Looking for some experienced with neural networks

Wish Lin

Veteran
Joined
Jun 18, 2019
Messages
673
Location
Taipei, Taiwan
WCA ID
2018LINW02
I have a project here on SS which is about a possibly next generation smart cube design. https://www.speedsolving.com/thread...ach-for-smart-cubes.75761/page-2#post-1350603
And I have now came up with a good hardware prototype but lacks the software expertise.

In my plan, this neural network
  • Will takes 8(or more) analog values THAT KEEP CHANGES ranging from 0~1023 as input and output cube moves(ie. R’).
  • Will have to be written in a language that can take inputs from a computers COM ports(from an Arduino)
  • Needs to work quite efficiently because it will take at least 400 inputs a second, or even more.
So, if you are familiar with neural networks, please PM me and I will share the rest of the details with you.
 
Last edited:
I have read the other thread but I really don't understand why you need neural networks with this
 
I have read the other thread but I really don't understand why you need neural networks with this
My new design uses 8 hall sensors to detect magnets placed in the 2x2. Although my design can indeed determine moves with some simple time delaying tricks, I don’t think it will work at all under high TPS and especially big corner cutting.

So, it’s better for me to have a neural network to learn the magnetic field changing patterns under that condition and determine moves.

But of course, since I have no neural networks experiences and only know the general concept on it, I’m not sure if this will work or not.
 
Hmmm, it still doesn't sound like the right tool to solve your issues but I'm no ML expert, so maybe I'm wrong
 
Back
Top