What a Deep Neural Network Thinks About Your #Selfie
Analysis
This article describes a fun experiment using a Convolutional Neural Network (ConvNet) to classify selfies. The author, Andrej Karpathy, plans to train a 140-million-parameter ConvNet on 2 million selfies to distinguish between good and bad ones. The article highlights the versatility of ConvNets, showcasing their applications in various fields like image recognition, medical imaging, and character recognition. The author's approach is lighthearted, emphasizing the potential for learning how to take better selfies while exploring the capabilities of these powerful models. The article serves as an accessible introduction to ConvNets and their applications.
Key Takeaways
- •The article describes an experiment using a ConvNet to classify selfies.
- •The experiment aims to train a ConvNet on a large dataset of selfies.
- •The article highlights the versatility and applications of ConvNets in various fields.
“We’ll take a powerful, 140-million-parameter state-of-the-art Convolutional Neural Network, feed it 2 million selfies from the internet, and train it to classify good selfies from bad ones.”