I work on optimization problems and machine learning experiments. Most of this is just personal curiosity about mathematical approaches that interest me.
These are experimental projects - some approaches work, others lead to interesting dead ends. Code and documentation are available for each project.
Working through optimization problems that don't have clean solutions. Mostly experimenting with different approaches to see what actually works.
Making smaller RL models learn from bigger ones. Useful when you can't afford to run massive networks.