How Nobel Laureates’ Work On AI, Machine Learning Define Modern Internet, Why They Warn Of ‘Consequences’ – News18

It wouldn’t be wrong to say that Artificial Intelligence (AI) has taken over the world. Today, from videos to fraud detection system, dictating a text while doing your other chores, all seem seamless. Thanks to the scientists, mathematicians and engineers who have revolutionised everyday lives.

Two of them – Princeton University physicist John J Hopfield and University of Toronto computer scientist Geoffrey Hinton – won the Nobel Prize in Physics for their pioneering work in AI.

The award comes with a prize sum of 11 million Swedish crowns ($1.1 million) which is shared by the two winners.

The artificial neural networks are modelled on biological neural networks, and both researchers’ work drew on statistical physics, hence the award is in physics.

The machine learning breakthroughs of Hopfield and Hinton “have showed a completely new way for us to use computers to aid and to guide us to tackle many of the challenges our society face,” the Nobel committee said.

Decoding Their Contributions in the Field of AI

The scientists’ research on neural networks in the 1970s and early 1980s paved the way for today’s deep-learning systems that have become part of the modern internet.

Neural networks are systems that learn skills by analysing data and are named after the web of neurons in the human brain. They are a part of everyday internet services, including Google, Apple’s Siri and OpenAI’s ChatGPT. These services are rooted in mathematics and computer science, not physics.

Hopfield used ideas from physics to study a particular type of recurrent neural network, now called the Hopefield network. Hopfield was a pioneer in using models from physics, especially those developed to study magnetism, to understand the dynamics of recurrent neural networks. He also showed that their dynamics can give such neural networks a form of memory, according to The Conversation.

“When you get systems that are rich enough in complexity and size, they can have properties which you can’t possibly intuit from the elementary particles you put in there,” he said in a press conference convened by Princeton. “You have to say that system contains some new physics.”

In the 1980s, Hinton extended Hopfield’s ideas to create a new class of models called Boltzmann machines, named after the 19th century physicist Ludwig Boltzmann.

Unlike Hopefield networks that could store patterns and correct errors, Boltzmann machines could generate new patterns, thus, setting the base for the modern generative AI revolution.

Hinton was also part of another breakthrough called backpropagation that happened in the 1980s. Backpropagation is part of machine learning and AI algorithm that trains neural networks and use weights based on the performance of the network.

In 2000s, Hinton and his co-workers used Boltzmann machines to train multilayer networks by fine-tuning algorithm on top of the pretrained network to further adjust weights, as per The Conversation. This led to deep learning revolution as we see it today.

About Hopfield and Hinton

John J. Hopfield, a Chicago native, is an emeritus professor at Princeton known for seminal discoveries in computer science, biology and physics. He is 91, and the third oldest Nobel physics laureate.

He began his career at Bell Laboratories in 1958 as a physicist studying the properties of solid matter. Then, he moved to the University of California, Berkeley, as an assistant professor in 1961 and joined the physics faculty at Princeton in 1964. Sixteen years later, he moved to the California Institute of Technology as a professor of chemistry and biology, and in 1997, returned to Princeton, this time in the department of molecular biology.

Geoffrey E. Hinton, born just outside London, has lived and worked mostly in the US and Canada since the late 1970s. Hinton, 76, began researching neural networks as a graduate student at the University of Edinburgh in the early 1970s.

Hinton, who is also called the ‘Godfather of AI’, joined Google after breakthrough in backpropagation in 2012. More recently, he received the Turing Award in 2018.

What About the Consequences of AI Technology?

After receiving the Nobel Prize, Hinton voiced concerns on the risks AI poses to humanity. Hinton predicted that AI would eventually have a “huge influence” on civilization, with profound benefits in areas like productivity and healthcare. “It would be comparable with the Industrial Revolution,” he remarked in a discussion with the Royal Swedish Academy of Sciences. However, despite these potential advantages, Hinton warned though there is no experience of “what it’s like to have things smarter than us; it’s going to be wonderful, but we also have to worry about a number of possible bad consequences, particularly the threat of these things getting out of control”, as quoted by the Business Insider.

Hinton feared that advanced AI systems could eventually surpass human intelligence and “take control”. He underlined the importance of ethical considerations and responsible development in AI, calling for increased collaboration between scientists, policymakers, and industry leaders to establish robust safeguards.

Hopfield also echoed Hinton’s views on the unknown potential and limits of AI. “One is accustomed to having technologies which are not singularly only good or only bad, but have capabilities in both directions,” he said, as quoted by Reuters.