AI - Text recognition and a Planner

During two different projects some text recognition algorithms was implemented in the first one and two planners was implemented in the second one. The report for the first project is available here.

Technical Skills

Java, Eclipse, Algorithms

Description

During seven weeks (half time) in the course Artificial Intelligence at Chalmers University of Technology me together with two other students implemented, tested and wrote reports about two different projects. The first one included implementing a few text recognition algorithms; Naive Bayes, Perceptron and Averaged Perceptron. We also tested a fourth algorithms that was already implemented; K-nearest neighbours. We tested the algorithms on six different review categories to see how well they would perform. Our tests included k-fold cross validation, out of domain tests, Mc Nemar test and other statistical analysis. We did almost every coding together, either in pairs or all three of us which means I have been involved in doing almost everything.

In the second project our task was to implement a planner that should be able to move blocks according to certain requests in a virtual Blocks World. However, it should be able to handle more advanced tasks then possible in the Blocks world, rather more like the SHRDLU. We did a simple planner with our own implementation and we also made a planner with the inspiration from the Partial Order Planning algorithm. However, it was not fully realised. Also in this project I have been involved in almost everything since we worked everyone together.

AI code