Finally someone who understands it. I can`t believe how people lack ground knowledge in plotting functions, but are already studying Maschine Learning. Thank you very much.
This is such a great explanation. So much important information is packed into this one video. Thank you very much for saving my time! keep up the great work. The programming community needs more people like you
Thank you so much, this really helped me grasp the concept. I always wondered how big inputs (or many inputs) result in 0-1 ranges when leaving the node, since they are all added and multiplied beforehand. I knew the activation function takes care of it, but had no idea how exactly
This was great. Had heaps of trouble finding something that explains why the function is written the way it is. Thanks. Any ideas why we specifically use Euler's Constant vs any other. Or is that just because it's commonly utilised as a "Natural" exponential function?
This video was an incredible help, ty for making it. Does anyone know why there's a limit for x=0? I thought limits were only for when x neared impossible values, but x=0 doesn't seem to break anything? My knowledge of calculus is still pretty low. Thanks.
We take limit at x=0 just to find the threshhold value of sigmoid function so that we can draw the border point or threshhold point on the basis of that value we can predict either our independent variabe lies in 1 category(in this case Cat) or in second category(Dog)). hope you understand.
Unfortunately, mathematicians cannot explain mathematics to non-mathematicians 😅. If i would explain the Sigmoid function clearly, I would start with the problem itself and what should be solved, before I start with any numbers and formulas.
so in other words, the sigmoid function should be used only WHILE training, then testing should just use a comparison operator to see if the value is more or less than 0.5?
It's so funny when you realize that videos with superficial concepts of AI have a lot of views, and videos with specifical topics just have a few views
