Meeting #5
Friday, 4/29/2022
Garbage Collection
RL Goals
- Have unit tests, documentation, and models running by the end of April
Network Splitting
- Growing neural network - start from null to initialization, followed by repeated iteration.
- Performs much worse than a standard neural network for now, but trying classification vs autoencoding.
Environment
- Doing well
Emergent Language
- Good results so far
- Good philosophical grounding
- Expect to get something ‘paper-publishable’ by end of the quarter