Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
Ultimate access to all questions.
As a Machine Learning Engineer at a leading financial institution, you've developed a binary classification model using Google Cloud's AutoML Tables to predict the likelihood of customers making timely loan repayments. This model plays a critical role in the loan approval process. Following a model's decision to reject a loan application, the risk management team requests a detailed explanation to understand the factors influencing this decision. The explanation must not only clarify the model's reasoning but also adhere to regulatory compliance requirements for transparency in automated decision-making. Given these constraints, which method should you employ to provide a comprehensive and compliant explanation for the model's decision? Choose the best option.
Explanation:
Local interpretability techniques like SHAP or LIME are the most effective for explaining individual predictions, as they provide insight into how each feature influenced the model's decision for a specific case. This approach meets the need for detailed, compliant explanations required by regulatory standards. Option E is also correct because combining global and local methods offers a comprehensive understanding of the model's behavior, both in general and in specific instances, enhancing transparency and trust in the model's decisions.