Introduces the ODUTQA-MDC task with a 25k-pair benchmark and MAIC-TQA multi-agent framework for detecting and clarifying underspecified open-domain tabular questions via dialogue.
hub Mixed citations
Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning
Mixed citation behavior. Most common role is background (62%).
42
Pith papers citing it
Background
62% of classified citations
abstract
A significant amount of the world's knowledge is stored in relational databases. However, the ability for users to retrieve facts from a database is limited due to a lack of understanding of query languages such as SQL. We propose Seq2SQL, a deep neural network for translating natural language questions to corresponding SQL queries. Our model leverages the structure of SQL queries to significantly reduce the output space of generated queries. Moreover, we use rewards from in-the-loop query execution over the database to learn a policy to generate unordered parts of the query, which we show are less suitable for optimization via cross entropy loss. In addition, we will publish WikiSQL, a dataset of 80654 hand-annotated examples of questions and SQL queries distributed across 24241 tables from Wikipedia. This dataset is required to train our model and is an order of magnitude larger than comparable datasets. By applying policy-based reinforcement learning with a query execution environment to WikiSQL, our model Seq2SQL outperforms attentional sequence to sequence models, improving execution accuracy from 35.9% to 59.4% and logical form accuracy from 23.4% to 48.3%.
hub tools
citation-role summary
background 5
dataset 2
method 1
citation-polarity summary
representative citing papers
Tiny language models under 10M parameters trained on a synthetic children's story dataset generate fluent, consistent, multi-paragraph English text with near-perfect grammar and reasoning.
Randomly replacing labels in in-context demonstrations barely hurts performance, showing that label space, input distribution, and sequence format drive in-context learning more than ground-truth labels.
GS-QA is a new benchmark of 2,800 QA pairs on 28 templates using OSM and Wikipedia data to evaluate LLMs on spatial predicates, multi-source reasoning, and diverse answer types including distances and counts.
LEAF-SQL uses level-wise exploration with adaptive fine-graining and dual agents to generate diverse SQL skeletons, reaching 71.6% execution accuracy on the BIRD benchmark and outperforming prior search- and skeleton-based methods.
RSAT uses SFT on verified traces followed by GRPO with NLI faithfulness rewards to make 1-8B models produce verifiable table reasoning with cell citations, raising faithfulness 3.7x to 0.826.
NL2SQLBench is a new modular benchmarking framework that evaluates LLM NL2SQL methods across three core modules on existing datasets, exposing large accuracy gaps and computational inefficiency.
The authors define a taxonomy for LLM-enhanced relational operators categorized into Select, Match, Impute, Cluster and Order, and release LROBench to evaluate single and multi-operator queries on semantic database processing.
Visual-TableQA is a new open-domain benchmark of rendered table images and complex QA pairs created via multi-LLM collaborative generation, with fine-tuned models showing robust generalization to external tests.
FLARE is a vision-language model family using text-guided vision encoding, context-aware alignment decoding, dual-semantic mapping loss, and text-driven VQA synthesis to achieve deep cross-modal integration, outperforming larger models with only 630 vision tokens at 3B scale.
Adapting large language models by training only a low-rank decomposition BA added to frozen weight matrices matches full fine-tuning while cutting trainable parameters by orders of magnitude and adding no inference latency.
MetaEvaluator applies meta-learning over reference models to deliver label-free performance estimates for unseen models across architectures and modalities on unlabeled datasets.
ξ-DPO rewrites the preference objective as minimizing distance to optimal margins and defines reward as a chosen-to-rejected ratio, yielding a bounded, interpretable margin ξ set directly from the initial reward-gap distribution.
FineStep adds step-level process rewards and credit assignment to tool-augmented Text-to-SQL, achieving 3.25% higher execution accuracy than GRPO on BIRD while cutting redundant tool calls.
FINER-SQL boosts 3B-parameter small language models to 67.73% and 85% execution accuracy on BIRD and Spider benchmarks via dense memory and atomic rewards in group relative policy optimization, matching larger LLMs at lower latency.
EGRefine optimizes column renamings via execution-grounded verification and view materialization to recover Text-to-SQL accuracy lost to schema naming issues while guaranteeing query equivalence.
Modular curriculum learning with tier-specific adapters outperforms standard fine-tuning on complex Text-to-SQL queries in Spider and BIRD benchmarks by avoiding catastrophic forgetting.
ReCoQA is a new large-scale benchmark for multi-step tool-augmented reasoning in real estate QA, accompanied by the HIRE-Agent hierarchical understand-plan-execute baseline.
A self-healing LLM pipeline for natural language to PostgreSQL translation achieves up to 9.3 percentage point accuracy gains on benchmarks through error diagnosis and anti-regression mechanisms.
AV-SQL uses a pipeline of LLM agents to generate intermediate CTE views that decompose complex Text-to-SQL queries, reaching 70.38% execution accuracy on Spider 2.0.
OmniTQA integrates LLM semantic reasoning as a first-class query operator with classical relational operators in a cost-aware planner for hybrid structured and semi-structured data.
A literature survey that taxonomizes methods, datasets, and evaluation practices for natural language interfaces to geospatial and temporal databases while identifying recurring trends and future directions.
TabXEval is a rubric-based two-phase framework using structural alignment (TabAlign) and semantic-syntactic comparison (TabCompare) to evaluate tables more precisely than standard metrics.
VPiT enables pretrained LLMs to perform both visual understanding and generation by predicting discrete text tokens and continuous visual tokens, with understanding data proving more effective than generation-specific data.
citing papers explorer
-
ODUTQA-MDC: A Task for Open-Domain Underspecified Tabular QA with Multi-turn Dialogue-based Clarification
Introduces the ODUTQA-MDC task with a 25k-pair benchmark and MAIC-TQA multi-agent framework for detecting and clarifying underspecified open-domain tabular questions via dialogue.
-
TinyStories: How Small Can Language Models Be and Still Speak Coherent English?
Tiny language models under 10M parameters trained on a synthetic children's story dataset generate fluent, consistent, multi-paragraph English text with near-perfect grammar and reasoning.
-
Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?
Randomly replacing labels in in-context demonstrations barely hurts performance, showing that label space, input distribution, and sequence format drive in-context learning more than ground-truth labels.
-
GS-QA: A Benchmark for Geospatial Question Answering
GS-QA is a new benchmark of 2,800 QA pairs on 28 templates using OSM and Wikipedia data to evaluate LLMs on spatial predicates, multi-source reasoning, and diverse answer types including distances and counts.
-
LEAF-SQL: Level-wise Exploration with Adaptive Fine-graining for Text-to-SQL Skeleton Prediction
LEAF-SQL uses level-wise exploration with adaptive fine-graining and dual agents to generate diverse SQL skeletons, reaching 71.6% execution accuracy on the BIRD benchmark and outperforming prior search- and skeleton-based methods.
-
RSAT: Structured Attribution Makes Small Language Models Faithful Table Reasoners
RSAT uses SFT on verified traces followed by GRPO with NLI faithfulness rewards to make 1-8B models produce verifiable table reasoning with cell citations, raising faithfulness 3.7x to 0.826.
-
NL2SQLBench: A Modular Benchmarking Framework for LLM-Enabled NL2SQL Solutions
NL2SQLBench is a new modular benchmarking framework that evaluates LLM NL2SQL methods across three core modules on existing datasets, exposing large accuracy gaps and computational inefficiency.
-
Large Language Model-Enhanced Relational Operators: Taxonomy, Benchmark, and Analysis
The authors define a taxonomy for LLM-enhanced relational operators categorized into Select, Match, Impute, Cluster and Order, and release LROBench to evaluate single and multi-operator queries on semantic database processing.
-
Visual-TableQA: Open-Domain Benchmark for Reasoning over Table Images
Visual-TableQA is a new open-domain benchmark of rendered table images and complex QA pairs created via multi-LLM collaborative generation, with fine-tuned models showing robust generalization to external tests.
-
FLARE: Fully Integration of Vision-Language Representations for Deep Cross-Modal Understanding
FLARE is a vision-language model family using text-guided vision encoding, context-aware alignment decoding, dual-semantic mapping loss, and text-driven VQA synthesis to achieve deep cross-modal integration, outperforming larger models with only 630 vision tokens at 3B scale.
-
LoRA: Low-Rank Adaptation of Large Language Models
Adapting large language models by training only a low-rank decomposition BA added to frozen weight matrices matches full fine-tuning while cutting trainable parameters by orders of magnitude and adding no inference latency.
-
Cost-Effective Model Evaluation with Meta-Learning
MetaEvaluator applies meta-learning over reference models to deliver label-free performance estimates for unseen models across architectures and modalities on unlabeled datasets.
-
$\xi$-DPO: Direct Preference Optimization via Ratio Reward Margin
ξ-DPO rewrites the preference objective as minimizing distance to optimal margins and defines reward as a chosen-to-rejected ratio, yielding a bounded, interpretable margin ξ set directly from the initial reward-gap distribution.
-
Every Step Counts: Step-Level Credit Assignment for Tool-Integrated Text-to-SQL
FineStep adds step-level process rewards and credit assignment to tool-augmented Text-to-SQL, achieving 3.25% higher execution accuracy than GRPO on BIRD while cutting redundant tool calls.
-
FINER-SQL: Boosting Small Language Models for Text-to-SQL
FINER-SQL boosts 3B-parameter small language models to 67.73% and 85% execution accuracy on BIRD and Spider benchmarks via dense memory and atomic rewards in group relative policy optimization, matching larger LLMs at lower latency.
-
EGREFINE: An Execution-Grounded Optimization Framework for Text-to-SQL Schema Refinement
EGRefine optimizes column renamings via execution-grounded verification and view materialization to recover Text-to-SQL accuracy lost to schema naming issues while guaranteeing query equivalence.
-
LeGo-Code: Can Modular Curriculum Learning Advance Complex Code Generation? Insights from Text-to-SQL
Modular curriculum learning with tier-specific adapters outperforms standard fine-tuning on complex Text-to-SQL queries in Spider and BIRD benchmarks by avoiding catastrophic forgetting.
-
ReCoQA: A Benchmark for Tool-Augmented and Multi-Step Reasoning in Real Estate Question and Answering
ReCoQA is a new large-scale benchmark for multi-step tool-augmented reasoning in real estate QA, accompanied by the HIRE-Agent hierarchical understand-plan-execute baseline.
-
SQL Query Engine: A Self-Healing LLM Pipeline for Natural Language to PostgreSQL Translation
A self-healing LLM pipeline for natural language to PostgreSQL translation achieves up to 9.3 percentage point accuracy gains on benchmarks through error diagnosis and anti-regression mechanisms.
-
AV-SQL: Decomposing Complex Text-to-SQL Queries with Agentic Views
AV-SQL uses a pipeline of LLM agents to generate intermediate CTE views that decompose complex Text-to-SQL queries, reaching 70.38% execution accuracy on Spider 2.0.
-
OmniTQA: A Cost-Aware System for Hybrid Query Processing over Semi-Structured Data
OmniTQA integrates LLM semantic reasoning as a first-class query operator with classical relational operators in a cost-aware planner for hybrid structured and semi-structured data.
-
Natural Language Interfaces for Spatial and Temporal Databases: A Comprehensive Overview of Methods, Taxonomy, and Future Directions
A literature survey that taxonomizes methods, datasets, and evaluation practices for natural language interfaces to geospatial and temporal databases while identifying recurring trends and future directions.
-
TabXEval: Why this is a Bad Table? An eXhaustive Rubric for Table Evaluation
TabXEval is a rubric-based two-phase framework using structural alignment (TabAlign) and semantic-syntactic comparison (TabCompare) to evaluate tables more precisely than standard metrics.
-
MetaMorph: Multimodal Understanding and Generation via Instruction Tuning
VPiT enables pretrained LLMs to perform both visual understanding and generation by predicting discrete text tokens and continuous visual tokens, with understanding data proving more effective than generation-specific data.
-
Encoding Database Schemas with Relation-Aware Self-Attention for Text-to-SQL Parsers
Relation-aware self-attention encodes schema structure for text-to-SQL, raising exact-match accuracy on Spider from 18.96% to 42.94%.
-
SecureMCP: A Policy-Enforced LLM Data Access Framework for AIoT Systems via Model Context Protocol
SecureMCP integrates RBAC with five sequential defense modules in an MCP server to achieve 82.3% policy compliance against adversarial LLM SQL queries in AIoT while preserving execution accuracy.
-
SCOPE:Planning for Hybrid Querying over Clinical Trial Data
SCOPE uses explicit multi-LLM planning to improve accuracy on 1,500 hybrid reasoning questions over clinical trial tables compared to zero-shot, few-shot, CoT, and agent baselines.
-
A Demonstration of SQLyzr: A Platform for Fine-Grained Text-to-SQL Evaluation and Analysis
SQLyzr is a new evaluation platform that adds diverse metrics, realistic settings, query classification, and analysis features to overcome the single-score limitations of existing text-to-SQL benchmarks.
-
FD-NL2SQL: Feedback-Driven Clinical NL2SQL that Improves with Use
FD-NL2SQL is a feedback-driven clinical NL2SQL system that decomposes questions, retrieves exemplars via embeddings, synthesizes SQL, and expands its example bank from user edits plus logic-based mutations to improve without new annotations.
-
Adapt to Thrive! Adaptive Power-Mean Policy Optimization for Improved LLM Reasoning
APMPO boosts average Pass@1 scores on math reasoning benchmarks by 3 points over GRPO by using an adaptive power-mean policy objective and feedback-driven clipping bounds in RLVR training.
-
Free Energy-Driven Reinforcement Learning with Adaptive Advantage Shaping for Unsupervised Reasoning in LLMs
FREIA applies free energy principles and adaptive advantage shaping to unsupervised RL, outperforming baselines by 0.5-3.5 Pass@1 points on math reasoning with a 1.5B model.
-
HFX: Joint Design of Algorithms and Systems for Multi-SLO Serving and Fast Scaling
HFX jointly designs scheduling and scaling for multi-SLO LLM serving, achieving up to 4.44x higher SLO attainment, 65.82% lower latency, and 49.81% lower cost than prior systems on multi-task workloads.
-
StarCoder: may the source be with you!
StarCoderBase matches or beats OpenAI's code-cushman-001 on multi-language code benchmarks; the Python-fine-tuned StarCoder reaches 40% pass@1 on HumanEval while retaining other-language performance.
-
CraftAssist: A Framework for Dialogue-enabled Interactive Agents
CraftAssist supplies a Minecraft bot, dialogue interface, and data-recording platform intended to support research on agents that execute tasks specified through conversation.
-
Knowledge Distillation for Low-Resource Open-source Text-to-SQL Model
A knowledge-aware Text-to-SQL framework constructs domain knowledge bases to generate synthetic data and enhance inference, claiming substantial gains on seven benchmarks especially in low-resource settings.
-
Retrieve Only Relevant Tables Whether Few or Many: Adaptive Table Retrieval Method
An adaptive thresholding mechanism combined with sliding-window reranking retrieves a query-dependent number of tables from large corpora, improving retrieval and downstream text-to-SQL performance on Spider, BIRD, and Spider 2.0.
-
M3: Conversational LLMs Simplify Secure Clinical Data Access, Understanding, and Analysis
M3 uses LLMs to translate natural language into SQL for the MIMIC-IV database, achieving 93-94% accuracy on benchmark questions with support for local privacy-preserving deployment.
-
Why Build an Assistant in Minecraft?
A rationale is presented for developing an assistant in Minecraft to advance natural language understanding and dialogue learning.
-
MLFriend: Interactive Prediction Task Recommendation for Event-Driven Time-Series Data
MLFriend enumerates prediction tasks for event-driven time-series data and interactively recommends useful ones, with evaluation on three datasets yielding 2885 tasks of which 722 were deemed useful by experts.
-
Table Question Answering in the Era of Large Language Models: A Comprehensive Survey of Tasks, Methods, and Evaluation
A survey that categorizes TQA benchmarks and LLM modeling strategies by challenges while identifying underexplored areas such as reinforcement learning.
-
Large Language Models: A Survey
The paper surveys key large language models, their training methods, datasets, evaluation benchmarks, and future research directions in the field.
-
Neural Machine Translating from Natural Language to SPARQL
Eight NMT models are evaluated for natural language to SPARQL translation, with CNN-based models reaching BLEU up to 98 and accuracy up to 94% on high-quality datasets.