Dr Biswa Sengupta

Dr Biswa Sengupta

Background - Research - Publications

Background - Research - Publications

Position: Senior Research Associate

Position: Senior Research Associate

Office Location: BN4-86

Office Location: BN4-86

Background

Background

Biswa Sengupta received a B.Eng. (2004) degree in Electronic and Computer Engineering and an MSc degree (2005) in theoretical computer science from the University of York, UK before moving to the Max Planck Institute for Biological Cybernetics (Depts. Schölkopf and Logothetis) for another MSc degree (2007) to work in the intersection of neuroscience and machine learning. He returned to University of Cambridge as a Dorothy Hodgkin fellow and obtained his doctoral degree studying tradeoffs between information encoding and energy consumption in single neurons. He was subsequently awarded a Wellcome Trust independent fellowship to work on cost-benefit problems in theoretical neuroscience. He developed this work at University College London (UCL) via two strands ? (a) understanding non-equilibrium steady states with a goal of establishing a Bayesian statistical theory for the nervous system, and (b) manufacturing computationally efficient inference algorithms for neuroimaging that are inspired by information geometry. With joint appointment at UCL, he is currently a senior research scientist at University of Cambridge where he works on Riemann geometric inversion methods for data emerging from fMRI/MEG/EEG at a systems level, along with understanding reduced models at single neuron level.

Biswa Sengupta received a B.Eng. (2004) degree in Electronic and Computer Engineering and an MSc degree (2005) in theoretical computer science from the University of York, UK before moving to the Max Planck Institute for Biological Cybernetics (Depts. Schölkopf and Logothetis) for another MSc degree (2007) to work in the intersection of neuroscience and machine learning. He returned to University of Cambridge as a Dorothy Hodgkin fellow and obtained his doctoral degree studying tradeoffs between information encoding and energy consumption in single neurons. He was subsequently awarded a Wellcome Trust independent fellowship to work on cost-benefit problems in theoretical neuroscience. He developed this work at University College London (UCL) via two strands ? (a) understanding non-equilibrium steady states with a goal of establishing a Bayesian statistical theory for the nervous system, and (b) manufacturing computationally efficient inference algorithms for neuroimaging that are inspired by information geometry. With joint appointment at UCL, he is currently a senior research scientist at University of Cambridge where he works on Riemann geometric inversion methods for data emerging from fMRI/MEG/EEG at a systems level, along with understanding reduced models at single neuron level.

Research Interests

Research Interests

Applied: Machine learning, Bayesian statistics and hierarchial networks Pure: Riemann geometry and optimal transport

Applied: Machine learning, Bayesian statistics and hierarchial networks

Pure: Riemann geometry and optimal transport

Selected Publications

Selected Publications

#JumpPublications

The list of publications is available at https://scholar.google.co.uk/citations?user=YZHhV9kAAAAJ&hl=en

The list of publications is available at https://scholar.google.co.uk/citations?user=YZHhV9kAAAAJ&hl=en