AI for Terminal

[xZeko]

Its been a long time since I have checked Terminal and last time I did it was quite difficult to get started with ML in Terminal because of the performance issues.

Interested to see what people have been upto and to see if ML or more specifically NN have been used effectively in this challenging game?

[IWannaWin]

I’m not aware of too many NN’s that were effectively used in the last couple of seasons. I know @Ryan_Draves made a decent one called ML_BOOTSTRAP, he made a thread about it if you are interested:

Mixed Results: DQN Experimentation

In preparation for the upcoming competition and the end of the semester, I’m probably shelving this project for the foreseeable future. But, I wanted to share what I’ve been working on for the past two months. Want to try this out too? Below is a short list of links of stuff I remember being helpful. It shouldn’t assume too much prior knowledge. I think starting with supervised learning helped a lot though.

Links

Game Engine <a href="https://terminal.c1games.com/rules#GameplayOverview&…

Further I know @paprikadobi made one based on my old algos xD, diskussed a bit at the end of this thread:

Season 2 Competition Prizes

Interesting, are you making a machine learning algo or something? Well I mostly use the demux, then my algo also has the ability to use a maze and a pingcannon. Theres one emergency very dynamic mode, mostly activates when opponent is a first row emp line like newton. Theres also one that rarely is chosen by my algo, but I think it fires for example against cakewalk. So that makes 5

Machine learning in general was talked about here:

The Ultimate Copy-Cat - ML Implementations

So I have personally begun the dive into Machine Learning (ML) with Terminal. It is my opinion that this strategy has taken a back seat since these discussions: Machine Learning How does the machine learning portion come into play? Machine learning libraries It is my goal with this topic to revamp some of the interest and discussion around ML. Personally, I hope to make ML a valid and viable option that can be successful in the competition. I’ll begin by discussing a little bit about what I…

But for now I know of no one who effectively got into the top 10 or something with a NN algo.
This might change this season though, since machine learning libraries were installed on the server, making it way easier to implement:

Python requested Libraries

Hey everyone could you supply the python libraries you would like to be included in our workers for your python algos? I can’t guarantee when we will get to including them but would be good to get a list. Here are some I already heard about and recall: numpy keras pytorch Scipy TensorFlow Pandas

I’m quite sure some people are at least part time working on some machine learning algos (including my self btw), so I expect to see some on the leaderboard soon (if they aren’t there already, I’m currently not that active and haven’t checked)

[xZeko]

Wow this is impressive really nice to see Terminal team add the libraries required, this is a good step
Thanks for the great overview @IWannaWin really appreciate it.

And if anyone is working on ML I think it would be awesome if you can share results / methods / status anything really just to give people an idea of the status of ML in Terminal.

[Isaac]

I’d love to work on a ML algo in full, particularly since life will be a lot easier with the new libraries being added. I tried some stuff before and most of the work was me just getting things set up properly. I will definitely share my progress once I get into it. Unfortunately (or fortunately?), my life has recently become very busy so I probably won’t be able to devote much time to this for a while :(.