Ari S. Morcos

A Research Scientist at Facebook AI Research, Ari works to understand the mechanisms underlying neural network computation and function and uses these insights to build machine learning systems more intelligently. Ari is also involved in understanding lottery tickets, properties predictive of generalization, methods to compare representations across networks, the role of single units in computation, and strategies that measure and induce abstraction in neural network representations. Ari previously worked at DeepMind in London, and earned a Ph.D. in neurobiology at Harvard University, where he used machine learning to study the cortical dynamics underlying evidence accumulation for decision-making.

Ari's Publications

April 07, 2020

RESEARCH

Insights on Visual Representations for Embodied Navigation Tasks

Recent advances in deep reinforcement learning require a large amount of data and result in representations that are often over specialized to the target task. In this work, we study the underlying potential causes for this specialization by measurin…

April 07, 2020