The Future of Computing is Physics

Story by:

• Michelle Franklin- [email protected]

Share This:

Article Content

Since its founding, UC San Diego has been at the forefront of physics research. And while you may wonder how neutrinos affect your everyday life, basic physics research shows up in the world in amazing ways, including health care and computer technology. The Department of Energy-funded Quantum Materials for Energy Efficient Neuromorphic Computing (Q-MEEN-C) is led by physics professors Ivan Schuller, Alex Frañó and Oleg Shpyrko, who is also department chair. Here Ivan and Oleg talk about artificial intelligence, the future of computing and why physics researchers play the long game.

Q: What is neuromorphic computing?

Ivan Schuller: Neuromorphic computing is an attempt to develop a computer that works like the human brain. To some extent, conventional computing already does this; however, scaling it up to the capabilities of the human brain would require prohibitive amounts of energy and water consumption.

Oleg Shpyrko: Traditional computing is still needed for precise calculations, but in our everyday lives, we use a lot of non-precise mental processes. Most of us don't factorize numbers on a daily basis, but we use image and pattern recognition when we say hello to familiar faces and know to pet a dog but not a snake. Pattern recognition is an easy task for humans, but a difficult task for machines.

If you think about self-driving cars, the car has to be able to distinguish between a person crossing the street or a garbage bag being blown by the wind. So the idea behind neuromorphic computing is that it can supplement - not replace - what is being done with traditional computers by helping us do the things that the brain is already good at.

Q: How can quantum materials and neuromorphic computing support the development of artificial intelligence (AI) technology?

Oleg Shpyrko: AI is on everybody's mind, but I think what people don't talk about enough is that there's a huge price to pay in terms of the energy required to even train something like ChatGPT 4.0 - much more than the energy required to use it, which in turn requires a lot more energy than a simple Google search.

Schuller: To some extent, efficiencies can be fine-tuned in the computing software, but it won't be enough. This is a hardware problem. That's where quantum materials can help.

Q: What do you see as the future of computing and of the AI revolution?

Shpyrko: To go back to what Ivan said about basic science research, if you think about engineering, they're often focused on incremental improvements: 10% faster, 3% smaller. And engineers are very good at this! But what we're looking for is scientifically-driven, fundamental change.

In the beginning of 20^th century, it was impossible to predict how replacing horses with automobiles would change every facet of life. That's where we are right now. We're looking for the same type of transformational change in how we approach computing. For that reason, I think it's very difficult to predict what the outcomes are going to be, but a revolution is brewing.

Q: Why is Geoffrey Hinton and John Hopfield's Nobel work so important?

Schuller: I think the 2024 Nobel Prize in physics is particularly exciting because it recognizes the convergence of physics with artificial intelligence. Their award highlights how fundamental physics concepts have been applied to create the neural networks that now underpin modern AI systems.

This is not just about theoretical advancements, but also about how these ideas have shaped the technologies we use daily. It's a testament to how breakthroughs in seemingly niche areas of physics can lead to transformative applications in other fields, including machine learning and neuroscience.