Why in the News?

The Royal Swedish Academy of Sciences awarded the 2024 Nobel Prize in Physics to John J. Hopfield and Geoffrey E. Hinton for their foundational contributions to artificial intelligence through machine learning and artificial neural networks. Their groundbreaking research from the 1980s has profoundly influenced AI, shaping technologies essential for modern data processing and pattern recognition.

John J. Hopfield

• He is credited with developing the Hopfield network, a type of recurrent neural network.
• Its neurons learn and process information based on Hebbian learning — an idea in neuropsychology that if one neuron repeatedly triggers a second, the connection between the two becomes stronger.
• The rules of a Hopfield network are based on the physics of a group of atoms, each producing its own small magnetic field.

John J. Hopfield

Geoffrey E. Hinton

• He made a breakthrough in the 2000s by developing a learning algorithm for a modified ANN called Restricted Boltzmann Machine (RBM).
• A layer of neurons could be trained as an RBM and multiple layers could be stacked, creating the first ANNs capable of deep learning.

Geoffrey E. Hinton

Artificial Intelligence (AI)

• It refers to the simulation of human intelligence in machines programmed to think and learn.
• It involves various techniques and algorithms that enable computers to perform tasks typically requiring human intelligence, such as understanding natural language, recognizing patterns, and making decisions.

Machine Learning (ML)

• It is a subset of AI that focuses on the development of algorithms that allow computers to learn from and make predictions based on data.
• Instead of being explicitly programmed for each task, they identify patterns and correlations within the data to inform their decisions.

Artificial Neural Networks (ANNs)

• They are a class of ML algorithms inspired by the structure and function of the human brain.
• They consist of interconnected nodes (neurons) organized in layers.
• They are trained using labeled datasets through a process called backpropagation which adjusts the weights of connections based on the error of the predictions.