Chunhua Liu
Postdoc Research Fellow and Associate Lecturer in NLP
Melbourne Connect
I am a Research Fellow in NLP at The University of Melbourne, working with Prof Eduard Hovy on topics under the umbrella of disinformation. I'm interested in building more reliable, inclusive and collaborative AI systems that align with the diverse cultures and values of global human societies.
Previously, I worked with
Dr Patanamon (Pick) Thongtanunam as a postdoc on empirical software engineering research, focusing on enhancing dataset quality, improving the quality of model generated review comments, and investigating LLM hallucinations in automated code review.
I was awarded a PhD in Natural Language Processing under the supervision of Prof Trevor Cohn and Dr Lea Frermann. My research interests center around understanding the structure and relationships between concepts through interdisciplinary perspectives, including NLP and cognitive psychology. With a specific emphasis on understanding the structure and reasons behind human word associations, evaluating the potential of word associations as a source of commonsense knowledge and incorporating commonsense knowledge to improve neural models' reasoning ability on tasks, such as commonsense question answering and natural language inferences.
I'm also exploring how cultural differences shape behaviors and norms, as well as how computational models learn and understand cultural differences. This is also a long-term research agenda that I’m eager to pursue further: understanding how cultural differences shape our daily behaviours and norms, and how these can be learnt and aligned by NLP models such as large-language models.
I am actively seeking academic job opportunities in the domains of NLP, software engineering, and cognitive psychology, including postdoctoral, lecturer, and research scientist positions. If you know of any available opportunities, please feel free to reach out via email (chunhua.liu1{at}unimelb.edu.au) or LinkedIn.
News
- July 2023: Attending ACL/StarSEM to present our paper Seeking Clozure Robust Hypernym Extraction from BERT with Anchored Prompts
Recent Papers
-
Chunhua Liu, Trevor Cohn, Simon Deyne and Lea Frermann
(2022).
WAX: A New Dataset for Word Association eXplanations.
In
Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, AACL/IJCNLP 2022 - Volume 1: Long Papers, Online Only, November 20-23, 2022.
