Shakir Mohamed
Program
Junior Fellow Academy
Appointment
Junior Fellow, Neural Computation and Adaptive Perception
Institution
University of British Columbia
Country
Canada 
Shakir Mohamed is a CIFAR Junior Fellow working under the supervision of Neural Computation and Adaptive Perception program Fellow Nando de Freitas and Associate Kevin Murphy in the Department of Computer Science at the University of British Columbia. Shakir completed his Ph.D. in 2010 in Statistical Machine Learning at St. John’s College, University of Cambridge, under the supervision of Dr. Zoubin Ghahramani. He holds an M.Sc. in Engineering with distinction and a B.Sc. in Electrical/Information Engineering with distinction from the University of the Witwatersrand, Johannesburg, South Africa. In 2009/2010, Shakir served as President of the Samuel Butler Room Society, the graduate society of St. John’s College, Cambridge. In 2006, he gave of his time as a careers role model for the Sci-Bono Science Centre in Johannesburg.
Shakir’s research focuses on statistical approaches to discovering structure in complex data. In his Ph.D., Shakir developed new models for factor analysis, a class of models that search for underlying factors that explain observed data. These underlying factors can correspond to preferences for film rental, factors explaining the expression of genes in genomic applications, or the low level features used by the visual system in interpreting a visual scene. Shakir’s research emphasizes Bayesian statistical approaches that allow for the uncertainty in our understanding of the world to be accounted for, and provides a principled mechanism for updating our beliefs of the world’s state as new evidence is accumulated. As a Junior Fellow, Shakir will develop Bayesian methods that will continue to help us understand complex data, whether this will be in understanding large social networks, in complex robotic systems or in the input to the visual system. Shakir will have a focus on analysis of visual systems involving research in generative modelling, sparse Bayesian learning, non-parametric Bayesian statistics, matrix factorization and latent variable modelling.
Shakir’s research focuses on statistical approaches to discovering structure in complex data. In his Ph.D., Shakir developed new models for factor analysis, a class of models that search for underlying factors that explain observed data. These underlying factors can correspond to preferences for film rental, factors explaining the expression of genes in genomic applications, or the low level features used by the visual system in interpreting a visual scene. Shakir’s research emphasizes Bayesian statistical approaches that allow for the uncertainty in our understanding of the world to be accounted for, and provides a principled mechanism for updating our beliefs of the world’s state as new evidence is accumulated. As a Junior Fellow, Shakir will develop Bayesian methods that will continue to help us understand complex data, whether this will be in understanding large social networks, in complex robotic systems or in the input to the visual system. Shakir will have a focus on analysis of visual systems involving research in generative modelling, sparse Bayesian learning, non-parametric Bayesian statistics, matrix factorization and latent variable modelling.