This repository contains my whole journey through Artificial Intelligence during my study.
Sudoku Solver
• An AI agent that solves the given sudoku in an efficient and quick way.
• The project is implemented using a nice GUI.
• Various stratergies that improve the speed of solution an reduce the expansion of state space have been implemented.
Isolation Board Game Playing AI
• An AI bot that plays the game of Isolation by itself.
• The large state space in the game has been pruned using stratergies like - Alpha Beta Pruning etc.
• The agent is utilizing the Minimax algorithm with various custom Heuristics to reduce the Horizon effect.
Pacman AI bot
• An AI bot that plays the Pacman game all by itself.
• It finds the path with the best score possible along with enemies/ghosts/food-pellets etc.
• Various search strategies along with heuristics were implemented such as Adverserial, A* etc.
Simulated Annealing
Uploading soon
Solving AIR CARGO PROBLEM via PDDL
• A Planning Problem - given different planes and cargo destinations and airports , find the optimal path for the travelling and transportation of cargos.
• Solved using a Planning Graph with PDDL.
• The Planning Graph helps in finding very good heuristics and reduces the exponential size of state space search tree.
• The project, describes a pipeline that can be used within a web or mobile app to process real-world, user-supplied images.
• Given an image of a dog, the algorithm will identify an estimate of the canine’s breed. If supplied an image of a human, the code will identify the resembling dog breed.
• Along with exploring state-of-the-art CNN models for classification, it focusses on important design decisions about the user experience for the app.
• A deep neural network that functions as part of an end-to-end automatic speech recognition (ASR) pipeline!
• Used the LibriSpeech dataset to train and evaluate models.
• The algorithm will first convert any raw audio to feature representations that are commonly used for ASR.
• Then moving on to building neural networks that can map these audio features to transcribed text.
• Final implementation was a Bidirectional RNN's with no batch normalization and GRU/LSTM units in different layers and with Time Distributed Dense Layers.