Projects I’m actively working on
- Building a robot using an ardunio controller and creating a control system using reinforcement learning algorithms
Projects I might work on one day
- Testing Style Transfer for the discriminator model trained in my efforts to get a GAN to create images of faces.
- Using a Multi Scale Gradient GAN to generate doodles in the style of me.
- Using one of policy gradient or actor critic to train bipedal walkers built in the genesistubs-physics-engine.
- I’ve decided to abandon this. The main reason being that I’m suffering sunk cost fallacy w.r.t the physics engine. I feel like becuase I’ve made it I need to use it but it’s turning into a giant time sink and I really want to focus on other things I can see greater value in.