FAQ

1. What is this project?

This is project is an endeavour between wakemaster39 and LPiner in an attempt to learn more about machine learning and give it a particle application to learn on. Both of us are professional developers in different fields and just stretching out legs in learning about a new field. The code, practices and ideas are no where near perfect but we welcome healthy discussions and suggestions so we can further our understanding and those around us.

2. What language is this written in?

We used the most appropriate language in this case of python, the language of sneks everywhere. On top of that we are using keras with the tensorflow back end. There are a hand full of other random libraries in there as well but those two are the most critical.

3. When will it do something?

That is a good question, and one we can't honestly answer. We are continuously making upgrades and improvements to both our knowledge and the snakes infrastructure. This at times requires model learning resets which removes all previous progress. This is for the better though as we hope that by making these improvements the final product will be much improved.

4. What type of learning is this?

Well depending on the time you check in, that could of changed. But we are using Reinforcement Learning (RL) of some nature. I suggest you take a look at the change log (add link) to see what the current status of it is.

5. Is the source code going to be released?

Short answer is yes, eventually. The program is continuously evolving and we don't want to just throw out a bad example for people to learn from. We ourselves are still learning too. So we will release it slowly over time as we become more comfortable that it is working and working well. We do not plan on releasing the connection to get the data, this should be relatively trivial for most people just it keeps script kiddies honest and us not releasing a functioning bot to the world.

__6. What backend does it use? Are you using OpenAI/Tensorflow?

We have currently written the bot using Keras and the Tensorflow backend. This was an arbitrary decision on our end with no particular reason for it. We likely could of used anything in the end and found a way to make it work. On the next project we will likely try something different so we can learn the pros and cons of each.