Explainable AI Explained for Developers • Joop Snijder & Willem Meints • GOTO 2022

I resumed my trips to fitness, where I usually do cardio on an elliptical trainer. And because pedaling is so boring, I usually get to watch a talk from a conference. That's how I saw it yesterday. Report on Explanatory AI The last GOTO conference was held in the Netherlands. I liked the report. Speakers told how ML is becoming ubiquitous Then they started talking about how ML can help a fake company that sells coffee by subscription. The story moved to the demo on using Auto ML in Azure to predict the outflow of subscribers And then the authors began to ask themselves what to do with the results of the predictions -- that's how they approached the stated theme, which is that it usually requires the ability to understand the results.

All the questions in the talk are addressed at the level of immersion that the average developer has without the deep ML details and all this:) And at the exit, listeners get a good understanding that now ML is also striving towards commodity. (Especially with the advent of Auto ML.)But the issue of data storage, their preparation, model deployment and interpretation of the results requires certain skills from engineers who are not yet solved with the help of Auto ML:)

#ML #Conference #Software