SQL: Robust Schema Linking for Text-to-SQL , author=. 2024 , eprint=

    2024

  22. [22]

    2023 , eprint=

    Enhancing Text-to-SQL Capabilities of Large Language Models: A Study on Prompt Design Strategies , author=. 2023 , eprint=

    2023

  23. [23]

    2024 , eprint=

    PURPLE: Making a Large Language Model a Better SQL Writer , author=. 2024 , eprint=

    2024

  24. [24]

    JAMIA Open , volume=

    Automating Pharmacovigilance Evidence Generation: Using Large Language Models to Produce Context-Aware Structured Query Language , author=. JAMIA Open , volume=. 2025 , publisher=

    2025

  25. [25]

    Proceedings of the VLDB Endowment , year=

    ValueNet: A Neural Text-to-SQL Architecture Incorporating Values , author=. Proceedings of the VLDB Endowment , year=

  26. [26]

    2025 , eprint=

    A Study of In-Context-Learning-Based Text-to-SQL Errors , author=. 2025 , eprint=

    2025

  27. [27]

    2024 , eprint=

    Large Language Models Cannot Self-Correct Reasoning Yet , author=. 2024 , eprint=

    2024

  28. [28]

    2022 , eprint=

    CodeT: Code Generation with Generated Tests , author=. 2022 , eprint=

    2022

  29. [29]

    Science , volume=

    Competition-Level Code Generation with AlphaCode , author=. Science , volume=. 2022 , publisher=

    2022

  30. [30]

    Proceedings of the Fifth ACM SIGPLAN International Conference on Functional Programming , pages=

    QuickCheck: A Lightweight Tool for Random Testing of Haskell Programs , author=. Proceedings of the Fifth ACM SIGPLAN International Conference on Functional Programming , pages=

  31. [31]

    and Huang, Fei and Cheng, Reynold and Li, Yongbin , booktitle=

    Li, Jinyang and Hui, Binyuan and Qu, Ge and Yang, Jiaxi and Li, Binhua and Li, Bowen and Wang, Bailin and Qin, Bowen and Cao, Rongyu and Geng, Ruiying and Huo, Nan and Zhou, Xuanhe and Ma, Chenhao and Li, Guoliang and Chang, Kevin C.-C. and Huang, Fei and Cheng, Reynold and Li, Yongbin , booktitle=

  32. [32]

    , booktitle=

    Xu, Wenbo and Zhu, Haifeng and Yan, Liang and Liu, Chuanyi and Han, Peiyi and Duan, Shaoming and Pan, Jeff Z. , booktitle=

  33. [33]

    Pourreza, Mohammadreza and Rafiei, Davood , journal=

  34. [34]

    arXiv preprint arXiv:2409.16751 , year=

    Cafero. arXiv preprint arXiv:2409.16751 , year=

    work page arXiv

  35. [35]

    Askari, Arian and Poelitz, Christian and Tang, Xinye , journal=

  36. [36]

    Cen, Jipeng and Liu, Jiaxin and Li, Zhixu and Wang, Jingjing , journal=

  37. [37]

    Shi, Jie and Xu, Bo and Liang, Jiaqing and Xiao, Yanghua and Chen, Jia and Xie, Chenhao and Wang, Peng and Wang, Wei , booktitle=

  38. [38]

    Teaching Large Language Models to Self-Debug

    Teaching Large Language Models to Self-Debug , author=. arXiv preprint arXiv:2304.05128 , year=

    work page internal anchor Pith review arXiv

  39. [39]

    Li, Zhenwen and Xie, Tao , journal=. Using

  40. [40]

    Advances in Neural Information Processing Systems , volume=

    Self-Refine: Iterative Refinement with Self-Feedback , author=. Advances in Neural Information Processing Systems , volume=

  41. [41]

    2025 , eprint=

    SQL-of-Thought: Multi-agentic Text-to-SQL with Guided Error Correction , author=. 2025 , eprint=

    2025

  42. [42]

    2024 , eprint=

    Before Generation, Align it! A Novel and Effective Strategy for Mitigating Hallucinations in Text-to-SQL Generation , author=. 2024 , eprint=

    2024

  43. [43]

    2024 , eprint=

    CHASE-SQL: Multi-Path Reasoning and Preference Optimized Candidate Selection in Text-to-SQL , author=. 2024 , eprint=

    2024

  44. [44]

    2024 , eprint=

    CHESS: Contextual Harnessing for Efficient SQL Synthesis , author=. 2024 , eprint=

    2024

  45. [45]

    2025 , eprint=

    XiYan-SQL: A Novel Multi-Generator Framework For Text-to-SQL , author=. 2025 , eprint=

    2025

  46. [46]

    2025 , eprint=

    AmbiSQL: Interactive Ambiguity Detection and Resolution for Text-to-SQL , author=. 2025 , eprint=

    2025

  47. [47]

    2025 , eprint=

    A Survey of Text-to-SQL in the Era of LLMs: Where are we, and where are we going? , author=. 2025 , eprint=

    2025

  48. [48]

    2025 , eprint=

    SQLens: An End-to-End Framework for Error Detection and Correction in Text-to-SQL , author=. 2025 , eprint=

    2025

  49. [49]

    2025 , eprint=

    RubikSQL: Lifelong Learning Agentic Knowledge Base as an Industrial NL2SQL System , author=. 2025 , eprint=

    2025

  50. [50]

    TTD - SQL : Tree-Guided Token Decoding for Efficient and Schema-Aware SQL Generation

    Sharma, Chetan and Narayanam, Ramasuri and Pal, Soumyabrata and Yeturu, Kalidas and Saini, Shiv Kumar and Mukherjee, Koyel. TTD - SQL : Tree-Guided Token Decoding for Efficient and Schema-Aware SQL Generation. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track. 2025. doi:10.18653/v1/2025.emnlp-industry.90

    work page doi:10.18653/v1/2025.emnlp-industry.90 2025

  51. [51]

    Jihyung Lee, Jin-Seop Lee, Jaehoon Lee, YunSeok Choi, and Jee-Hyong Lee

    Lee, Jihyung and Lee, Jin-Seop and Lee, Jaehoon and Choi, YunSeok and Lee, Jee-Hyong. DCG - SQL : Enhancing In-Context Learning for Text-to- SQL with Deep Contextual Schema Link Graph. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025. doi:10.18653/v1/2025.acl-long.748

    work page doi:10.18653/v1/2025.acl-long.748 2025

  52. [52]

    Small LLM s Are Weak Tool Learners: A Multi- LLM Agent

    Shen, Weizhou and Li, Chenliang and Chen, Hongzhan and Yan, Ming and Quan, Xiaojun and Chen, Hehong and Zhang, Ji and Huang, Fei. Small LLM s Are Weak Tool Learners: A Multi- LLM Agent. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024. doi:10.18653/v1/2024.emnlp-main.929

    work page doi:10.18653/v1/2024.emnlp-main.929 2024

  53. [53]

    and Zhang, Hao and Gonzalez, Joseph E

    Zheng, Lianmin and Chiang, Wei-Lin and Sheng, Ying and Zhuang, Siyuan and Wu, Zhanghao and Zhuang, Yonghao and Lin, Zi and Li, Zhuohan and Li, Dacheng and Xing, Eric P. and Zhang, Hao and Gonzalez, Joseph E. and Stoica, Ion , title =. Proceedings of the 37th International Conference on Neural Information Processing Systems , articleno =. 2023 , publisher =

    2023

  54. [54]

    2025 , eprint=

    LLM-as-a-qualitative-judge: automating error analysis in natural language generation , author=. 2025 , eprint=

    2025

  55. [55]

    heightened creativity at dawn

    Tian, Yuan and Kummerfeld, Jonathan K. and Li, Toby Jia-Jun and Zhang, Tianyi , title =. Proceedings of the 37th Annual ACM Symposium on User Interface Software and Technology , articleno =. 2024 , isbn =. doi:10.1145/3654777.3676368 , abstract =

    work page doi:10.1145/3654777.3676368 2024

  56. [56]

    and Zhang, Tianyi

    Tian, Yuan and Zhang, Zheng and Ning, Zheng and Li, Toby Jia-Jun and Kummerfeld, Jonathan K. and Zhang, Tianyi. Interactive Text-to- SQL Generation via Editable Step-by-Step Explanations. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.1004

    work page doi:10.18653/v1/2023.emnlp-main.1004 2023

  57. [57]

    Proceedings of the 30th International Conference on Intelligent User Interfaces , pages =

    Tian, Yuan and Lee, Daniel and Wu, Fei and Mai, Tung and Qian, Kun and Sahai, Siddhartha and Zhang, Tianyi and Li, Yunyao , title =. Proceedings of the 30th International Conference on Intelligent User Interfaces , pages =. 2025 , isbn =. doi:10.1145/3708359.3712083 , abstract =

    work page doi:10.1145/3708359.3712083 2025

  58. [58]

    ACM Trans

    Ning, Zheng and Tian, Yuan and Zhang, Zheng and Zhang, Tianyi and Li, Toby Jia-Jun , title =. ACM Trans. Interact. Intell. Syst. , month = dec, articleno =. 2024 , issue_date =. doi:10.1145/3650114 , abstract =

    work page doi:10.1145/3650114 2024

  59. [59]

    URL http://dx.doi.org/10.1145/3581641.3584066

    Ning, Zheng and Zhang, Zheng and Sun, Tianyi and Tian, Yuan and Zhang, Tianyi and Li, Toby Jia-Jun , title =. Proceedings of the 28th International Conference on Intelligent User Interfaces , pages =. 2023 , isbn =. doi:10.1145/3581641.3584067 , abstract =

    work page doi:10.1145/3581641.3584067 2023

  60. [60]

    Zhang, Tianshu and Qian, Kun and Sahai, Siddhartha and Tian, Yuan and Garg, Shaddy and Sun, Huan and Li, Yunyao , title =. Proc. VLDB Endow. , month = jun, pages =. 2025 , issue_date =. doi:10.14778/3748191.3748222 , abstract =

    work page doi:10.14778/3748191.3748222 2025

  61. [61]

    2026 , eprint=

    ALL-FEM: Agentic Large Language models Fine-tuned for Finite Element Methods , author=. 2026 , eprint=

    2026

  62. [62]

    2025 , eprint=

    ALLOY: Generating Reusable Agent Workflows from User Demonstration , author=. 2025 , eprint=

    2025

  63. [63]

    2026 , eprint=

    Attention-Aligned Reasoning for Large Language Models , author=. 2026 , eprint=

    2026

  64. [64]

    2025 , eprint=

    Selective Prompt Anchoring for Code Generation , author=. 2025 , eprint=

    2025

  65. [65]

    2025 , eprint=

    Supporting Construction Worker Well-Being with a Multi-Agent Conversational AI System , author=. 2025 , eprint=

    2025