Question: An artificial intelligence design team performs the six steps of model development. Which of the following are steps of model development? Select all that apply.
- Collect data with which to train the model
- Fill in missing data or remove biased data that might mislead the model
- Testing the model one time for a complete, comprehensive understanding of its performance
- Assess performance to ensure accurate predictions will be made outside of the training environment
Explanation
Model development begins with collecting data that is relevant to the task the model is being trained to perform. Data preparation can include fixing missing values, removing poor-quality examples, and addressing bias that could affect model behavior. Training uses prepared data so the model can identify patterns and improve predictions. Performance evaluation checks whether the model works beyond the training data and can generalize to new examples.
Why the other options are incorrect
Testing the model one time is incorrect because model performance requires evaluation beyond a single check.
Source for verification
https://developers.google.com/machine-learning/crash-course/overfitting/dividing-datasets
https://developers.google.com/machine-learning/crash-course/overfitting/interpreting-loss-curves
The answer(s) to the question is highlighted in the BOLD text above. You can also find more questions and answers related to the exams on the "Generative AI for Educators" page.