The Loihi Neuromorphic Research Chip

This chapter focuses on Intel's Loihi neuromorphic chip. Although Loihi is a digital chip, this avenue of research pushes beyond conventional von Neumann architectures. The chapter describes Loihi as an exploratory AI research chip inspired by the principles of neural computation in nature. Loihi explores a very different paradigm compared to the von Neumann and matrix-acceleration architectures in wide use today. The chapter gives an industry perspective on neuromorphic computing, indicating that the research is not limited to universities, but is also being pursued by industrial research organizations. The Loihi architecture integrates self-modifying processes allowing autonomous adaptation and learning in response to discrete events and an evolving statistical environment. The chip supports sparse and irregular communication between its units in the form of packetized event-driven "spikes". Loihi is implemented using a novel asynchronous design methodology that allows fully exploiting activation sparsity.