Ontology learning with large language models (LLMs) for task-oriented dialogue (TOD) systems
This project focuses on ontology learning with large language models (LLMs) for task-oriented dialogue (TOD) systems. Building on the work of Lo et al. (2024), who fine-tuned LLMs for ontology construction using ArXiv and Wikipedia data, we aim to adapt similar methods for TOD ontology learning. Using datasets like MultiWOZ (Budzianowski et al., 2018) and schema-guided dialogue (SGD; Rastogi et al., 2020), we will explore automatic ontology construction and relation extraction. Additionally, ontology evaluation remains an important open research question, and this project will investigate new approaches for assessing the quality of learned ontologies.
References
Pawel Budzianowski, Tsung-Hsien Wen, Bo-Hsiang Tseng, I˜nigo Casanueva, Stefan Ultes, Osman Ramadan, and Milica Gašić. 2018. MultiWOZ - A Large-Scale Multi-Domain Wizard-of-Oz Dataset for Task-Oriented Dialogue Modelling. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 5016–5026, Brussels, Belgium. Association for Computational Linguistics.
Andy Lo, Albert Q Jiang, Wenda Li, and Mateja Jamnik. 2024. End-to-End Ontology Learning with Large Language Models. In Advances in Neural Information Processing Systems, volume 38. Curran Associates, Inc.
Abhinav Rastogi, Xiaoxue Zang, Srinivas Sunkara, Raghav Gupta, and Pranav Khaitan. 2020. Towards Scalable Multi-domain Conversational Agents: The Schema-Guided Dialogue Dataset. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34.05, pages 8689–8696.