Ilenna Jones
Kempner Research Fellow
She/Her
Contact Information
Areas I Research:
About
Ilenna Jones is a Research Fellow at the Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University. She received her BA in Neuroscience at Dartmouth College in 2015, with a focus on cellular and molecular neuroscience. She began her Neuroscience PhD at the University of Pennsylvania in 2017, where she pivoted her research focus to computational neuroscience and neuro-AI with her advisor Konrad Kording. There, she was one of 45 students awarded the three-year Howard Hughes Medical Institute Gilliam Fellowship for Advanced Study in 2020. In 2020 and 2021 she was a TA for the Neuromatch Academy Computational Neuroscience and Deep Learning courses, and in 2023 she was a TA for the Imbizo Computational Neuroscience summer school in Cape Town, South Africa. She received her PhD and began her position as a Kempner Research Fellow in 2023. Ilenna often spends her time learning latin dance or playing board games.
Research Focus
Jones’ current research focuses on neural computation, neural biophysics, and optimization. In the brain, is the most fundamental scale for learning and computation at the multi-neuronal level, the single-neuron level, or at the level of a dendritic arbor? Jones investigates how the implementable details of single neurons impact our understanding of what the capabilities of single- and multi-neuronal systems should be. With a background in deep learning, cellular biophysics and dendritic computation, her research vision has three aspects:
(1) How do the biological details of single neurons, such as dendritic branching, ion channel distributions, and synaptic clustering, impact how a neuron performs tasks?
(2) How can we use constraints from multi-neuronal system phenomena to identify normative tasks neurons should perform or objectives neurons should follow?
(3) Which neuronal plasticity rules modify neuronal biological details so that neurons learn to perform normative tasks?
Jones also aims to explore how neuronal details implemented in deep learning architectures impact learning, expressivity, and generalization.