To teacher and TA
Here is the result about "The Second Experiment".
The code is in folder "new_code". There are annotations in the code
Following information are other detail things.
AI-Project
Codes
A simple model
We start with a simple model. The model has six states which is mentioned in paper "Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation".
Our new codes are in folder "random_idea". We test methods such as drawing subgoal randomly, drawing the farthest state.
Another simple model
New code is in folder "new_code".
Other codes
We copy and modify some existing experiments, such as Flappy Bird (in folder "DeepLearningFlappyBird-master") and Data-Efficient HRL (in folder "efficient-hrl").
Reports
Proposal is in folder "proposal".
Progress report is in folder "progress".
References
Some references are in folder "refs", "stochastic NN" and "other_refs (maybe not so related)".
Game
The game MontezumasRevenge is in folder "game". It supplies to be played by humans.