Designing an agent which can navigate the procedurally generated, ascii dungeons of Nethack 🏆
- Develop a policy that favors the agent almost all the time.
- Model based vs Model free, battle for best
- Mix N Match different policies
- Make three independent agents where the third is a mix of the two.
- If using NN, optimize it using Neural Magic - Reduces inference time
- Maybe use stable baselines
- Working on creating a player using path finding algorithms (SSSP, ASSP etc) 💥
- Not done
- Figure out how to deal with secret passage (Tentatively dealt with!) ▪️ 🔻
- Add tests