Brain-like power: Quantum material paves way for energy-efficient computing

Synthetic Brain? Quantum material's discovery redefines computer efficiency

Web Desk Updated: August 09, 2023 15:26 IST

A team of researchers  from the Quantum Materials for Energy Efficient Neuromorphic Computing (Q-MEEN-C) consortium, led by the University of California San Diego, has unlocked a hidden potential in quantum materials. This potential, known as "non-local" behavior, holds promise in mimicking the intricate workings of the human brain and revolutionizing energy-efficient computing.

The recent strides in quantum material research have unearthed the potential for energy-efficient computing that emulates the brain's functions. The discovery of non-local behavior marks a critical milestone on the path to creating brain-like computers with minimal energy requirements, heralding a new era of computing innovation.

Traditionally, computers have outperformed humans in tasks such as complex calculations and quick data recall. However, the human brain remains unparalleled in its ability to process complex information layers swiftly, accurately, and with minimal energy consumption. Recognizing faces after a single encounter or distinguishing between various objects, like mountains and oceans, showcases the brain's astonishing efficiency.

The journey toward creating brain-like computers with minimal energy requirements has taken a significant stride forward, thanks to the recent discovery of non-local behavior in quantum materials. The breakthrough has opened the door to more energy-efficient neuromorphic computing, where devices emulate the brain's intricate functions.

Professor Alex Frañó, co-director of Q-MEEN-C, explained that non-local interactions, which are common and effortless in the human brain, have been a rarity in synthetic materials until now. This newfound capability paves the way for the creation of more efficient machines capable of sophisticated learning processes.

The research, published in Nano Letters, builds on the consortium's earlier work to emulate brain elements like neurons and synapses using quantum materials. The latest findings reveal that electrical stimuli transmitted between neighboring electrodes can also influence non-neighboring electrodes. This phenomenon of non-locality holds immense significance, marking a critical milestone in the journey toward brain-inspired computing.

The breakthrough was not without its challenges. The pandemic-induced closure of physical lab spaces compelled the researchers to devise innovative methods. Utilizing arrays of devices to simulate neurons and synapses, they confirmed the theoretical possibility of non-locality in quantum materials.

Upon the reopening of labs, the team collaborated with UC San Diego's Jacobs School of Engineering Associate Professor Duygu Kuzum, who helped translate simulations into actual devices. The process involved manipulating a thin film of nickelate, a "quantum material" ceramic with rich electronic properties. By introducing hydrogen ions and applying an electrical signal, the researchers induced a memory-like configuration change in the material. This configuration remained even after the signal was removed, representing a breakthrough step toward creating more conductive pathways for electricity flow.

Innovatively, the design concept deviates from conventional circuitry. Drawing inspiration from the non-local behavior observed in the experiment, the Q-MEEN-C researchers realized that circuit wires need not be continuously connected. This design similarity to a spider web, where movement in one part resonates across the entire structure, offers a more efficient and cost-effective approach to creating networks for electricity transport.

While artificial intelligence (AI) software has made remarkable progress in simulating brain-based activities, the ultimate potential can only be realized when advanced hardware complements the software's capabilities. This hardware-software synergy is vital for pushing the boundaries of AI. Professor Frañó envisions a hardware revolution parallel to the ongoing software advancements, and the successful reproduction of non-local behavior in synthetic materials brings us a step closer to this realization.

As the research progresses, the next phase involves crafting more intricate arrays with additional electrodes, driving us closer to achieving a new paradigm in the realm of artificial intelligence. The convergence of hardware and software innovation promises to unlock a future where machines can learn through physical properties, mirroring the complexity and efficiency of the human brain.