Depth First Learning retweeted
I LOVE this talk by @ShriramKMurthi on how to design a curriculum to teach programming.
youtube.com/watch?v=5c0BvOlR…
It's tempting to jump to solution X. He pulls back to focus on *constraints*, just as you'd do in an eng design problem. 🧵
Depth First Learning retweeted
Very informative interview about @colinraffel's motivations and the work behind T5.
I also want to give a shout out to @pseudovirtual, @ellenchisa , and Michael McCanna, without whom we wouldn't have been inspired to do this series.
We're back! With T5 by Colin Raffel et al. We're trying out a different format, with an audio (and transcript) interview between @cinjoncin and @colinraffel.
Check out the interview here: depthfirstlearning.com/2021/…. Some choice quotes in the thread 👇
Depth First Learning retweeted
Stein Variational Gradient Descent Tutorial reddit.com/r/MachineLearning…
Depth First Learning retweeted
Amazing to see the hard work of @JamesAllingham here! 👏
We are super proud to release our fifth depth-first curriculum. It’s on WGAN, by @JamesAllingham, and by far our best yet. Check it out at depthfirstlearning.com/2019/…. We also want to give a big shout out to the @DeepIndaba for enabling this collaboration. Highly recommend attending!
Depth First Learning retweeted
If anyone has any questions about @DepthFirstLearn and what it is like to create a curriculum, the WGAN curriculum, or about @DeepIndaba - please let me know! Also, if anyone has feedback on the curriculum please us know so we can make it as awesome as possible!
Depth First Learning retweeted
Welcome to ICLR!
I'll be giving a talk tomorrow (Monday) at 3:40pm at the reproducibility workshop on how to structure research code for /understanding/, with a focus on Colab notebooks and
@DepthFirstLearn.
Happy to chat afterwards
Full schedule here: sites.google.com/view/icml-r…
Depth First Learning retweeted
I had an extremely positive experience with @DepthFirstLearn - highly recommend! Great structured way to study something very specific in great depth. I especially enjoyed getting elbow-deep into a topic that was quite far from my usual interests!
Want to go deep into notable ML papers and fundamentals?
Our 2019 fellows are running intimate online groups on:
• Stein Variational Gradient Descent
• Spherical CNNs
• Resurrecting the sigmoid through dynamical isometry
• Normalizing flows
👉 goo.gl/forms/hAJCMeGXxvUaJEe…
Depth First Learning retweeted
I would be discussing Spherical CNNs by @TacoCohen and @wellingmax #ICLR2018 #bestpaper. Explore #CNNs #GroupTheory #calculus If interested in this or one of the other groups, do sign up here: depthfirstlearning.com/2019/… @JaneStreetGroup @DepthFirstLearn #happylearning
Depth First Learning retweeted
Very excited to be an inaugural 2019 DFL Jane Street Fellow along with @skroon @vinayramasesh @rileyfedmunds Piyush & Bhairav. Thanks to @JaneStreetGroup,@DepthFirstLearn for this wonderful initiative to make research papers more accessible to the readers.
Replying to @DepthFirstLearn
Dr Sandhya Prabhakaran (Spherical CNNs): Research Fellow at Memorial Sloan Kettering Cancer Centre in NYC after obtaining her PhD at the Univ of Basel. @sandhya212
Depth First Learning retweeted
Thanks to @JaneStreetGroup and @DepthFirstLearn for this exciting initiative! I'm looking forward to exploring @shakir_za and @DeepSpiker's #normalizingflows with enthusiastic students - sign up if you're interested in this or one of the other groups at depthfirstlearning.com/2019/…
Depth First Learning retweeted
Depthfirstlearning is very useful if you want to dive into advanced topics 🧐. I use it to understand TRPO. Definitely check out their curriculum! Looking forward to more topics. depthfirstlearning.com @DepthFirstLearn
Depth First Learning retweeted
Attending #ICML2018? Visit the Reproducibility Workshop (goo.gl/h993at) to hear new work from the Google Brain team on a practical “cookbook” for #GAN research (arxiv.org/abs/1807.04720), along with AI Resident work on new educational framework, depthfirstlearning.com















