October 16 2024
By Minhoo Jeong
The Nobel Prize in Physics for 2024, awarded by the Royal Swedish Academy of Sciences on October 8, went to John J. Hopfield and Geoffrey E. Hinton in recognition of their work in artificial intelligence first in the history of the prestigious physics prize. The Nobel Committee cited the pair for enabling of machines to have cognitive abilities inspired by the human brain, thus laying the foundation for modern machine learning and AI. Hopfield developed the theory of the associative memory networks that store and recall patterns. Hinton invented techniques that allow the autonomous extraction of features from data. Both scientists contributed much to deep learning, which literally swept many areas, such as self-driving cars, speech recognition, and medical diagnostics.
Hopfield's invention-the "Hopfield Network" used human memory and played a pivotal role in pattern recognition and optimization. Hinton worked on the backpropagation algorithm in 1986 that can let efficient learning take place in deep neural networks. Further, he developed more state-of-the art technologies in deep learning using his Restricted Boltzmann Machines and dropout techniques to enhance neural network performance and generalization. His more recent work, such as capsule networks, aimed to overcome some limitations of convolutional neural networks and push the boundaries even further with regard to AI.
Their innovations have ranged from impacting narrow areas like AI to interdisciplinary ones like climate science, medicine, finance, and biology. While giving his acceptance speech, Hinton showed surprise but seemed more terrified because he was able to demonstrate that AI is getting smarter faster than expected and called for caution in the creation and usage of AI. On the other hand, Hopfield stressed the necessity of the importance of inter-disciplinarity and underlined what physics contributes to the evolution of AI.