Comments on Leo Breiman's paper 'Statistical Modeling: The Two Cultures' (Statistical Science, 2001, 16(3), 199-231)

TL;DR

This paper critically examines Breiman's call for embracing model-free learning, highlighting the current challenges in understanding the inferential aspects of modern machine learning methods amidst their practical success.

Contribution

It provides a reflective analysis on the evolving role of statistical inference in the era of deep and machine learning, emphasizing unresolved questions about understanding and trust.

Findings

Modern machine learning excels in prediction but lacks interpretability.

The role of inference and explanation in machine learning remains an open challenge.

Statistical community is reconsidering the balance between prediction and understanding.

Abstract

Breiman challenged statisticians to think more broadly, to step into the unknown, model-free learning world, with him paving the way forward. Statistics community responded with slight optimism, some skepticism, and plenty of disbelief. Today, we are at the same crossroad anew. Faced with the enormous practical success of model-free, deep, and machine learning, we are naturally inclined to think that everything is resolved. A new frontier has emerged; the one where the role, impact, or stability of the {\it learning} algorithms is no longer measured by prediction quality, but an inferential one -- asking the questions of {\it why} and {\it if} can no longer be safely ignored.