model-free

Curvature Corrected Moving Average: An accurate and model-free path smoothing algorithm.

ExORL: Exploratory Data for Offline Reinforcement Learning

SUNRISE: A Simple Unified Framework for Ensemble Learning in Deep Reinforcement Learning

Implementation of Soft Actor Critic and some of its improvements in Pytorch

Customisable Unified Physical Simulations (CUPS) for Reinforcement Learning. Experiments run on the ai2thor environment (http://ai2thor.allenai.org/) e.g. using A3C, RainbowDQN and A3C_GA (Gated Attention multi-modal fusion) for Task-Oriented Language Grounding (tasks specified by natural language instructions) e.g. "Pick up the Cup or else"

Code release for SceneReplica paper.

Code for IEEE MLSP 2021 paper titled "Model-Free Learning of Optimal Deterministic Resource Allocations in Wireless Systems via Action-Space Exploration"

A Pytorch implementation of The Visual Centrifuge: Model-Free Layered Video Representations.

Deep Reinforcement Learning implementation in Keras of an AI controlling the popular Flappy Bird videogame, using Asynchronous Advantage Actor Critic (A3C)

Network inference via Event Space Linearization (ESL)

Project for the course "Foundations of Reinforcement Learning" 2021 at ETH Zurich

Sample Policy Gradient

A Liquid RL framework for small scale architecture research

A certifying, model-free, minimal-intervention runtime shield for LLM agents: it repairs an unsafe tool-call trajectory with the least-cost enactable edit instead of blocking, and certifies every intervention for re-verification without the model.

A compact PyTorch implementation of the REINFORCE algorithm enhanced with a learned baseline (value network).This repository trains a policy network on OpenAI Gym's CartPole-v1 environment using Monte Carlo returns and a value function baseline to reduce variance.