AWS Certified AI PractitionerAI/ML and Generative AI FundamentalsEasy
A team is developing a generative AI model to create new musical compositions. They want the generated music to exhibit a high degree of creativity and unexpected variations, rather than strictly adhering to common musical patterns. Which parameter should they adjust to encourage more diverse and surprising outputs?
- ATemperature
- BNumber of Epochs
- CBatch Size
- DLearning Rate
Show answer & explanationAnswer & explanation
Correct answer: A. Temperature
Temperature is a parameter used in generative models, especially language models, to control the randomness of the output. A higher temperature value (e.g., closer to 1.0 or higher) makes the model's output more diverse, creative, and sometimes unpredictable, while a lower temperature (closer to 0) makes it more deterministic and conservative.
Why the other options are wrong
- B. Number of epochs refers to how many times the entire training dataset is passed through the model; it affects how well the model learns, not the creativity of its generation at inference time.
- C. Batch size determines the number of samples processed before the model's internal parameters are updated; it impacts training stability and speed, not directly output creativity.
- D. Learning rate controls the step size at which the model's weights are updated during training; it affects convergence, not output creativity.
Temperature Parameter (Generative AI)
A parameter that controls the randomness and creativity of a generative AI model's output, typically applied during sampling from probability distributions.
- Higher temperature = more random/creative/diverse output.
- Lower temperature = more deterministic/conservative/focused output.
- Often used in Large Language Models (LLMs) and other generative models.
- Affects the probability distribution over possible next tokens/outputs.
Memory trick: Control the flow for creative glow.