AlexNet
AlexNet is a landmark deep convolutional neural network that revolutionized computer vision in 2012. Developed by Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton, it won the ImageNet competition by a dramatic margin, demonstrating that deep learning could dramatically outperform traditional feature extraction methods.
The network's architecture—featuring eight layers with convolutional and pooling stages—was trained on GPUs, making large-scale image processing feasible. AlexNet used neural network innovations like ReLU activation, dropout regularization, and data augmentation to prevent overfitting on millions of images.
This breakthrough catalyzed the modern deep learning era, showing that scaling up networks and data could unlock previously impossible capabilities in image recognition, object detection, and beyond. It proved that neural networks deserved renewed investment after decades of relative stagnation, fundamentally reshaping artificial intelligence research and practice.
AlexNet's success triggered a cascade: competitors raced to improve on its design, hardware manufacturers optimized for neural network workloads, and academia embraced deep learning as the dominant paradigm for pattern recognition and machine learning.