Wide, ML Interviews, Llama 4 Live
New Releases (4/18)
Wide
I just finished a new chapter of the India Edition of our popular Deep Learning Math Workbook. Here’s the preview. This chapter teaches you how a deep neural network grows "wide."
Machine Learning Interviews
Got a question about interviews for ML jobs, don't ask me and read this book instead, because as a professor I never applied to an ML job. I read this book and recorded a short video to highlight a few of my favorite figures.
Recently, I had the pleasure of joining Susan Shu Chang, the author of this wonderful book, on a live event about Llama 4. Susan offered insightful take on Llama 4 from a developer career perspective. If you are looking to get an ML job, you should check out her book.
Llama 4 Live
Thank you for the HUGE turnout to our Llama live event. If you missed it, you can watch the video recording we just released.
Mohsena Ashraf, my PhD student, did a fantastic job hosting the event.
I took the opportunity to go over different attention mechanisms using a highlighter by hand. ✍️
I also learned a ton from my follow panelists Susan Shu Chang and Eduardo Ordax about evaluation, job opportunities, open source contributions, and industrial impacts. I believe many attendees learned a lot too, judging from the active engagement in the Q/A.
Will we get together again? I'd love to.
A few comments on Chapter 10 PDF:
* The answer for box #6 is given as 0, but it looks like the correct answer should say 3.
* The are multiple solutions for boxes 10 and 11. I was expecting unique answers for these boxes.
* The are multiple solutions for boxes 12 and 13. I was expecting unique answers for these boxes.
* The answers for boxes 48, 49, and 50 are given as 2, 0, and 7, respectively, but it looks like the correct answers should be 9, 8, and 0, respectively. On the same page, the value for B2 is given as 9, while in fact it should say 6.