Project Title: Machine Learning as a Service for Multi-agent systems developed in NetsBlox
Project Owner: Tamas Kecskes
Related Class: DAI – CS6366
Class instructor: Dr. Julie A. Adams
Student’s advisor: Ákos Lédeczi
Goal: The primary goal of the project is to provide access to machine learning capabilities to beginning programmers by creating a service-based extension for the NetsBlox visual programming language. This extension should also be generic enough solution that it can be leveraged in a variety of contexts. The main focus will be on the development of “AI as a service” for use in a blocks-based programming environment. The service should be flexible enough to play a variety of games effectively. (detailed description and schedule available here)
Specification: The first step into understanding our game and how the different agents can behave we created their state machine representations (bumperGame and bumperPlayer). As a further step, we described the different messages that will be used during the client-server and server-learner communications.
Regarding our learner, we made our initial considerations and picked our implementation strategy (that might needs to be changed as a result of our initial tests).
DemonstrationWebSite: Here you can always check out how the service looks like in real life (look for the example game ‘BumperGame’ – currently it is just a new deployment of the Netsblox system).
ProjectReports: