Narayan Srinivasa

As optimization problems grow in scale and complexity across industries,
conventional computing approaches face fundamental limits in energy
efficiency and scalability. Neuromorphic computing, inspired by the
structure and dynamics of the brain, offers a radically different paradigm—
one where computation emerges from distributed, stochastic, and event-
driven neural processes. This lecture introduces how neuromorphic
systems, such as Intel’s Loihi and Loihi2, can transform optimization by
mapping problems into spiking neural networks that naturally explore
solution spaces via probabilistic dynamics. We demonstrate how this
approach enables efficient solutions to NP-hard problems with significant
gains in energy and time-to-solution. Through examples spanning
combinatorial optimization and real-world scheduling, we highlight both
the promise and the practical challenges of this emerging technology, and
outline a path toward scalable, deployable neuromorphic optimization
systems.

Address:Chief AI Scientist, Arch Systems LLC, United States