My research at the University of Toronto involves the study of flexibility in reinforcement learning. The goal is to build a system that is trained in a particular task to suddenly be good at another task, given a small set of data in that new task. My research will be directly applied towards the detection of cognitive diseases involving speech. I am also exploring one-shot learning techniques to detect these diseases. All my work will involve sparse and multi-modal data and I am building machine learning models that are apt for this data with a direct impact on the possible prevention of cognitive diseases.

Currently, I am involved in the following projects: 1) Improving/Adaptation of GANs for text 2) Augmenting LDA with word2vec for detection of Alzheimer's disease 3) Ethics of AI in healthcare & Explainable ML 4) Dialogue generation & HCI adaptations of smart assistants for younger population 5) Detection of ICD-10 disease codes from doctor's notes in French, Hungarian and Italian.

I am also interested in using deep learning for standard NLP tasks like sentiment analysis, document classification and similarity. Stock Analysis based on historic data using LSTMs is another research project that I am pursing in my free time.

---