- Autoencoders: In this notebook, we train two Autoencoders: 1) to Reconstruct Fashion MNIST images 2) to detect Anomalies in the ECG500 Dataset.
- Cartpole Environment Steps vs Environments Reward Evaluation: In this notebook, we simulate the Cartpole Environment and compare the total reward accumulated by an Agent per Step and Episode.
- Six Degrees of Kevin Bacon: In this notebook, we find the distance between two nodes in a graph using Breadth First Search and Depth First Search.
- Gaming Agent & Negamax: In this notebook, we implement a Tic-Tac-Toe game with one human player. We train an AI agent using Negamax, and then we solve the game using Depth First Search & Iterative Deepening and compare the solutions.
- Multi Armed Bandits: In this notebook, we create a Custom Environment that inherits the Multi Armed Bandit Environment and initialize its Observation Space, Action Space, Reward Structure, and Policy.
- Epsilon Greedy & Upper Confidence Bound: In this notebook, we optimize the Ad Placement by applying the Epsilon Greedy and Upper Confidence Bound approaches on the Click-Through-Rate data.
- LinUCB, Thompson Sampling, & Neural Epsilon Greedy: In this notebook, we generate Movie Recommendations by applying the LinUCB, Thompson Sampling, & Neural Epsilon Greedy approaches on the MovieLens dataset and analyzing the ratings given by users to different movies.
- Markov Decision Process & Dynamic Programming: In this notebook, we compare the policies generated by the Policy Iteration and Value Iteration on the Frozen Lake Environment.
- Monte Carlo Methods: In this notebook, we compare the performances of the Monte Carlo Exploring Starts and Monte Carlo Epsilon Soft policies on the Cliff Walking Environment.
- Temporal Difference: SARSA, Expected SARSA, and Q-Learning: In this notebook, we compare the performance of SARSA, Expected SARSA, and QLearning on the Taxi Environment.
khusheekapoor/ArtificialIntelligenceProjects
Mini Projects on Artificial Intelligence including Reinforcement Learning, Gaming Agents, Multi Armed Bandits, Markov Decision Process, Dynamic Programming, Monte Carlo Methods
Jupyter Notebook