SDS 444: Future-Proofing Your Career — with Jon Krohn

Interpretability Tools

In the realm of data science, model interpretability is gaining significant attention. Jon Krohn highlights the importance of using open-source tools like LIME and SHAP to enhance the transparency of machine learning models. These tools help data scientists ensure that their models are not only accurate but also understandable and free from biases.

Automl may recommend an extravagant, deep learning model as optimal for accurately solving a problem. But if interpretability is paramount to your application, say, because your model will approve people for a credit card or determine the length of their prison sentence, then a simple regression model with marginally less accuracy, perhaps, but completely interpretable model weights might be much more appropriate.

--- Jon Krohn

Ultimately, the responsibility lies with the data scientist to choose the right balance between model complexity and interpretability 1.

   

Bias and Fairness

Ensuring fairness and eliminating bias in machine learning models is a critical task for data scientists. Jon Krohn discusses how interpretability tools can aid in identifying and mitigating biases, particularly those affecting demographic groups. By leveraging these tools, data scientists can create models that are not only effective but also equitable.

Ultimately, the data scientist themselves is responsible for ensuring that a model is sufficiently interpretable and doesn't include unwanted biases, such as those against a particular demographic group.

--- Jon Krohn

This approach is crucial when models are used in sensitive applications, such as credit approval or legal sentencing, where fairness is paramount 1.