Barak Pearlmutter is a professor in the Department of Computer Science at Maynooth University (NUI Maynooth) in Ireland, where he leads the Brain and Computation Lab. His research sits at the intersection of machine learning, neuroscience, and programming language design, driven by a long-standing interest in adaptive systems and how computation can help us understand the brain.
Pearlmutter received his PhD in Computer Science from Carnegie Mellon University, writing his thesis on the gradient descent process in neural networks. He subsequently held a faculty position at the University of New Mexico before moving to Ireland.
He is widely known for his contributions to automatic differentiation, a mathematical technique that plays a central role in training modern AI systems. His research has also addressed how the brain processes and separates sounds, how neural signals can be used to control computers directly, and the theoretical foundations of machine learning. He holds patents in automatic differentiation and acoustic source separation, and his early software contributions include Oaklisp, an object-oriented programming language, and a garbage collector incorporated into the Linux kernel.
