koobas.hobune.stream
Question is off-topic and not specific to Unity
How can I train a simple supervised learning model and use it in my game? (Classification)
Hey all,
I need to use machine learning to allow my game to classify a large set of data as one thing or another. How can I do this?
I've looked at ML-Agents but seems like it's overkill, I don't need any decisions made in realtime, I just want a model that can tell me that...
[set of floats] is "FLOWER",
[new set of floats] is "CUP" (for example)
I would train it ahead of time and then no new learning would happen after that.
p.s I'm new to Machine Learning so I may be thinking about something in the wrong way here...
I have doubts that a simple ANN will do the job you require. For instance, I tried to make a cat versus dog classifier, and it was right only 50% of the time -- this is the same as if you randomly guess, which is not good at all. Best of luck to you. $$anonymous$$y C++ code is at https://github.com/sjhalayka/ann_xor -- it is 95% compatible with C#, you just need to switch from a vector to a List.
