Neuromorphic NeuRRAM Chip for AI Developed, Performs Computations in Memory Without Network Connectivity: Details

Photo Credit: David Baillot/UC San Diego Jacobs School of Engineering

Scientists have developed a neuromorphic chip to run AI applications that can perform computations directly in memory without the need for a network connection to the cloud. In addition, the chip consumes a small amount of energy which makes it more efficient than other chips. The discovery is expected to enable AI to be used in a variety of devices where it can perform many sophisticated tasks without relying on a central server.

The NeuRRAM chip has proven to be more efficient than compute-in-memory chips and can provide results as accurate as conventional chips. Additionally, the chip may have applications in tasks such as image recognition and reconstruction and voice recognition.

AI computing requires both power and computational capacity. Most AI applications on edge devices require data to be moved from the device to the cloud, where it is processed. The data is then transmitted back to the device. This is because most edge devices are battery powered and have limited power that they use for computing.

Developed by engineers at the University of California, the NeuRRAM chip reduces this power consumption, making edge devices smarter, more robust and accessible. Moreover, it also enhances data security as transferring data from the device to the cloud involves some data privacy risks.

The process of transferring data is considered a cumbersome task. "It's like commuting eight hours for a two-hour work day," explained Weir Vaughn, a Stanford University PhD graduate who worked on the chip at UC San Diego. He is also a co-author of a study published in Nature.

The team used a type of non-volatile memory called resistive random access memory that enables in-memory calculations without the need for a separate computer unit. While compute-in-memory is not a new approach, the new RRAM chip is different because it offers optimal performance and flexibility for diverse AI applications while maintaining uniform accuracy.

The researchers demonstrated the chip's capabilities by running various tasks on it and observed impressive results that were comparable to existing digital chips.