tmabraham

This behavior is not exclusive to OpenAI. NVIDIA did this too. Originally StyleGAN3 was published under the name "Alias-free GAN" and the paper itself uses that terminology.

I love that the author highlighted Prof. Robert Ghrist's great material. I took his calculus classes through Coursera maybe about 8-9 years ago. He just makes everything exciting and his visuals are just beautiful. For example, within the first few lectures, he made it feel like Taylor series was like the coolest thing ever. Highly recommend checking his lectures out. Check out his website: https://www2.math.upenn.edu/~ghrist/

The fastai book actually makes a nice comparison between the systems described in PDP and modern deep learning.

> In fact, the approach laid out in PDP is very similar to the approach used in today's neural networks.

From: https://github.com/fastai/fastbook/blob/master/01_intro.ipyn...

Haha yesterday in #off-topic I called you the "grandfather of EleutherAI"

How is this better than Voila, which is developed by the Jupyter Notebook team? https://github.com/voila-dashboards/voila

Tqdm is awesome and I use it all the time!

Another great alternative is fastprogress: https://github.com/fastai/fastprogress

It often works better in Jupyter Notebooks.

> Given that this is just structured data behind the scenes, it will remain a puzzle as to why Spotify decided not to enable Wrapped in the browser.

There is no puzzle here, they are 100% trying to force people to use the app. I suspect this is because it is much more likely people will pay for subscriptions since the app is severely limited, especially in comparison to the web/desktop version. IIRC you can't even see the songs in a playlist and it's fixed on shuffle for the mobile version.

It's important to also credit Alfredo Canziani who organized much of the course.

I use https://fastpages.fast.ai/ for my blogging platform, so I use Utterances through that... It's a great system IMO

The visualizer looks fun: https://twitter.com/minimaxir/status/1447679798822649856

haha this a project my current Ph.D. advisor worked on (he's the lead author of that study).

He likes to say AI=avian intelligence...

The website is down, I am getting a 404 page.

This is bringing back memories of my college physics class studying! Great website with simple bite-sized explanations of key topics in physics...

Also, they have for other fields too:

http://hyperphysics.phy-astr.gsu.edu/hbase/chemical/chemcon....

http://hyperphysics.phy-astr.gsu.edu/hbase/geophys/geophys.h...

http://hyperphysics.phy-astr.gsu.edu/hbase/biology/biocon.ht...

They have some of that already: http://hyperphysics.phy-astr.gsu.edu/hbase/Chemical/chemcon.... http://hyperphysics.phy-astr.gsu.edu/hbase/Biology/biocon.ht...

If you follow through with the entire fastai course/book, then you will learn how to build an ML framework (the fastai library) from scratch. Yes, it starts out by saying "you can do deep learning with 4 lines of code" but it goes through all the inner workings, and talks about some pretty advanced stuff. Of course, no book can cover everything, but it also functions quite well as a jumping off point for more advanced self-study.

This is quite interesting, and I hope the course material will be made available for others (including myself) to check out.

> None of my students had a course in abstract algebra before. But they didn't need it to get their hands dirty and start playing around.

This really resonates with me, as a deep learning researcher who started out by taking the fast.ai course. The fast.ai course highlighted the same ideas of getting your hands dirty and exploring in a top-down approach. It's great to see how this sort of approach translates to other fields.

Yeah it doesn't seem to be a profitable scam, but rather probably just revenge or to protect themselves...

Wow, this is shocking that the scammers out-scammed Jim Browning, a well-known YouTuber fighting cyber-scams. I hope the channel is restored as soon as possible! I wonder what recourses are available when you delete a YouTube channel yourself...

I just wanted to say that your deprocrastination.co Chrome extension is quite good and very helpful (I am just using the free version). Your blog posts are also quite nice. Quite exciting to see these course materials, excited to delve into it!

Memory palaces are amazing! A while back I tried memorizing a deck of cards. I spent several days trying to memorize via rote memorization, but it was quite difficult for me to memorize more than ~20 cards. One day I tried memory palaces and immediately was able to memorize 40-50 cards and finally was able to master memorization of 52 cards after a couple of practice rounds.

I think I might have aphantasia, but a while back I learned successfully learned how to use a memory palace...

Ah that's unfortunate, I didn't know.

That said, my point still stands. Additionally, the comparison they did is massively unfair. I'm pretty sure Kaggle competitions have most of the feature they claim it doesn't have (ex: multiple phases, custom metrics, evaluation in environments).

IDK why this is being downvoted, it was indeed published two year ago, they just apparently repackaged this as an easier-to-use tool: https://arxiv.org/abs/1903.07291

To me this looks more like a competitor to CodaLab than Kaggle, and seems to target the ML research competition market. It seems unlike Kaggle, it doesn't seem to support things like notebooks, discussions, etc.

I might add that the comparison to Kaggle is added by the OP, I don't see it mentioned on the website anywhere.

A recent paper on this: https://www.nature.com/articles/s41551-020-0534-9

I took the fast.ai course and now I am doing a Ph.D. in Biomedical Engineering focused on applying deep learning to microscopy.

I don't think fast.ai is enough if you want to do theoretical research in deep learning, but it certainly provides enough to work on practical problems with deep learning. That said, many of us in the fastai community are able to delve deep into, understand, and implement recent deep learning papers and even develop novel techniques. So I think with a little extra studying, one could go easily transition to core deep learning research.

I found a paper from 2019 in the references of this current paper: https://arxiv.org/abs/1903.09760

overall electronics supply chain issues, and I heard something about cryptocurrency mining also being involved with supply issues for hard drives.

Biomedical Engineering is the way to go...

A good question, one of the paper's authors had the following responses on Twitter:

https://twitter.com/manorlaboratory/status/13887170085444198... https://twitter.com/manorlaboratory/status/13887291512893153...