Quantum Materials Unlock the Potential for Energy Efficient Brain-Like Computers

Creating Brain-Like Computers with Minimal Energy Requirements

In the world of computing, we often rely on computers to complete complex tasks quickly and accurately. However, when it comes to certain tasks that are simple for us humans, computers struggle. Recognizing faces or distinguishing between a mountain and the ocean requires a lot of processing power and energy for computers, and even then, the results may not be perfect.

That’s why the Department of Energy has funded the Quantum Materials for Energy Efficient Neuromorphic Computing (Q-MEEN-C) consortium, led by the University of California San Diego. The goal of this research is to create brain-like computers with minimal energy requirements, which could revolutionize various aspects of modern life.

The first phase of the research focused on finding ways to create or mimic the properties of a single brain element, such as a neuron or synapse, in a quantum material. UC San Diego Assistant Professor of Physics Alex Frañó, along with Professor of Physics Robert Dynes and Professor of Engineering Shriram Ramanathan, played significant roles in this phase.

Now, in the second phase, the Q-MEEN-C team has made an exciting discovery. They found that passing electrical stimuli between neighboring electrodes can also affect non-neighboring electrodes. This phenomenon, known as non-locality, is an essential milestone in the development of new devices that can mimic brain functions, also known as neuromorphic computing.

The brain operates with frequent and minimal exertion non-local interactions, and replicating these behaviors in synthetic materials is challenging. However, the team at Q-MEEN-C ran calculations during the pandemic and found that non-locality in quantum materials was theoretically possible.

When the labs reopened, they enlisted the help of UC San Diego Jacobs School of Engineering Associate Professor Duygu Kuzum, who assisted in turning a simulation into a real device. They achieved this by using a thin film of nickelate, a quantum material ceramic with rich electronic properties. By inserting hydrogen ions and applying an electrical signal, they created a memory-like device that can retain new configurations even after the signal is removed.

The design concept from Q-MEEN-C simplifies the creation of networks that transport electricity. Unlike traditional complicated circuits with continuous connection points, the non-local behavior allows all the wires in a circuit to communicate without direct connections. It’s similar to how movement in one part of a spider web can be felt across the entire web.

The brain learns in complex layers, creating connections in multiple areas, enabling us to differentiate various things. Currently, software programs like ChatGPT and Bard can simulate these pattern recognition tasks, but they require advanced hardware support to reach their full potential.

The hardware revolution in the field of AI is needed to complement the ongoing software advancements. The ability to reproduce non-local behavior in synthetic materials brings us one step closer to that revolution. The next phase of the research will involve creating more complex arrays with more electrodes and elaborate configurations.

This important step in understanding and simulating brain functions opens up the possibilities for a new paradigm in artificial intelligence. To truly unlock the potential of AI, we need hardware that can perform tasks alongside the software.

The research conducted by Q-MEEN-C is primarily supported by Quantum Materials for Energy Efficient Neuromorphic Computing, funded by the U.S. Department of Energy. This work marks a significant advancement in the field, and the researchers are optimistic about the future of brain-like computers.

Source link

Stay in the Loop

Get the daily email from AI Headliner that makes reading the news actually enjoyable. Join our mailing list to stay in the loop to stay informed, for free.

Latest stories

You might also like...