DePasquale Lab

BIO

Kavya is an undergraduate at Boston University studying math and computer science, and a Trustee Scholar and Harold C. Case Scholar. She is a research assistant in the DePasquale Lab and the NOC Lab.

Work in the lab

Kavya joined the lab as an undergraduate research assistant and was awarded a UROP for the fall of 2026.

Her project is Investigating the Flexibility-Noise Tradeoff in Multifunctional Spiking Neural Networks, and she works on it with Ryan and Zach. The brain uses the same circuits for many different computations, yet neural responses vary from one trial to the next, and the origins of that variability are not fully understood. One idea is that some of it is the price of flexibility: the more tasks a circuit learns, the noisier it becomes. She studies this in spiking neural network models trained to perform many tasks. Real circuits seem to keep different computations in nearly non-overlapping subspaces of activity, so she is developing a new learning method, built on the extended Kalman filter, that forces a network to be strictly orthogonal. This lets her test whether doing so reduces noise but also limits how many tasks the network can handle.

In the news