Across
- 3. – PyTorch component for automatic differentiation
- 5. – When a model performs well on training data but poorly on new data
- 7. – Regularization technique that randomly ignores neurons during training
- 8. – Small group of data points used in training
- 13. – Technique that reduces overfitting
- 16. – A property/column used in prediction
- 18. – Method used to update weights in neural networks
- 19. – Activation function named for hourglass-shaped curve
- 20. – Common activation function that outputs zero for negative values
Down
- 1. – Algorithm that adjusts model parameters to minimize loss
- 2. – Weight adjustment method during optimization
- 4. – Algorithm that combines predictions from multiple models
- 6. – Activation function used in the output layer for multi-class classification
- 9. – Algorithm for classifying data into clusters
- 10. – Method to choose the best hyperparameters
- 11. – One full cycle of training
- 12. – The answer we want to predict
- 14. – What machine learning learns from
- 15. – Algorithm that finds patterns in data
- 17. – Widespread open-source DL framework