Towards Harnessing Large Language Models for Comprehension of Conversational Grounding
This work addresses the problem of enhancing dialogue systems for better handling of conversational grounding, but it is incremental as it focuses on analyzing existing challenges and proposing future improvements.
The study investigated large language models' ability to classify dialogue turns and predict grounded knowledge in information-seeking conversations, revealing challenges in these tasks and discussing ongoing efforts to improve comprehension through pipeline architectures and knowledge bases.
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.