sherjilozair

9 ms·

31.
by sherjilozair 10y ago
The real question is, where do you get this training data from?
32.
by sherjilozair 10y ago
Can Vulcan be used for Deep Learning?
33.
by sherjilozair 10y ago
The TF program does learn to sovle FizzBuzz. The above code was used to generate data with which to train the network.
34.
by sherjilozair 10y ago
It has been popular belief that although deep learning has been successful in a lot of domains, its biggest shortcoming is high sample complexity, i.e. even for simple problems you need tens of thousands of samples. Thus, deep learning was
35.
by sherjilozair 11y ago
Part of the problem in writing a deep learning book, is that very little that warrants being in a book, is actually known. The mainstream deep learning academic community welcomes theoretical work but articles on new techniques which beat S
36.
by sherjilozair 11y ago
The FSF does not have a monopoly over the word "free". For a lot of people, "free" means as in "free beer", which this book is, in the online format.
37.
Baidu Research and Peel Collaborate on Voice-Enabled Smart Home Products(finance.yahoo.com)
3 points by sherjilozair 11y ago|0 comments
38.
by sherjilozair 11y ago
IMO, anything that be done with an HMM can now be done with an RNN. The only advantage that an HMM might have is that training it might be faster using cheaper computational resources. But if you have the $$$ to get yourself a GPU or two, t
39.
Self-Driving Cars Won’t Work Until We Change Our Roads – And Attitudes(wired.com)
13 points by sherjilozair 11y ago|6 comments
40.
Recognizing Both English and Mandarin with a Single Algorithm(svail.github.io)
16 points by sherjilozair 11y ago|1 comments
41.
Ride-Along in One of China’s First Self-Driving Cars(technologyreview.com)
4 points by sherjilozair 11y ago|0 comments
42.
by sherjilozair 11y ago
Might be worth noting that it is very likely that Google, Facebook and Baidu did not even participate in this ImageNet. It is widely believed in the vision community that this dataset is at the end of its run, i.e. any subsequent improvemen
43.
by sherjilozair 11y ago
Well, what I really wanted to post was his quote on "Fear of AI", not really the whole video, which is https://youtu.be/qP9TOX8T-kI?t=3836
44.
Andrew Ng on What's Next in Deep Learning [video](youtube.com)
76 points by sherjilozair 11y ago|11 comments
45.
by sherjilozair 11y ago
I only get 5. I'm happy. And they are probably people who 'liked' the page much before he announced candidacy.
46.
Deep Speech 2: End-To-End Speech Recognition in English and Mandarin(arxiv.org)
105 points by sherjilozair 11y ago|19 comments
47.
TensorFlow Benchmarks(github.com)
98 points by sherjilozair 11y ago|55 comments
48.
by sherjilozair 11y ago
> Allegedly this practice goes on in America, without the consumer's knowledge. source?
49.
by sherjilozair 11y ago
Why do you need a CUDA account? You don't need one to install CUDA.
50.
by sherjilozair 11y ago
This right here made me stop as well. I used to keep a box in plain view. That facilitated the thought of "It's right here, I can smoke any time I want. I'll just not smoke right now, maybe later if I still feel like it, the
51.
by sherjilozair 11y ago
You can't be further from the truth. A vast magntitude of bugs are due to the system running as if nothing has happened operating on garbage data, and giving you garbage results. If your system is robust to garbage data, say neural net
52.
Baidu explains how it’s mastering Mandarin with deep learning(medium.com)
11 points by sherjilozair 11y ago|0 comments
53.
by sherjilozair 11y ago
Sorry, but the problem is how to use a pen to draw machine readable shapes. Machines find reading QR codes and OP's triangles easier than reading letters.
54.
by sherjilozair 11y ago
Not necessarily. You can overload @ to do some other kind of object combination, which fits your paradigm better. You're not tied to using it for matrix multiplication.
55.
by sherjilozair 11y ago
This is Google we're talking about. I think they will spend time to generalize this situation as much as possible.
56.
by sherjilozair 11y ago
This is why it isn't 'by default'.
57.
by sherjilozair 12y ago
def a(b=[]): b.append(1) print b Now if you do, a() a() you might expect to see [1] [1] but you will actually see [1] [1, 1] The list is not created at function call time, but at function defin
58.
by sherjilozair 12y ago
Slightly off-topic, but is anyone aware of an similarly capable library for hand-written text recognition, i.e. ICR?
59.
by sherjilozair 12y ago
Why not? DNN-HMM and GMM-HMM won;t have done any better even if trained for 2300 hours.
60.
Baidu claims deep learning breakthrough with Deep Speech(gigaom.com)
7 points by sherjilozair 12y ago|1 comments

More ›