Across
- 2. Input given to an AI
- 4. Information used to teach a model
- 6. Learning with correct answers provided
- 9. Process of improving model performance
- 11. Software that performs intelligent tasks
- 12. Result produced by a model
- 18. AI that creates new content
- 21. Information sent into a system
- 22. Type of model that works with text
- 23. Direction used to update model learning
- 26. Increasing model size or capacity
- 27. Learning through rewards and penalties
- 29. Process of cleaning noisy data
- 31. Repeated structure found in data
Down
- 1. Model guess about what comes next
- 2. Adjustable values that influence model behavior
- 3. Numerical representation of meaning
- 5. Likelihood that a prediction is correct
- 6. Ordered set of tokens processed by AI
- 7. Confident but incorrect AI output
- 8. Extra information used by a model
- 10. Step-by-step procedure for solving a problem
- 13. Finding patterns without answers
- 14. Model that removes noise from data
- 15. Trained system that makes predictions
- 16. When a trained model generates an output
- 17. Mathematical representation used in models
- 19. Software used to complete tasks
- 20. Process of learning from examples
- 24. Method of learning by hiding data parts
- 25. Structured instruction to a system
- 28. Small unit of text processed by AI
- 30. Representation used for tokens