I work in the applied research team at Mila, where we solve real-world problems with deep learning. As an example, I worked with Dialogue to develop automatic symptom detection systems, publishing our work at NeurIPS 2022, in the main track and the dataset track.
These days I am most interested in certain topics in NLP, including decoding algorithms, controllable generation, evaluation, retrieval, and interpretability. I also keep an eye on healthcare/bio applications, such as protein design.
I earned my master degree at McGill University, and I did my undergrad at Wuhan University, China. My research focused on machine learning and NLP for healthcare. For instance, I worked on COVID-19 media news surveillance for public health measures, constructing a large medical NLP pre-training dataset, among others.
I listen to jazz in the morning with coffee, post-rock or Pink Floyd when I use my brain (e.g. reading papers or coding), and shoegazing whenever I feel the need. In my free time I (used to) play soccer and (more recently) learn guitar.
- NeurIPSTowards Trustworthy Automatic Diagnosis Systems by Emulating Doctors’ Reasoning with Deep Reinforcement LearningThirty-sixth Conference on Neural Information Processing Systems, 2022
- PatternsInferring global-scale temporal latent topics from news reports to predict public health interventions for COVID-19Patterns, 2022
- EMNLP WorkshopMeDAL: Medical Abbreviation Disambiguation Dataset for Natural Language Understanding PretrainingIn Proceedings of the 3rd Clinical Natural Language Processing Workshop, 2020
- Nat. Commun.Inferring multimodal latent topics from electronic health recordsNature communications, 2020