What's Next for Fastai/ with Jeremy Howards

 There's a very interesting and informative recent interview with fastai's Jeremy Howard by Sam Charrington of TWIML AI.  I first heard it on the TWIML AI Podcast, which you can access here.

It's also available in a slightly longer format as a video if you want to see everyone's smiling faces in the zoom interview.


Jeremy gets into several really interesting topics, which we're going to briefly talk about below.  We'll expand out to provide a little more info for some of thee topics.  But you should really watch the keynote, or listen to the podcast to hear the whole thing.


The interview starts off with some info on Jeremy's background, and then dives into a conversation focused on the fastai v2 release, and it's associated book "Deep Learning for Coders with fastai and PyTorch: AI Applications Without a PhD'.

The Future of fastai is then discussed.  As we alluded to in a recent HTC Seminar Series post that was an interview with Chris Lattner, Chris leaving Google has put somewhat of a dent into Jeremy's current thinking on Swift as a possible alternative for a future version of fastai.  He also talks about the Julia language a little bit.  And some recent static designs trends in the Python language community he's not sure are a good thing.

One really interesting  future research direction at fastai discussed in the podcast that is already happening is self-supervised learning.  Self-supervised learning has been looked into quite a bit in the NLP world.  Many of the most exciting recent directions in NPL have been driven by this approach.  It is a hot current research topic in computer vision and imaging.

For more information on this topic:

Here's a quick overview of the general concept of Self-Supervised Learning.



Here's a post on 'Self-supervised Learning and Computer Vision' by Jeremy.  As you would expect, the clearest explanation of the 3 links i'm providing here for more info.

Here's a medium post on 'Self-Supervised Representation Learning' in image space.  

A recent paper on 'Self-supervised Domain Adaption for Computer Vision Tasks'.


Nbdev is a library that allows you to develop a python library in Jupyter Notebooks.  This lets you put all of your code, tests, and documentation in one place.  So you can think of it as a modern implementation of a lot of Donald Knuth's hopes for literate programming.


Nbdev provides a lot of useful features, including automatically generating docs from Jupyter notebooks, and a robust two-way synch between notebooks and source code.  This allows you to use an IDE if you wish for code editing and navigation.  It also provides a way to easily host your documentation for free using GitHub pages.

You can learn all about nbdev here.


Jeremy never ceases to amaze me.  We're all anxiously awaiting the upcoming fastai v2 Part 2 course series of lectures to come into existence at some point in the hopefully not too far away future.

The HTC Education Series: 'Getting Started with Deep Learning' course is based on the fastai v2 course.  We include additional material in the HTC course to try and supplement what is presented in the fastai course.

If you have been taking the course, then after reading Jeremy's post above on 'Self-supervised Learning and Computer Vision' , you will realize you already know a lot more about this subject then you might have expected.


Comments

Post a Comment

Popular posts from this blog

CycleGAN: a GAN architecture for learning unpaired image to image transformations

Image
 Like the Doublemint Twins touting the joys of Doublemint gum, 2 GANs are surely better than 1 GAN.  Especially if we package them together inside of one meta GAN module.  And this is exactly what the CycleGAN architecture does. Have you ever harbored dark secrets of turning a horse into a zebra?  The CycleGAN was developed to do just that. Learn how to turn a horse into a zebra.  And more. Now right away you can notice a difference between the image to image transformation GAN architectures we've been discussing over the last few posts.  Those last few posts described systems that learn from a database of matched input-output image pairs. And if your goal is to turn an edge representation into a nicely filled in continuous tone image, it's easy to build your database of matched input-output image pairs that your GAN system can then learn off of.  Take a continuous tone image (which will be the output of the database pair entry), then run it through an edge detector algorithm. 
Read more

Pix2Pix: a GAN architecture for image to image transformation

Image
 I thought following up yesterday's TraVelGAN post with a Pix2Pix GAN post would be useful to compare what is going on in the 2 architectures.  Two different approaches to the same problem. I stole this Pix2Pix Overview slide below from an excellent deeplearning.ai GAN course (note that they borrowed it from the original paper) because it gives you a good feel for what is going on inside of the Pix2Pix architecture.    Note how the Generator part is very much like an auto-encoder architecture, but rebuilt using the U-Net architecture features (based on skip-connections) that fastai has been discussing in their courses for several years before it became more widely known to the deep learning community at large  (and which originally came from an obscure medical image segmentation paper) . So the Generator in this Pix2Pix GAN is really pretty sophisticated, consisting of a whole image to image auto-encoder network with U-Net skip connections to generate better image quality at highe
Read more

Smart Fabrics

Image
"What makes smart fabrics revolutionary is that they have the ability to do many things that traditional fabrics cannot, including communicate, transform, conduct energy and even grow."  This is a quote from Rebeccah Pailes-Friedman, who is a professor at Pratt with a research focus on developing wearable technology and smart textiles. Enhancing textiles can be viewed through different lens.  Enhancements can be aesthetic, or performance boosting, or durability boosting. 
Read more