Letter #20 - In which we’re not really curious at all.
We spend our entire lives trying to tell stories about ourselves — they’re the essence of memory. It is how we make living in this unfeeling, accidental universe tolerable.
“Now I hate to leave you bare
If you need me I’ll be there
Don’t you ever let me down”
The definition that OpenAI team used for artificial curiosity was relatively simple: The algorithm would try to predict what its environment would look like one frame into the future. When that next frame happened, the algorithm would be rewarded by how wrong it was. The idea is that if the algorithm could predict what would happen in the environment, it had seen it before.
Wait until it learns how to eat chips :). Full paper is here.
Text classification offers a good framework for getting familiar with textual data processing without lacking interest, either. In fact, there are many interesting applications for text classification such as spam detection and sentiment analysis. In this post, we will tackle the latter and show in detail how to build a strong baseline for sentiment analysis classification. This will allow us to get our hands dirty and learn about basic feature extraction methods which are yet very efficient in practice.
Quick and dirty intro to text classification using the IMDB Movie Reviews dataset. Doesn’t go into the state of the art techniques, but that’s very much ok.
Not sure what you just read? Take a look at this post.
Enjoyed it? Subscribe to my tiny newsletter and get new issues straight to your inbox. Every week, on Tuesdays.