Across
- 2. Foundation of symbolic AI reasoning.
- 5. Structured body of text data.
- 6. Rule-of-thumb for solving problems.
- 8. Systematic error in model predictions.
- 9. Autonomous actor in reinforcement learning.
- 12. Breaking input into discrete units.
- 13. Generalization without task-specific training.
- 15. Reducing model complexity post-training.
Down
- 1. Competing networks for generating data.
- 3. Direction of steepest model improvement.
- 4. Output of neuron after input processing.
- 7. One full pass through training data.
- 10. Measure of uncertainty in predictions.
- 11. Final layer in sequence generation.
- 14. Function enabling nonlinear decision boundaries.