AI AI AI
Across
- 2. Making predictions without any task-specific training data.
- 3. The process of adapting a pre-trained model to a specific task.
- 7. Grouping similar data points together without predefined categories.
- 12. A measurable property or characteristic of data used in training AI models.
- 13. The time delay before an AI system produces a result.
- 14. Assigning categories or labels to data using AI models.
- 15. Text-to-Speech, converting written text into spoken words using AI.
- 17. A mathematical representation of a problem, trained on data to make predictions or decisions.
- 19. Splitting text into smaller units like words or subwords for processing.
- 20. Assessing the performance of an AI model.
- 22. AI systems capable of processing and understanding multiple types of data (e.g., text and images).
- 23. A piece of text, such as a word or subword, used in language processing.
- 24. Input given to an AI model to generate a response or output.
Down
- 1. When a model performs well on training data but poorly on new, unseen data.
- 4. A measure used to evaluate the success of a model, like accuracy or precision.
- 5. A sequence of steps for processing data or training models in AI workflows.
- 6. A numerical representation of words or data for use in AI models.
- 8. A type of neural network architecture that powers many LLMs.
- 9. A computing system inspired by the human brain, used in deep learning.
- 10. Automatic Speech Recognition, converting spoken language into text.
- 11. A set of rules or steps used by computers to solve problems or perform tasks.
- 16. Reducing words to their base or root form for NLP tasks.
- 18. A mechanism in AI models that helps focus on the most relevant parts of input.
- 21. A mathematical object used to represent data or features in AI.