Hacker News new | comments | show | ask | jobs | submitlogin
Ask HN: Best introductory video courses on ML and Deep Learning?
323 points by rayalez 9 months ago | hide | past | web | 40 comments | favorite






fast.ai Making neural nets uncool again. fast.ai is dedicated to making the power of deep learning accessible to all.[0]

Udacity. Intro to Machine Learning: Pattern Recognition for Fun and Profit[1]

MIT 6.S099: Artificial General Intelligence[2]

Lecture Series on Artificial Intelligence by Prof. P. Dasgupta, Department of Computer Science & Engineering, Indian Institute of Technology, Kharagpur[3]

DeepMind. Reinforcement Learning Course by David Silver[4]

Stanford Winter Quarter 2016 class: CS231n: Convolutional Neural Networks for Visual Recognition.[5]

[0] http://www.fast.ai/

[1] https://eu.udacity.com/course/intro-to-machine-learning--ud1...

[2] https://agi.mit.edu/

[3] https://www.youtube.com/watch?v=eLbMPyrw4rw&list=PL6EE0CD029...

[4] https://www.youtube.com/watch?v=2pWv7GOvuf0&list=PL7-jPKtc4r...

[5] https://www.youtube.com/watch?v=NfnWJUyUJYU&list=PLkt2uSq6rB...


One important skill you will need is feature engineering. Formal methods for it aren’t typically in ML ciriculums, but it’s worth understanding if you’re interested in applications if ML.

Deep learning addresses it to some extent, but isn’t always the best choice if you don’t have image / text data (eg tabular datasets from databases, log files) or a lot of training examples.

I’m the developer of a library called Featuretools (https://github.com/Featuretools/featuretools) which is a good tool to know for automated feature engineering. Our demos are also a useful resource to learn using some interesting datasets and problems: https://www.featuretools.com/demos


I am playing with an encrypted dataset, so this comes very useful. Any tips on how to take advantage of it or about its strengths and weaknesses?

Not a video course, but an excellent book: "Hands-On Machine Learning with Scikit-Learn and TensorFlow" by Aurélien Géron. Provides a gentle and high quality introduction to Machine Learning with practice from the very beginning. I wish I had this book when I was starting. It explains everything from data engineering, through how algorithms work, to practical applications. Everything in Python 3, covering pandas, scikit-learn, tensorflow. It is absolutely wonderful!

[1] http://shop.oreilly.com/product/0636920052289.do


As a person who did a lot of Deep Learning... learning in 2017, I think this was my favorite.

Fast.ai was fine, but I felt like most of my learning for the things I cared about came from reading research papers, watching Karpathy's CS231n lectures, and blog posts that went into detail on particular concepts.

But when at certain points I felt confused on certain concepts, Geron's book did a pretty good job explaining things slowly and in great detail, especially with respects to the code he wrote. It's still a book I'll pick up for 20-40 minutes every other day to help my mind recall about how something works.

Funnily enough, I've spent the last few months reading Sutton/Barto's Intro to Reinforcement Learning (along with Silver's lectures on DeepMind's YouTube Channel) and only realized Geron touches upon RL a little bit in the latter part of the ML book.


Have you tried Andrew Ng's course? And I don't mean the one from the start of Coursera, but the original Stanford ML Course he taught at the blackboard. This ML course has all the math done by hand instead of skipped over.

https://www.youtube.com/view_play_list?p=A89DCFA6ADACE599


bad link?


I'll give another recommendation for http://www.fast.ai/

I went through the first phase of the course as an intro to AI/DL and thought it was really great from a high-level perspective. If you have a decent understanding of Python you'll have a working model running on AWS within the first few hours of the course which is very rewarding.

It does a better job than I expected explaining the underlying intuition of the math, but doesn't dive deep into the actual formulas. There are obviously tradeoffs to this approach and if you want to continue in the field you'll need to do something to fill in this background, but as far getting your hands dirty and understanding the basics I really liked the fast.ai approach.


As a beginner I cannot recommend this class in its current form. The first lecture with its setup walk through is outdated and I have trouble understanding how to do a work around.

Part 1v2 was just released, it uses pytorch. Check it out, I was in the fall fellowship and can heartily recommend it.

This is welcomed news, thank you for the update. Can you link to v2? On the site I just see part 1, v1 with the AWS set up. This is the beginner unfriendly/outdated one I am referring to, and I don't see the updated version you're referring to.

I just checked the course forum, and Jeremy asked us not to share on high traffic sites until theyve finished the new website. I think it'll be out within a week or so, so check the fast.ai site for updates. Should be soon!

Okay, I will check in a week or so. Thanks.

I liked Yaser Abu-Mostafa's Caltech ML course. A bit dated (2012), but solid introduction to the basics (such as VC dimensions). However, given your stated preferences, other recommendations (such as fast.ai) are probably better suited.

https://work.caltech.edu/telecourse.html


I found this course pretty interesting, since it gives some perspective on what sorts of guarantees people want that DL doesn't provide, but I wouldn't necessarily recommend starting with it unless you have a theory bent.


If you're interested in learning Reinforcement Learning, then I can't recommend David Silver's lecture series highly enough. Youtube videos and slides are available for the entire thing. http://www0.cs.ucl.ac.uk/staff/D.Silver/web/Teaching.html

I wouldn't really classify this as an introductory course. Some background in A* search, Markov processes etc is required. David has a fantastic narration style and this is one of those courses where I had to step out because I had other responsibilities and not because I got bored with the content.

If you want to see projects coded in Python (random forests, word2vec, convolutional neural nets), I recently finished producing this set of videos: https://www.packtpub.com/big-data-and-business-intelligence/...


As much as I love Ng, I vastly prefer Hinton’s lectures. Something about that English accent.

If you speak french, I would recommend Hugo Larochelle channel:

https://www.youtube.com/user/hugolarochelle


I'll add another vote for course.fast.ai since I'm currently working through it. The instructors are serious about delivering a practical course that will get you right into the process of training and playing around with the code. Also, one of the best parts of their courses (they're offering a few simultaneously) is the community of learners that hang around on the forums.

Fast.ai, you can checkout the forum here: http://forums.fast.ai/

I found the "Bay Area Deep Learning School" 2-day lecture series to be very good:

https://www.youtube.com/watch?v=eyovmAtoUx0

https://www.youtube.com/watch?v=9dXiAecyJrY


A pretty short introduction to machine learning created by myself: http://lausbert.com/2018/01/14/the-shortest-introduction-to-...

Ideal to practice, learn, and teach machine learning:

https://bigml.com/ml101 https://bigml.com/education/videos



getting your feet wet

andrew ng's machine learning course: https://www.coursera.org/learn/machine-learning

to get up to date on convnet architecture

Fei-Fei Li and Karpathy's cs231n: https://cs231n.github.io/

if you want to go deep

geoff hinton's neural networks for machine learning coursera: https://www.coursera.org/learn/neural-networks


Andrew Ng's deeplearning.ai course is pretty accessible for beginners.

How about Microsoft's AI School https://aischool.microsoft.com

this talk really enlightened me: https://www.infoq.com/presentations/neural-networks-introduc...

it is not a full course, but more an introduction.


Fast.ai

Can we name the links so that we know what each is? @mindcrime

I would have but I was half asleep when I first posted that. Luckily someone else came along and labeled them in a reply. :-)

Just recently I have written a "Machine Learning for Web Developers in JavaScript" blog post [0]. If you are a web developer, it might be interesting for you. It outlines my approach of learning it and gives a couple of great resources for JavaScript enthusiasts. Otherwise, I will just post a couple of the materials I used below. It's not only video courses, because I believe it's always useful to stimulate all senses.

- [0] https://www.robinwieruch.de/machine-learning-javascript-web-...

Podcast:

- http://ocdevel.com/podcasts/machine-learning

Courses:

- https://www.coursera.org/learn/machine-learning

- https://de.udacity.com/course/machine-learning-engineer-nano...

- https://www.coursera.org/specializations/deep-learning

- http://course.fast.ai/

Books:

- https://www.amazon.com/gp/product/B014X01SS0/

- http://www.deeplearningbook.org/

- http://neuralnetworksanddeeplearning.com/

- https://www.safaribooksonline.com/library/view/deep-learning...

Math:

- http://www.fast.ai/2017/07/17/num-lin-alg/

- https://www.khanacademy.org/math/linear-algebra

- https://www.khanacademy.org/math/statistics-probability

- https://www.khanacademy.org/math/calculus-home

JavaScript ML:

- https://bri.im/

- https://github.com/javascript-machine-learning


tried fast.ai?

I took the coursera specialisation one week ago. It takes you from the very basics to some more complex modules like keras or tensorflow. If you are into it and have time, the whole 4 courses can be done in the free week: https://www.coursera.org/specializations/deep-learning



Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: