The human brain is the most powerful and efficient computer on the planet. Even after decades of technological advancements, no computer has ever been able to outperform the brain in terms of efficiency, power consumption, and a variety of other factors.
Will neuromorphic computers be capable of accomplishing this?
The exact sequence of events that occur when we perform a specific task on our computer, or on any other device, is entirely dependent on its architecture. It depends on the solid state structure of the computer’s various components, such as the processor and memory.
The Von Neumann architecture, which was first introduced in the late 1940s, is used in almost all modern computers.
The processor is in charge of executing instructions and programmes, and the memory is in charge of storing those instructions and programmes. When you consider your body as an embedded device, your brain serves as both a processor and a memory. There is no distinction between the two because of the architecture of our brain.
Doesn’t it make sense to modify computer architecture so that it functions more like our brain, given that we know the human brain is superior to every computer that exists? Many scientists, beginning with Carver Mead, an American scientist and engineer, realised this in the 1980s.
Now fast forward to the present
Almost every company has a dedicated team to neuromorphic computing. Multiple research organizations and universities are conducting groundbreaking research. It’s safe to say that neuromorphic computing is gaining traction and will continue to do so as new technologies emerge.
It’s worth noting that, despite the fact that this is a specialised field with prerequisites in solid-state physics, VLSI, neural networks, and computational neurobiology, undergraduate engineering students are very interested in it.
Last year, Dr Shubham Sahay, Assistant Professor of Electrical Engineering at IIT Kanpur, introduced a course on neuromorphic computing. Despite the fact that it was a post-graduate course, he saw a lot of interest from undergraduates. “They were very interactive throughout the course.” The large number of B.Tech students in my class attests to the fact that undergrads are passionate about this subject. “I believe that in the future, this (neuromorphic computing) could be included as a core course in the undergraduate curriculum,” he says.
