Past Event! Note: this event has already taken place.
Generative Machine Learning Models Enabled Intelligent Data Analytics
January 21, 2019 at 1:30 PM to 3:00 PM
| Location: | 5345 Herzberg Laboratories |
| Cost: | Free |
Abstract: Modern data analytics faces three grand challenges among others: (1) how can we characterise the behaviours of all input features in a massive high-dimensional dataset of many classes? (2) how are multi-modal data (data captured from different perspectives) integrated for better decision making or knowledge discovery? and (3) how can the representation of data in hierarchical hidden space be interpretable in the use of deep learning? Through modelling joint distributions, generative models have versatile functionality that discriminative models are unable to achieve. In this talk, the speaker will discuss his successful design of generative models to wisely address these three challenges. First, Bayesian matrix factorisation methods are proposed to comprehensively discover ubiquitous and specific feature patterns from large-scale multi-class data through enforcing structured sparsity. Second, exponential family generative models and derived deep generative models are presented to integrate multi-modal data. Third, novel capsule (shallow and deep) generative models are investigated for the understanding and visualisation of data representation in the hidden space.
Bio: Yifeng Li is a Research Officer within the Scientific Data Mining Team of Digital Technologies Research Centre, National Research Council of Canada (NRC). He is the co-founder of the Ottawa-AI Alliance and co-organiser of the Ottawa-AI Workshop. He received the NRC’s Rising Star Award from the NRC President in 2018 in recognition of his outstanding research accomplishments and leadership. Prior to his joining to NRC, Dr. Li was a post-doctorate supported by the NSERC Postdoctoral Fellowships Program at the Wasserman Laboratory of the Centre for Molecular Medicine and Therapeutics, University of British Columbia. He obtained his PhD from the School of Computer Science, University of Windsor, in 2013. His doctoral dissertation was recognised by the Governor General’s Gold Medal. His research interests include theory & applications of machine learning & deep learning, data science & data analytics, computational biology & bioinformatics, cancer research, gene regulation, large-scale optimisation, computational intelligence, etc. He has published 18 journal articles and 26 conference papers of high quality in these areas.