AI Terminology

1234567891011121314
Across
  1. 5. mechanism to focus on parts of input (Transformer core)
  2. 8. Rectified Linear Unit activation function
  3. 9. algorithm (e.g. Adam) that updates model weights
  4. 10. using a trained model to predict on new data
  5. 12. basic processing unit in a neural network
  6. 13. trained function mapping input → output
  7. 14. one full pass over the training dataset
Down
  1. 1. adapt a pretrained model on new data
  2. 2. reduce overfitting by adding penalty/constraints
  3. 3. gradient method used to train neural networks
  4. 4. activation that produces a probability distribution
  5. 6. to fit a model to data
  6. 7. an attention-based neural-network architecture
  7. 11. dense vector representation of discrete items