Dear MLPack-Team, though I know that the official application deadline for Google Summer of Code applications was in early April, would I still like to add some late information to my proposal as I have done some significant work towards my proposed project.
Unfortunately do I not see some sort of GSoC ID to identify myself with, so I hope my Name (Paul Pauls) and my e-mail address will suffice ([email protected]), which I also stated in the application. The title of my application proposal was "Implementing ‘NeuroEvolution of Augmented Topologies’(NEAT) in MLPack and integrating it into the MLPack ecosystem." By now have I done some significant research into Neuroevolution in general and Neuroevolution of Augmented Topologies in specific. I am currently writing a seminar paper about it in order to grasp the topic as best as I can and therefore hopefully deliver the best GSoC project as I can. Though the paper is still in the making and changes daily, have I created a github repository for it [1] so you can convince yourself about my familiarity with the subject. I also created a repository that trains an agent to play the SNES game 'Super Mario World' utilizing the NEAT algorithm when training the neural net [2]. Though it is still in heavy development it can already create an agent that successfully beats the first level (Yoshi Island 1). It utilizes the NEAT-Python library and judges its input by actually analyzing the Videogame Screen via OpenCV, unlike the popular MarIO [3], which gets its input by reading the right points in the ROM memory map. I hope - even though it is quite late in the process - that this email finds its way to the people judging my GSoC application. Either way would I like to thank you for your time and consideration. With my best regards, Paul Pauls [1] https://github.com/PaulPauls/Neuroevolution_of_Augmenting_Topologies_Paper [2] https://github.com/PaulPauls/SuperMarioWorld-NEAT-Agent [3] https://www.youtube.com/watch?v=qv6UVOQ0F44 _______________________________________________ mlpack mailing list [email protected] http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack
