Zoom: https://illinois.zoom.us/j/88360282620?pwd=sViMRaJaBnBZIczc9GLj0g0fgtOn3b.1
Refreshments Provided.
Abstract:
The nervous system transforms sensory inputs into intelligent behavior through the dynamic interactions of neurons. To understand this process requires meeting two outstanding challenges: systematically deciphering high-dimensional sensory representations and inferring mechanistic dynamics models.
In this seminar, I will outline three directions of research in my lab that address these challenges.
Deep neural network (DNN) models of sensory systems achieve high predictive accuracy. Yet, concrete insights into neural encoding have been slow to emerge from these complex models. I will describe an approach, 'artiphysiology', where electrophysiological experiments are reproduced in a DNN. Subsequently, analysis of the DNN provides well-motivated hypotheses.
Recent advances in large-scale neural recording technology have spurred new inquiries into the high-dimensional geometry of the neural code. However, characterizing this geometry from noisy neural responses—particularly in datasets with more neurons than trials—poses major statistical challenges. We address this problem by developing new tools for the accurate estimation of high-dimensional signal geometry.
Electron microscopy has enabled the reconstruction of an entire nervous system down to the nanometer scale. It is unclear how to relate this structural information to the function of the organism. We approach this issue by leveraging natural symmetries in the nervous system to pose inference of structure-function relationships as an unsupervised learning problem.
Bio:
Dean is an Assistant Professor of Computational Neuroscience at the University of Illinois Urbana-Champaign in the Department of Psychology. Dean was a postdoctoral fellow at the Princeton Neuroscience Institute working with Dr. Jonathan Pillow. He received his PhD in Neuroscience at the University of Washington with Dr. Wyeth Bair.
Part of the Siebel School Speakers Series. Faculty Host: Mohammed El-Kebir
Meeting ID: 883 6028 2620
Passcode: csillinois
If accommodation is required, please email <erink@illinois.edu> or <communications@cs.illinois.edu>. Someone from our staff will contact you to discuss your specific needs