CSE 571 Artificial Intelligence

Collection of Artificial Intelligence Algorithms implemented on various problems

implement basic and contextual MAB algorithms for recommendation system

Python implementation of UCB, EXP3 and Epsilon greedy algorithms

Offline evaluation of multi-armed bandit algorithms

Implementations of basic concepts dealt under the Reinforcement Learning umbrella. This project is collection of assignments in CS747: Foundations of Intelligent and Learning Agents (Autumn 2017) at IIT Bombay

Machine Learning based Load Balancing with RYU OpenFlow Controller

Multi-Agent Deep Recurrent Q-Learning with Bayesian epsilon-greedy on AirSim simulator

Repository Containing Comparison of two methods for dealing with Exploration-Exploitation dilemma for MultiArmed Bandits

The goal of this project is to build an RL-based algorithm that can help cab drivers maximize their profits by improving their decision-making process on the field. Taking long-term profit as the goal, a method is proposed based on reinforcement learning to optimize taxi driving strategies for profit maximization. This optimization problem is fo…

Implementation of greedy, E-greedy and Upper Confidence Bound (UCB) algorithm on the Multi-Armed-Bandit problem.

Greed is Good: Exploration and Exploitation Trade-offs in Bayesian Optimisation

Deep Recurrent Q-Network with different exploration strategies for self-driving cars (using AirSim)

A multi-armed bandit (MAB) simulation library in Python

Epsilon-Greedy Q-Learning in a Multi-agent Environment

Public repository for a paper in UAI 2019 describing adaptive epsilon-greedy exploration using Bayesian ensembles for deep reinforcement learning.

A multi agent reinforcement learning environment where two agents controlled by DRQNs play a custom version of the pursuit-evasion game.

Implementation of the Q-learning and SARSA algorithms to solve the CartPole-v1 environment. [Advance Machine Learning project - UniGe]

See a program learn the best actions in a grid-world to get to the target cell, and even run through the grid in real-time! This is a Q-Learning implementation for 2-D grid world using both epsilon-greedy and Boltzmann exploration policies.