Proposes a hierarchical world model (HWM) which yields improved sample efficiency and final performance of model-based RL, while building toward a human-cogn...
CleanRL is an open-source library that provides high-quality single-file implementations of Deep Reinforcement Learning algorithms. It provides a simpler yet...
Code-level optimizations, which are low-level optimization techniques used in the implementation of algorithms, have generally been considered as tangential ...
Reinforcement learning methods achieve performance superior to humans in a wide range of complex tasks and uncertain environments. However, high performance ...
An attempt at expanding upon the theory and motivation behind the Soft Actor-Critic algorithm for continuous and discrete action space, as well as their resp...
Diving into the autoregressive models for unsupervised reinforcement learning with a focus on the MADE algorithm. Aims a exposing the theory in an intuitive ...
Going over the theory and implementation of the DDPG method. Also provides some benchmark on a handful of environments, and the effect of some specific imple...
Investigating the impact of various exploration schemes on the DQN performance and learning efficiency.
Build log and notes for the clear cased Charybdis Ergonomic Mechanical Keyboard with integrated Trackball
Motivation, building logs and notes of the ErgoDash rev1.2 Split Mechanical Keyboard
An automation-friendly, custom OpenAI Gym environment wrapper of the Torcs Racing Car Simulator for Reinforcement Learning research.
