Skip to main content

Stern is a theoretical physicist studying biological networks, with a focus on neural networks. She shows how modifications in a network’s local connectivity impact their global dynamics and, as a result, their task-performing abilities.

Stern employs modeling frameworks adapted from statistical mechanics to investigate network dynamics and their task performances. Previous attempts to explain tasks performed by networks were made by repeatedly training the networks on the same task. While this method resulted in rising performance successes, the networks grew so complex that their mechanisms were not well understood. To remedy this, Stern employs analytical tools to study biological networks, including nonlinear dynamics methods, random matrices, and mean-field theory. Alongside the analytical tools, she uses large-scale simulations and data analysis to utilize, test, and inform analytical findings. Through collaborations with other researchers at Rockefeller, Stern compares her modeling results to findings from experimental labs, ensuring the plausibility and relevance of her results.