October 15, 2024
📰 FEATURE STORY
Were the Physics and Chemistry Nobel Prizes to AI researchers the right call?
If you look into some basic definitions of Artificial Intelligence (AI), you’ll probably think it comes under technology or computer science. But is it ‘traditional’ science? Is it a science like biology, physics, or chemistry? If this year’s Nobel Prizes in physics and chemistry are anything to go by, AI can be seen as a scientific subject. But that has not been widely accepted yet.
In the aftermath of the announcements by the Nobel Committee, there’s a debate on whether pioneers in AI should be awarded the Physics Prize. It’s one thing to use AI technology as a tool in research. But recognising work on the building blocks of something like AI has ruffled some feathers. Did the Nobel Committee make the right call?
Context
Google is no doubt proud following last week’s announcements. Three researchers with links to the tech giant won Nobel Prizes for their work on AI. It may cement the company’s position as a leader in this technology. Two of the three people who won the prize in chemistry – Demis Hassabis and John Jumper – are scientists of Google’s AI lab called DeepMind. Geoffery Hinton, part of the duo that won the Physics prize, was a VP at Google until last year. Often called the godfather of AI, he famously quit to warn about the potential dangers of AI.
Let’s look at what they won for. The Nobel Prize in Physics went to John J. Hopfield and Geoffrey E. Hinton for their “foundational work in neural networks.” They published papers in the 1980s that described rudimentary neural networks. It wasn’t as advanced as the networks used for modern AI tools like ChatGPT, but their ideas were pathbreaking.
Neural networks are a model of computing that mimics the human brain’s structure and functions to process data and information. One of the components is called “backpropagation” which allows scientists to train neural networks.
The Nobel Prize in Chemistry was given to biochemist David Baker and computer scientists Demis Hassabis and John Jumper. The latter two are from DeepMind in the UK.
Both prizes have a connection. Hinton helped develop an approach used by DeepMind to make its breakthrough in predicting the shapes of proteins.
For better or worse, the fact that some of the world’s most prestigious scientific awards were given to private-sector researchers reflects a paradigm shift. It seems the Nobel Committee sees AI as important and shows us what the future of scientific research looks like.
But that begs the question: Is AI physics? Is it chemistry? The Nobel Committee saw it fit to award people whose work was instrumental in shaping AI today. But does that mean we should treat AI on par with sciences like physics and chemistry?
VIEW: There are connections
Through the decades, the Nobel Prizes for Physics and Chemistry have been awarded to ‘traditional’ and worthy accomplishments in particle physics, quantum mechanics, cosmology, etc. But science seldom stands still, and technology has crept into all our lives. As both continue to evolve, there’s a compelling case to expand the scope of the Nobel Prizes to include contributions to AI and other emerging technologies. It’s because interdisciplinary research is becoming increasingly important. Breakthroughs occur at the intersections of traditional fields.
Neural networks draw inspiration from the physical structure and functions of brains. Many of the mathematical models underlying artificial neural networks have parallels with systems in statistical physics. One in particular is called the ‘spin glass theory’, a branch of statistical mechanics. Another concept is the ‘energy landscape’. It helps describe the space of complex physical systems and finds application in the training dynamics of neural networks.
The biology prize also highlights how AI and science intersect. It’s a real-world example of how AI is helping people today with discoveries in protein structure. Google’s parent Alphabet developed an AI and machine learning model called AlphaFold that can predict the structure of proteins in just minutes. It can help scientists decide which bits of protein to put in a vaccine that is effective against diseases caused by shape-shifting parasites, like malaria, for example. If you’re a traditionalist, then the Nobel Committee expanded the field of science too broadly. But it’s clear which way the wind is blowing.
COUNTERVIEW: Loose connections at best
After the Committee announced the winners, some in the scientific field felt their territory was being crept upon. Jonathan Pritchard, an astrophysicist at Imperial College London, said, “The Nobel Committee got hit by the AI hype”. He didn’t understand how neural networks were a physics discovery. Sabine Hossenfelder, a physicist at the Munich Center for Mathematical Philosophy in Germany, felt the same way and said the work comes under computer science.
That’s the sense among some in the scientific field. The work done by the researchers was noteworthy. It’s just not rooted in traditional science. Even if there’s inspiration from physics, no new theory was developed, or a long-standing physics problem or equation wasn’t solved. Physics, for example, is often seen as tied to something physical. It’s a branch of science concerned with nature and the properties of non-living matter and energy. That’s not what AI is – which is related to something happening in the mind of a computer and not a physical being.
Hinton himself said in an interview that his work would be more appropriate for a Nobel Prize in computer science. The concept of computer science didn’t exist when the awards were created over 100 years ago. The Nobel Foundation hasn’t indicated that it will change the traditional categories. The 1968 addition of an award for economic sciences was an exception. Perhaps that’s what the Committee should look at – a new category instead of folding in different works.
Reference Links:
- With AI warning, Nobel winner joins ranks of laureates who’ve cautioned about the risks of their own work – CNN
- When AI looked at biology, the result was astounding – The Washington Post
- AI was central to two of 2024’s Nobel prize categories. It’s a sign of things to come – The Conversation
- Why the Nobel Prize in Physics Went to AI Research – IEEE Spectrum
- Google’s Nobel prize winners stir debate over AI research – Reuters
- AI comes to the Nobels: double win sparks debate about scientific fields – Nature
- A.I.’s Sweep Reignites Debate Over Whether It’s Time to Modernize the Nobel Prizes – Observer
What is your opinion on this?
(Only subscribers can participate in polls)
a) The Physics and Chemistry Nobel Prizes to AI researchers were the right call.
b) The Physics and Chemistry Nobel Prizes to AI researchers were the wrong call.
Previous poll’s results:
- Prashant Kishor’s Jan Suraaj party can succeed in Bihar: 62.5% 🏆
- Prashant Kishor’s Jan Suraaj party can’t succeed in Bihar: 37.5%
🕵️ BEYOND ECHO CHAMBERS
For the Right:
Government should reform a regime of highly ineffective food and fertiliser subsidies
For the Left:
BJP has changed the destiny of J&K. Abrogation of Article 370 led us here