Stephen Keeley

Bio:
Dr. Stephen Keeley studies machine learning in the contexts of problems in neuroscience using primarily Bayesian statistical methods. However, some of the work in understanding brain activity can be used to understand high dimensional data in other settings, such as deep neural networks. As such, Dr. Keeley has a project that learns how to characterize information as it is transmitted across a deep neural network.

Abstract:
We introduce a nonlinear variational framework with an alignment objective inspired by stochastic neighbor embedding. Our method enables explicit control over the degree of alignment between latent spaces of high dimensional activity via a tunable hyperparameter, allowing representations to remain independent or become aligned to facilitate interpretability. We demonstrate the approach on both synthetic data from a four-layer deep neural network and multi-region neural recordings from mouse visual cortex. Across both settings, the method successfully aligns latent spaces and reveals how manifolds transform across layers or brain regions, providing a flexible tool for probing neural information transformations