TARGA: Targeted Synthetic Data Generation for Practical Reasoning over Structured Data

TL;DR

TARGA is a framework that generates targeted synthetic data for semantic parsing, improving reasoning and generalization in knowledge base question answering without manual annotation.

Contribution

It introduces a dynamic, automated synthetic data generation method that enhances semantic parsing models' performance and generalization without manual labeling.

Findings

Outperforms existing non-fine-tuned methods on KBQA datasets.

Achieves significant F1 score improvements (+7.7, +12.2).

Demonstrates superior sample efficiency and robustness.

Abstract

Semantic parsing, which converts natural language questions into logic forms, plays a crucial role in reasoning within structured environments. However, existing methods encounter two significant challenges: reliance on extensive manually annotated datasets and limited generalization capability to unseen examples. To tackle these issues, we propose Targeted Synthetic Data Generation (TARGA), a practical framework that dynamically generates high-relevance synthetic data without manual annotation. Starting from the pertinent entities and relations of a given question, we probe for the potential relevant queries through layer-wise expansion and cross-layer combination. Then we generate corresponding natural language questions for these constructed queries to jointly serve as the synthetic demonstrations for in-context learning. Experiments on multiple knowledge base question answering (KBQA) datasets demonstrate that TARGA, using only a 7B-parameter model, substantially outperforms existing non-fine-tuned methods that utilize close-sourced model, achieving notable improvements in F1 scores on GrailQA(+7.7) and KBQA-Agent(+12.2). Furthermore, TARGA also exhibits superior sample efficiency, robustness, and generalization capabilities under non-I.I.D. settings.