Robust and Reliable Computation using Heterogeneous Devices and Circuits

For many practical tasks, that involve real-time processing of sensory data and closed-loop interactions with the environment, conventional artificial intelligence neural network accelerators cannot match the performance of biological ones.<br/>Neuromorphic intelligence aims to emulate the principles of computation of animal brains, exploiting the physics of computation of electronic and memristive devices.<br/>Similar to their biological counterparts, neuromorphic devices and circuits are affected by variability and inhomogeneities.<br/>In this talk I will present brain-inspired methods and strategies to achieve reliable and robust computation using an unreliable and inhomogeneous computing substrate, and present examples of neuromorphic systems that use these methods to solve edge-computing and sensory processing applications.