Yeah. Even I am of the opinion. Schmidhuber has been one of the geniuses with LSTMs etc coming out from his Lab. It's hard to imagine he would needlessly interrupt.
Ex and Ez are The Cost Discriminators trying to Maximize on the outcome! Helloo!? Make no mistake: the discriminators are simply the negative of what the generators are trying to put out there. And, don't forget the generator minimizes the log probability of the discriminator being correct.
When answering the question of 1:25:44 he told that the guy he mentored at Google was working on something related. Did someone get the name of his former intern? I am interested in his project!
hello sir, i'm doing a study having 19 predictors. 2 are continuous variable and 17 are categorical variable but my outcome is categorical variable such as 1 and 0. but now i have great interest in this newly model such as GAN.. i'm asking is it possible to use this kind of model in my study? thanks
I have only a slight understanding of calculus. What exactly is d(x) and g(x). Derivatives? Derivatives of what? What is the way into understanding these long equations?
![The moment we stopped understanding AI [AlexNet]](http://i.ytimg.com/vi/UZDiGooFs54/mqdefault.jpg)
If anyone wants to listen to the confrontation between Goodfellow and Schmidhuber, here it is: 1:02:58
Goodfellow handled that like a ninja.
thanks. i was searchiung for thgat
Yeah. Even I am of the opinion. Schmidhuber has been one of the geniuses with LSTMs etc coming out from his Lab. It's hard to imagine he would needlessly interrupt.
thank you a good fellow
Thanks so much for sharing !! Its amazing to hear from Ian himself...
Did anyone notice that the person who introduced Ian was Aaron Courville..!
Dude. I love nerds.
Any idea, at 33:28, What do Ex and Ez represent in the objective function J of the discriminator ?
Ex and Ez are The Cost Discriminators trying to Maximize on the outcome! Helloo!? Make no mistake: the discriminators are simply the negative of what the generators are trying to put out there. And, don't forget the generator minimizes the log probability of the discriminator being correct.
What an exciting time.
When answering the question of 1:25:44 he told that the guy he mentored at Google was working on something related. Did someone get the name of his former intern? I am interested in his project!
It sounded like Xi Chen - it's possibly this guy: www.linkedin.com/in/peter-xi-chen-25178356/. Here's his site: peterchen.io/.
Pramod Beligere Nice! Thank you so much, Pramod!
It's Tianqi Chen (homes.cs.washington.edu/~tqchen/)
hello sir, i'm doing a study having 19 predictors. 2 are continuous variable and 17 are categorical variable but my outcome is categorical variable such as 1 and 0. but now i have great interest in this newly model such as GAN.. i'm asking is it possible to use this kind of model in my study? thanks
thank you for this!!!!!
Thanks for sharing
I have only a slight understanding of calculus. What exactly is d(x) and g(x).
Derivatives? Derivatives of what? What is the way into understanding these long equations?
D(x) is the output of the Discriminator and G(x) is the output of the Generator.
How GAN works starts at ua-cam.com/video/AJVyzd0rqdc/v-deo.html
Slightly better video quality: ua-cam.com/video/YpdP_0-IEOw/v-deo.html
It seems that they are both 720p.