SDS 539: Interpretable Machine Learning — with Serg Masís

Emerging Trends

Emerging trends in machine learning are reshaping the landscape with innovations like no-code solutions and AutoML. Serg Masís envisions a future where machine learning becomes more accessible through drag-and-drop interfaces, reducing the need for extensive programming. He highlights the integration of these tools with legal and technical frameworks, emphasizing the importance of testing and trusting models rigorously 1.

I see a lot of the machine learning in the future is going to be drag and drop. So that kind of frees our hands from all the programming stuff that we have to do day to day and can lead it to more productive things.

--- Serg Masís

Jon Krohn adds that the future of data science will involve new professions like AI auditors, ensuring models remain effective and ethical even after deployment 2.

   

Ethics & Accountability

Ethics and accountability are becoming central to AI practices, with a focus on fairness, accountability, and transparency. Serg Masís describes interpretable machine learning as a pyramid, where transparency is the base, but fairness and accountability are crucial at the top 3. He is co-authoring a book on responsible AI, emphasizing bias mitigation and detection as key challenges in the field.

I prefer to see interpretable machine learning as a pyramid. It's often seen as having three layers. You have fairness, accountability, and transparency.

--- Serg Masís

Jon Krohn notes that understanding how models make decisions is vital, but addressing fairness and accountability is even more critical for ethical AI development.