Hector Kohler's repositories
LunarLander-v2
Using Stable-Baselines 3 library to train a LunarLander-v2 agent. With Yannis Elrharbi-Fleury : https://github.com/sohio92
aeirl-pytorch
A simple implementation of Auto-Encoding Inverse Reinforcement Learning with PyTorch
gym-decision-trees
Simple continuous-states mdp generator for which the optimal policy is a decision tree of given depth. Interface with openAI Gym.
MorrisPool-ReinforcementLearning
Code and PDF report about reproducing the Morris Pool experiment: model rats behaviour in a pool using reinforcement learning
Bayesian-network-learning
Comparing learning algorithms based on constraints (in progress).
CrossEntropyMethod-VS-PolicyGradient
Simple library based on https://github.com/osigaud/Basic-Policy-Gradient-Labs made for the Master's project: Comparison of Evolutionnary Methods and Reinforcement Learning methods on classical control benchmarks. Supervised by Prof Olivier Sigaud with Damien Legros.
High-Dim-Covariance-Matrix-Estimation
A small library to study some Covariance Matrix Estimators in R. Designed in 2019 as part of my Bachelor's Degree thesis.
MarioNette-pong
Trying MarioNette on Pong
Pareto-Local-Search
PLS algo for knapsack problem.
Prioritized-experience-replay
Implementation of Dyna-Q according to Sutton and Barto. Implementation of Mattar's learning model: https://www.nature.com/articles/s41593-018-0232-z
Segment-Anything-Gym-Wrapper
A gymnasium wrapper for objects states detected with the Segment Anything Model.
TREX-Tree-Reward-EXploration
Using Tree estimators of the MDP models to then count leaves grouping similar transitions and do count-based exploration.
asym-rlpo
Asymmetric methods for partially observable reinforcement learning
centrale-sdm-23-24
Sequential Decision Making in Central, based on RLSS 2019
Forecasting-RNN
A simple time series forecasting library using RNNs
MNIST-DeepLearning
Code image classification of MNIST dataset using different architectures: simple linear NN, autoencoder, and highway network
primenet
Predicting prime numbers as list of bits.
rlberry
An easy-to-use reinforcement learning library for research and education.