A neural network, also called an artificial neural network (ANN), is an AI architecture loosely inspired by the structure of the human brain, made up of layers of interconnected nodes, or "neurons," that process and pass information to one another. Each connection between nodes has an adjustable weight, and the network learns by repeatedly tuning those weights based on how close its outputs come to the desired result.
Neural networks form the foundation of deep learning, and different architectures of neural network are suited to different kinds of problems: convolutional neural networks excel at image-related tasks, and transformer-based neural networks, built around the attention mechanism, are what power today's large language models like ChatGPT. Despite the brain-inspired terminology, a neural network's "neurons" are simple mathematical functions, not biological cells, and the analogy is best understood loosely rather than literally.
