Toxic Language Detection is a *hard* problem. The reason why it is hard has to do with the fact that it is very easy to point out a flaw that GridSearch does not capture. In this video, I'll build and explore a classification pipeline for this task and we'll highlight a very common failure scenario.
You can find the code for this video here;
https://github.com/RasaHQ/algorithm-whiteboard-resources/tree/master/toxic
You can find the code for this video here;
https://github.com/RasaHQ/algorithm-whiteboard-resources/tree/master/toxic
- Category
- Lookbadass
Comments