Final project report : GSoC 2019 #472
Description
This issue lists the work done throughout the GSoC 2019 and highlights the work that can be done in the future.
Work done:
- Added notebooks for the Classical searches and the Non-classical searches explaining the various search methods given in the book and demonstrating how the code can be extended to solve various search-related problems. These notebooks can be viewed here and here.
- Added a notebook for the Adversarial Search illustrating the use cases of the adversarial package of the code repository.
- Added a notebook for Constraint Satisfaction Problems.
- Explored various deep learning libraries for java (ND4J and DL4J) and wrote a tutorial for the same in the form of a notebook.
- Wrote a notebook for the Deep learning chapter to demonstrate the techniques mentioned in the book. Used DL4J and Tensorflow for java for the same.
- Implemented the First Order Inductive Learner and tested the same on the family tree given the textbook. The implementation can be viewed here.
- Implemented the Monte Carlo Tree Search algorithm and tested it on the game of Tic-Tac-Toe. The implementation can be viewed here.
- Added testing suite for testing the algorithms of the adversarial package.
- Wrote a notebook for the Reinforcement Learning chapter illustrating the various uses of reinforcement package of the code repository.
I've included detailed progress of my GSoC and the overall experience with aimacode in my blog posts. Please give them a read.
Link to my PRs till date: here
Future possibilities:
- There's still a huge scope of contribution in the notebook section. A general approach to any notebook is discussed here.
- Angelic Search is yet to be implemented in the AIMA3e branch.
- There are plenty of APIs which are yet to be implemented in the AIMA4e branch.
- Besides this, @samagra14 has discussed the possible additions to the project in the issue Miles to go : GSoC 2018 #438.