No More Hustleporn: How to Work on Machine Learning

Tweet by Brandon Rohrer

Data science. Algorithm and feature design. he/him. Currently at iRobot, formerly of Facebook, Microsoft, DuPont Pioneer, Sandia Natl Labs, MIT.

Feeling bored with machine learning?
Here's how to bring back the magic.

1. Choose a problem family. Bonus points if it's something other than image classification or next-word prediction.

2. Choose an application area. Bonus points if it's something other than e-commerce.

3. Choose a problem to solve. Bonus points if it makes someone's life better.

4. Pull together your own data set. Bonus points if you have to make some of the measurements yourself.

5. Try to solve the problem. Existing methods probably won't work well off the shelf. Customize them to your application. Modify and extend them with abandon.

6. Repeat.