Towards Harnessing Large Language Models for Comprehension of Conversational Grounding

TL;DR

This paper investigates the ability of large language models to understand conversational grounding by classifying dialogue turns and predicting knowledge elements, highlighting current challenges and potential improvements.

Contribution

It introduces an experimental analysis of large language models' capabilities in conversational grounding tasks and discusses strategies to enhance their understanding through pipeline architectures and knowledge bases.

Findings

Large language models face challenges in classifying grounding-related dialogue turns.

Predicting grounded knowledge elements remains difficult for current models.

Ongoing research aims to improve models' comprehension using structured architectures and knowledge bases.

Abstract

Conversational grounding is a collaborative mechanism for establishing mutual knowledge among participants engaged in a dialogue. This experimental study analyzes information-seeking conversations to investigate the capabilities of large language models in classifying dialogue turns related to explicit or implicit grounding and predicting grounded knowledge elements. Our experimental results reveal challenges encountered by large language models in the two tasks and discuss ongoing research efforts to enhance large language model-based conversational grounding comprehension through pipeline architectures and knowledge bases. These initiatives aim to develop more effective dialogue systems that are better equipped to handle the intricacies of grounded knowledge in conversations.