This is my (Andrés del Campo Novales) prize winning submission for GoodAI's General A.I. Challenge (Round 1: Gradual Learning).
You can read more about it in General A.I. Challenge - Gradual Learning Results You may also check the specifications or the whitepaper submitted to the challenge outlining the instructions for compiling and running the agent, as well as its high-level design.
Note that this project uses NetMQ to communicate to the AI environment.
The code in the Source folder is uploaded AS IT WAS delivered to the challenge with the exception of adding license and cleaning up the solution of extra files. It was developed in C# under Windows. Note that due to time constraints and due to the prototyping nature of the challenge, the code is not necessarily as clean as I would have liked it to be.
You can also find an agent already compiled in Learner.zip, which can be tested using the Environment Instructions against the training task set included -or any evaluation set created separately.