SUQL: Conversational Search over Structured and Unstructured Data with Large Language Models

TL;DR

This paper introduces SUQL, a novel language for conversational agents that enables seamless querying of hybrid structured and unstructured data, supported by a semantic parser based on large language models, demonstrated on real-world datasets.

Contribution

The paper presents SUQL, the first formal language for hybrid data access in conversational agents, and a semantic parser that handles large, real-world databases with mixed data types.

Findings

Achieves near state-of-the-art accuracy on HybridQA dataset.

Effective in real-world restaurant knowledge base with high user satisfaction.

Applicable to large databases and free-text corpora.

Abstract

While most conversational agents are grounded on either free-text or structured knowledge, many knowledge corpora consist of hybrid sources. This paper presents the first conversational agent that supports the full generality of hybrid data access for large knowledge corpora, through a language we developed called SUQL (Structured and Unstructured Query Language). Specifically, SUQL extends SQL with free-text primitives (summary and answer), so information retrieval can be composed with structured data accesses arbitrarily in a formal, succinct, precise, and interpretable notation. With SUQL, we propose the first semantic parser, an LLM with in-context learning, that can handle hybrid data sources. Our in-context learning-based approach, when applied to the HybridQA dataset, comes within 8.9% exact match and 7.1% F1 of the SOTA, which was trained on 62K data samples. More significantly, unlike previous approaches, our technique is applicable to large databases and free-text corpora. We introduce a dataset consisting of crowdsourced questions and conversations on Yelp, a large, real restaurant knowledge base with structured and unstructured data. We show that our few-shot conversational agent based on SUQL finds an entity satisfying all user requirements 90.3% of the time, compared to 63.4% for a baseline based on linearization.