>.<
![\frac{\partial{G}}{\partial{w_{ij}}} = -\frac{1}{R}[p_{ij}^{+}-p_{ij}^{-}]](https://lh3.googleusercontent.com/blogger_img_proxy/AEn0k_uQiovfCkncsf0nJlO_wOUanA8jWrN5NB8vI7KlLMFNj3m52LBDmYvHsRpP26tdpP3vYqrk4cfHaS-XJlHLQQdFowaSq5dC1eV3LMDsMi0ns7Dstfx2CLlMCxIjzVQR9oVzAsj3p6YKMFAuvbRP=s0-d)
![\frac{\partial{G}}{\partial{\theta_{i}}} = -\frac{1}{R}[p_{i}^{+}-p_{i}^{-}]](https://lh3.googleusercontent.com/blogger_img_proxy/AEn0k_uMyAGjZVl-jz15hh4P2bbGcgjGaZeEsnDkGkzd7uXWHpJQy4Obe_S4bY10GTjpjit1YvdX_ES7yJLNBCrCCUetwHbXUKNQdJInM4dXWL1pGyARY7z9tLV867oZOsbetubMQR3w1fry7vVQCywH4Q=s0-d)
I figure that I'd have to explode these out to larger decompositions, but most of these symbols I don't even know what they mean.
Thursday, October 14, 2010
Smartness
So I've made some real progress on mah C# ANN WPF app (say that three times fast)... I found this cool thing at Microsoft's DevLab called Rx, which is a Reactive library that replaces Events in .NET with functional-type methods. Combine that with PLINQ, and I'll have a very familiar environment to work with. The only thing that I am pondering now is how to get the calculus done for it. Hinton uses a thermo-dynamics model for training the A.I., but I'll be buggered if I can actually codify it into a method/function. Guess I'll need to hit half-price books for a dummy's guide to calc.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment