GUWAHATI: The Royal Swedish Academy of Sciences has awarded the 2024 Nobel Prize in Physics to John J. Hopfield and Geoffrey E. Hinton for their groundbreaking contributions that have laid the foundation for modern machine learning using artificial neural networks. Their discoveries, originating from principles in physics, are at the core of today's powerful machine-learning technologies.
ALSO READ: Florida Braces For The Worst As Hurricane Milton Intensifies
John Hopfield developed a form of associative memory capable of storing and reconstructing images and other data patterns, while Geoffrey Hinton pioneered techniques that enable machines to autonomously identify features within data, such as recognising elements in images.
Artificial intelligence (AI) often refers to machine learning that operates through artificial neural networks. These networks were originally inspired by how the brain functions. In such systems, neurons are represented by nodes with varying values that interact through connections, similar to synapses in the brain. These connections can strengthen or weaken over time. The neural network is trained by reinforcing stronger connections between nodes that consistently show high values. Since the 1980s, both Hopfield and Hinton have made significant contributions to this area.
John Hopfield’s invention revolves around a network that can save and recreate patterns. Imagine the nodes in his network as pixels. The Hopfield network uses physics principles that describe the properties of a material based on atomic spin, a feature that makes each atom behave like a tiny magnet. The network as a whole can be understood in terms of the energy associated with the spin system in physics. The network is trained by adjusting the connections between nodes so that stored images correspond to states of low energy. When fed an incomplete or distorted image, the Hopfield network updates the nodes’ values step by step to lower its overall energy, gradually reconstructing the original saved image that closely matches the distorted input.
