Carragee Consulting

Machine Learning changed how I see the world

Mindful Modeler – Substack

This definition is very broad and covers many common tasks:

– Time series forecasting: The past is available, but the future is not.

– Image classification: The pixels are available, but the label is not.

– Text-to-text generation: The prompt is available, but the answer is not.”

This short piece from Christoph Molnar has synthesized a bunch of my recent thoughts on the influence of education.

When I think back on my education I value the broad scope of my liberal arts degree, with a major focus on discussion and expository writing.

Choosing an economics major felt like electing to understand the world around me through models, sometimes mathematical and more often through logic and decision flows.

I’ve now been away from formal education for longer than I participated in it, but having learned to learn within a set of frameworks has set me up to seek out and process new information through a particular lens, constantly trying to assimilate new information into an existing mental model.

Continuing to read across disciplines and enjoying learning from others cross threads my thinking and compounds the issues I face in simplifying and expressing my thoughts to others.

Being one of the people who can bridge across professional paradigms and try problems out from various perspectives is itself a valuable skill, but almost ensures my expertise lacks technical depth.

That you can consciously change your way of thinking and understanding the world around you through formal education and training is exactly the point, but somehow the way Christoph frames the results in action seems novel.

Beginning to understand the mechanisms at work in Machine Learning, the common flaws and limitations start s to feel like wearing X-ray goggles. You can see the skeletons underpinning many problems and diagnose ways to approach them.

Is this another signal to reread ‘Algorithms to Live By: the Computer Science of Daily Living” to recognize the patterns around me? Or a call to action to delve deeper into the coding, modeling and application of ML for the problems and data as I already view them?