Hey chummers,
Is there any way to explore mathematics in a completely different way or perspective when compared to traditional/mainstream math..
Do you have any idea..
Feel free to rant about it..
What do you mean?
Kind of on this subject. I think neural networks and back propagation are retarded. The math behind them is not, but classifying them like that is dumb. It's better to think of a neural network as a linear function. The function takes linear variables (you can only use each of them only once) in and outputs other variables. You can think of training a neural network as having a computer trying to write a function that tries to minimize the result of some cost function. Back propagation is simply auto differentiation.
The math you learn in school is always "here's how you do a thing". I'm currently taking a course on statistics, and this is all it is. You build complex mathematical frameworks, but only to serve the end goal of munging data. As such, the course tries it's best to avoid having to teach people what all the math means. The prof put the definition of the students t distribution up briefly, but there was no discussion of where is came from, or what it means. What annoys me especially is how much of this could be done ten times easier with a monte carlo simulation, probably with fewer assumptions and higher accuracy. We're learning how to execute algorithms we don't understand, despite the last half century having been spent perfecting algorithm executing machines.
I'm not telling that mathematics is useless but the way we're introduced to it is not helping..
So i asked for a new perspective
>>1186 Exactly the problem im facing, thats why i never paid much attention in lectures
>>1187 Yeah but in a much disciplined way..thats why i said mathematics
Hey chummers,
Is there any way to explore mathematics in a completely different way or perspective when compared to traditional/mainstream math..
Do you have any idea..
Feel free to rant about it..