Week 8 updates
This week has been all about taking our project to the next level. First, we upgraded our decision boundary visualization—moving from a 2D graph to a 3D decision boundary graph inspired by a method from a StackOverflow post. This allowed us to plot and understand three state variables at once, which has already helped us interpret the model’s behavior in more depth.
We also successfully found an optimal policy for our current CartPole model. With that in place, we decided it was time to scale up. Rather than continuing to build and train everything from scratch, we began using pre-trained models from Hugging Face’s model hub and GitHub repositories. This gives us the freedom to focus on interpreting and analyzing models more efficiently—and to explore new environments.
Some of the environments we’re planning to work with include Pong, Lunar Lander, and other more complex simulations. This transition will help us test how well our interpretability tools and techniques translate across different challenges.
Alongside this, I’ve started drafting the research paper and outlining the final presentation. It’s exciting to see everything come together, and I’m looking forward to sharing our progress in a more polished format soon.
Thanks again for following along!
Comments:
All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.