GANs & Roses + A Resource Treasure Chest
Handpicked creme-de-la-creme AI resources just for you
Here are a few quick updates before this week's goodies.
Beta reading for AI for the Rest of Us is going well, and I've received a lot of helpful feedback so far. So if you signed up to help and have yet to receive an email from me, don't worry. I'm definitely counting on you. You'll surely receive an email in the next few weeks as I juggle editing and writing.
Speaking of writing, I've written about 20k words in the last week alone. At a certain point, I hit a rut and couldn't think of how to phrase a concept in simpler terms. To overcome that, I wrote a piece of short fiction about the imaginary world of Neural Nets. You can read that if you're into fiction here.
Illustrations are going on in full swing as well! Here's a short behind-the-scenes timelapse of one of the illustrations from the book:
Also, thanks to the many readers who've left feedback on what they'd like to see in upcoming editions. I'll be sure to plan my topics to cover what you'd like to see. Have feedback? Use the form at the end of each edition. It only takes a minute 🙂.
AI is progressing at a crazy pace, and this week, I've collected some resources that I found interesting to learn from. I hope you find value in them too!
This Week on Gradient Ascent:
[Consider reading] Drag Your GAN - Step Aside Photoshop 🎨
[Check out] LLM University from Cohere 🧑🎓
[Consider reading] Joint Image & Video Transformers from Google 📸
[Check out] Incredibly Well-Curated Computer Vision Resources 📚
[Definitely Watch] A talk on AI ethics 🗨️
[Check out] A series on prompt engineering 🔡
Resources To Consider:
Drag Your GAN - What Magic is This?
Paper: https://arxiv.org/abs/2305.10973
In this paper, the authors explore a powerful way to control GANs. A user can precisely deform an image by simply "dragging" points on an image. In fact, you can manipulate shape, pose, expression, and layout with ease. I know it sounds crazy, but watch the GIF and let me know if your jaw didn't drop.
Cohere's LLM University
Link: https://docs.cohere.com/docs/llmu
Cohere's LLM curriculum covers everything from how LLMs work to how you can use them to build and deploy apps. Definitely check this resource out. It's packed with helpful information.
Google's Transformer for Video Understanding
Link: https://ai.googleblog.com/2023/05/sparse-video-tubes-for-joint-video-and.html
Google's new paper reimagines how a Vision transformer (ViT) can be used for Video understanding. Using a series of "sparse video tubes," this model can be used for video and image-only tasks. The blog article above covers the high-level detail of this method works, so it's worth checking out.
Curated Computer Vision Resources
Link: https://ground-truth.beehiiv.com/p/ground-truth-computer-vision-newsletter-33
A gazillion newsletters cover AI and a large chunk of them aggregate resources for their readers. However, only a few of them curate resources intentionally and well. Dasha Gurova's "Ground Truth" newsletter does an incredible job of filtering out the signal from the noise. She focuses entirely on Computer Vision, so if you're working in this field, definitely consider subscribing to it. I particularly liked the issue linked above.
Must Watch Talk on AI Ethics
With the recent proliferation of AI, ethics, and safety have become supremely important. Dr. Rachel Thomas, the co-founder of fast.ai, is an expert in AI ethics. Please watch her talk on this subject. It's only 20 minutes and very eye-opening.
Every time I attempt a reasoned reading of your pieces, my head's spinning and I have to take them a bite at a time. But then I feel this awesome sense of accomplishment for having learned something new, so out there relative to my own circle of competences, but fascinating and useful. Thank you, Sairam. This is like a mission, and you're accomplishing it with flying colors :)