CPMAI_V7 practice question 12 of 35

Your team is using a neural network algorithm to generate a Machine Learning Model. What specific artifacts need to be included? (Select all that apply.)

Full set: 100 questions

Answer options

A: The algorithm code Correct answer

B: Supporting training data Correct answer

C: Bias-variance tradeoff

D: Hyperparameter settings Correct answer

Correct Answer:

A. The algorithm code; B. Supporting training data; D. Hyperparameter settings

Explanation

A neural-network model must be reproducible. Essential artifacts therefore include: (A) the precise algorithm/architecture code, (B) the complete training data set or immutable pointer to it, and (D) the chosen hyper-parameter values (e.g., learning rate, batch size). Together these allow independent rebuilding, auditing, and performance validation of the model. The bias-variance trade-off is an analytical concept, not a storable artifact.

Why Incorrect

C. Bias-variance trade-off is an evaluation concept, not a discrete file or configuration that can be version-controlled as an artifact.

References

1. Cognilytica. “CPMAI v7 Methodology Handbook,” §5.2.3, pp. 112-113.

2. Goodfellow, Bengio & Courville. “Deep Learning,” MIT Press, 2016, §6.4, p. 208-210 (hyper-parameters).

3. MIT OCW 6.036, “Reproducibility and ML Pipelines,” Lecture notes, Spring 2020, pp. 4-6.

Keep practicing

Take the free CPMAI_V7 practice test

Ten exam-style questions with answers and explanations, plus the exam facts and study guides.

Start the free practice test