If links are broken for you, then turn of your adblocker, because he is measuring clicks with Google analytics and he's JS is broken thanks to the missing ga function. #issue reported
thanks for reporting. It should be fixed now
A Guide to Deep Learning
21–30 of 35 posts
Re: A Guide to Deep Learning
#22Earlier quoted context omitted.
I will mention that you can get surprisingly far with just a little bit of labeled training data now. Transfer learning is a pretty powerful technique.
You still need to train a model on what you're transferring from . It's not like you can manufacture trained models out of thin air.
I think the idea is that, for example, you start with some pre-trained visual object-identification network, remove the final logistic-regression layer, and replace with a layer which you train your problem on.
Re: A Guide to Deep Learning
#23The primary thing missing from all of these guides is that you need to have two things for ML: 1. A purpose for utilizing it 2. A data set to train/act on Without that, all you get are a bunch of shovels and picks, but no idea of what kind of wood/bricks you need or a plan for the house.
Thanks for the feedback, Andrew!
This guide is designed mostly for those who already know what problems they want to solve, but don't know how to start or where to look for high quality and up-to-date educational resources. Also this is designed for those who want to do research in this area and just want to develop better shovels (that was me 2 years ago).
Regarding datasets, each course we suggest in the guide has its own way of dealing with datasets. Most of them teach how to work with MNIST, which is pretty good for many purposes.
I agree that another guide on datasets could be useful for some people.
Re: A Guide to Deep Learning
#24I'll bite. We see these being posted every week. Why?
Many people (especially students from universities) keep asking me and my coworkers at YerevaNN about good educational resources. They have different levels of math background, some want to study theory and watch visualizations, others want to play with the code before reading formulas..
At some point we understood it's better to spend some time and build a guide that will cover most of these questions (and, as always, we spent a lot more time on this than we expected)
Re: A Guide to Deep Learning
#25I'll bite. We see these being posted every week. Why?
Many people (especially students from universities) keep asking me and my coworkers at YerevaNN about good educational resources. They have different levels of math background, some want to study theory and watch visualizations, others want to play with the code before reading formulas.. At some point we understood it's better to spend some time and build a guide that will cover most of these questions (and, as alway…
I think OP is asking, why spend time writing another guide when there's plenty that suffice, including full books and papers on the topic?
Re: A Guide to Deep Learning
#26Earlier quoted context omitted.
Everyones into Graphics programming, but what would I actually do with it? The uses are myriad, you can build Star Wars Movies, you can build computer games. You can do CAD and 3D modelling and printing, you can visualize fluid simulations But each is an entire field. The same with Deep Learning - You can do image classification for medical diagnosis. Self Driving Cars. Realtime Translation. OCR. You can model chemic…
One of my partner's tasks as a legal assistant is to go through mountains of OCRed PDFs and classify them and extract pieces of data so that lawyers and paralegals can go through them more easily. Do you imagine deep learning would be an appropriate means of automating that, or is it overkill?
I have built deep learning systems for similar problem domains. I am happy to chat to you about your example. Email me if you get a chance.
Re: A Guide to Deep Learning
#27The primary thing missing from all of these guides is that you need to have two things for ML: 1. A purpose for utilizing it 2. A data set to train/act on Without that, all you get are a bunch of shovels and picks, but no idea of what kind of wood/bricks you need or a plan for the house.
Thanks for the feedback, Andrew! This guide is designed mostly for those who already know what problems they want to solve, but don't know how to start or where to look for high quality and up-to-date educational resources. Also this is designed for those who want to do research in this area and just want to develop better shovels (that was me 2 years ago). Regarding datasets, each course we suggest in the guide has…
Yes, thank you for the resources and my comment is not an indictment on your page as I think it's well organized.
Moreso I'm commenting broadly on the recurrence of similar guides and how they are generally not accessible to people with no exposure to ML.
All the best.
Re: A Guide to Deep Learning
#28Re: A Guide to Deep Learning
#29Everyone's into deep learning, but what would I actually do with it? With some other field, like computer graphics, one can fairly quickly get a 3D cube spinning on their screen and know it has some relation to the special effects in the Star Wars movie they just saw. No one makes it obvious what the hobbyist can expect to do with deep learning or how it relates to the broader world.
Everyones into Graphics programming, but what would I actually do with it? The uses are myriad, you can build Star Wars Movies, you can build computer games. You can do CAD and 3D modelling and printing, you can visualize fluid simulations But each is an entire field. The same with Deep Learning - You can do image classification for medical diagnosis. Self Driving Cars. Realtime Translation. OCR. You can model chemic…
You can ride your bike with no handlebars. ;)
Re: A Guide to Deep Learning
#30Earlier quoted context omitted.
Many people (especially students from universities) keep asking me and my coworkers at YerevaNN about good educational resources. They have different levels of math background, some want to study theory and watch visualizations, others want to play with the code before reading formulas.. At some point we understood it's better to spend some time and build a guide that will cover most of these questions (and, as alway…
I think OP is asking, why spend time writing another guide when there's plenty that suffice, including full books and papers on the topic?
I was looking for something that would answer common questions for some time, but couldn't find anything. Pointing students to research papers unfortunately doesn't work in most of the cases. Deep learning book is too hard for some and they want to see a comparison of different sources.