Fake AI and ML Tipps – Friday Faves

This week in the Friday Faves, we’re looking at some AI snake oil and a tip for training better models. A bit shorter, as I’m updating my website and my CV after handing in my PhD thesis.

AI Snake Oil

First slide of the AI Snake Oil talk by Arvind Narayanan

So if you want to call something that solves a problem somewhat and has an if statement, you can call it AI. But maybe don’t do it on things that basically amount to “sorting humans”. Sure people always tried to do it, but phrenology was bad, eugenics was bad, and 20% on an assessment on the big five is a random number generator (which possibly outputs nicer numbers for some people than others).

https://www.cs.princeton.edu/~arvindn/talks/MIT-STS-AI-snakeoil.pdf

Stop the Augmentation

If you’re training your model, you may want to consider stopping the augmentation for the last couple of epochs to improve the final performance.

The following two tabs change content below.
... is a geophysicist by heart. He works at the intersection of machine learning and geoscience. He is the founder of The Way of the Geophysicist and a deep learning enthusiast. Writing mostly about computational geoscience and interesting bits and pieces relevant to post-grad life.

Latest posts by Jesper Dramsch (see all)

Posted in Computer Science and tagged , , , , .