This Repository contains my various attempts and Experiments while working on the projects "Generative AI for Reinforcement Learning Trajectories" and "Generative AI for Predicting Stream Cipher Keystreams".
Some of the files need Keras 3.0 to run while others need Keras 2.0. The files that need Keras 3.0 are all files implementing diffusion. All other files need Keras 2.0.
Most important files are:
Diffusion_D4RL.ipynb - The Final Code of RL task, diffusion model being applied to most complex environment tested which is D4RL locomotion Maze2d large.
LFSR_final_graphs_and_experiments_calculation.ipynb - Calculations and graphs for the final report and presentation of Stream Cipher Part. Also Contains MLP implementation of model
5len_LFSR_Transformer.ipynb - Transformer model implementation of Stream Cipher Part
All other files are for experimentation of various different choices of models and environments. In each one, we have either tried a new implementation or experimented with different things.