Glossary

This glossary contains key machine learning and deep learning terms, with each entry on its own page for detailed explanations.

  • Logit - direct result of the linear operations (weighted sums of inputs plus bias) in the final dense layer of a neural network
  • Mini-Batch Gradient Descent - Processing data in batches to optimize neural network training
  • Normalization - The process of centering data and equalizing variance across features
  • ReLU - Simple and widely-used activation function
  • Tensor - Mathematical concept representing multi-dimensional data structures