My field of research is in Natural Language Processing (NLP) — a sub-field of Artificial Intelligence — and the goal of NLP is to develop computational models to understand human languages. There are a host of NLP applications that are commonly used in our everyday lives: machine translation (e.g. Google Translate) and chatbot (e.g. Siri) are two examples. My research focuses on building computational models in an unsupervised or semi-supervised setting, i.e. a learning scenario where the supervision signal for model training is not available or scarce, and is characterised by a diverse flavour of applications, e.g. topic models, lexical semantics, text generation and misinformation detection. Some of my research has a broader community interest beyond academia — e.g. my work in text generation and influence operations has been covered by science magazines (New Scientist) and mainstream news media (Guardian and BBC).
Prior to joining the University of Melbourne, I spent over 3.5 years as an industry scientist at IBM Research, developing solutions for clients in application domains from education to government.
Affiliations: The University of Melbourne’s NLP Group and Information and Influence Hub.
Prospective Students
- If your query is related PhD supervision, it would help to: (1) have a look over my publications to identify areas of overlapping interests; and (2) send a thoughtful research proposal to suggest potential projects. As a student I don’t expect that you have a perfect understanding of the gaps in the field and can identify research questions that fit the scope of a PhD, but the proposal would serve to kickstart a discussion about potential project ideas.
- If your query is about internship, please note that I do not take interns (and so any queries of this nature would not receive a response).
News
Visited Mentone Girls’ Grammar for a gifted student event to talk about AI, ethics and creative generation.
Interviewed by SBS Mandarin News about Generative AI and workforce.
Our Linkage partner, Justice Connect, has won the Not-For-Profit Technology Innovator of the Year Award at Infoxchange’s Australian Technology Awards 2024 for our AI project.
Our paper on understanding the relation between empathy and toxicity won the Best Paper Award at ALTA 2023.
Our parallel sentiment dataset for Indonesian languages received the Outstanding Paper Award at EACL 2023.
Our art+science chatbot project is now being exhibited at NGV Australia (Fed Square)! More information in this article (first project).
Our paper on a new natural language inference dataset was a runner up for Best Paper Award at AACL 2022.
Our paper on understanding commonsense stories in Indonesian won the Best Paper Award at CSRR 2022.
Together with Georgia Banks we will be presenting an art+science collaboration project for ANAT Spectra 2022, where we build a chatbot with Georgia’s personality to explore ego/self.
3 papers accepted for ACL2022! One Country, 700+ languages, The Patient Is More Dead Than Alive, and An Interpretable Neuro-Symbolic Reasoning Framework for Task-Oriented Dialogue Generation.
Our work on understanding mass influence activities online is covered by Guardian.
Our paper on conversation disentanglement won the Best Paper Award at ALTA 2021!
- Our paper on learning cross-lingual embeddings for low resource languages won the Best Paper Award at MRL 2021!
- Talked about AI and creativity on the radio show PassW0rd (starts at around the 43 minute mark).
- Our IEEE Spectrum article “The AI Poet” is up.
- I’m presenting a talk on “Creativity, Machine and Poetry” at a public forum on language [video].
- Our Shakespearean sonnet generator was covered by New Scientist, Times UK, Daily Mail, and others. More information can be found here.