Neuromorphic Computing May Be the Next Big Thing

This piece, by Onno Berkan, was published on 11/05/24. The original text, by Schuman et al., was published by Nature Computational Science on 01/31/22.

Neuromorphic computing represents a new frontier in computer technology that aims to mimic how the human brain processes information. Neuromorphic Computers have become increasingly important as traditional computing methods reach their physical limits.

This Oak Ridge National Laboratory study suggests that the field has evolved significantly since its inception in the 1980s when it primarily focused on mixed analog-digital systems. Several major neuromorphic computing systems have been developed, including prominent examples like IBM's TrueNorth and Intel's Loihi.

These systems have shown promise in various applications, such as keyword recognition, medical image analysis, and object detection. However, the field faces several challenges. One significant issue is that current benchmark tests, which use datasets like MNIST and CIFAR-10, don't fully utilize the temporal processing capabilities that make neuromorphic computers unique.

The research community is exploring various approaches to improve these systems, including methods inspired by biological processes in the brain. While neurons and synapses have been the primary focus, other neural components like glial cells might also be valuable for computation.

One of the exciting opportunities in this field is the potential for "co-design," where all aspects of the system– from materials to applications– can influence each other directly. This approach could lead to more efficient and effective systems than the current bottom-up development method.

A significant challenge facing the field is the need for better development tools and simulators. Current simulators are often slow and limited in scale, making it difficult to evaluate and improve new approaches rapidly.

Another significant hurdle is the need for standardized programming methods. Currently, programmers must design very detailed systems, making the process time-consuming and prone to errors.

The researchers suggest that instead of focusing on a single benchmark or challenge problem, the field would benefit from developing a suite of diverse challenges that could better demonstrate the full capabilities of neuromorphic systems. This approach ensures that advances in the field have broader applications rather than being too narrowly focused.

Neuromorphic computing is one of the most interesting and exciting fields out there! If you’re interested, Prof. Joshua Yang at Viterbi works on Neuromorphic computing systems. There’ll be a faculty spotlight on him soon.

Want to submit a piece? Or trying to write a piece and struggling? Check out the guides here!

Thank you for reading. Reminder: Byte Sized is open to everyone! Feel free to submit your piece. Please read the guides first though.

All submissions to berkan@usc.edu with the header “Byte Sized Submission” in Word Doc format please. Thank you!