Blog post on adversarial reprogramming: rajatvd.github.io/Exploring-Adve…

I gave a talk at #PySangamam on Generators in Python, find the slides here: rajatvd.github.io/PySangamam/

Check out my new post on Generating Words from Embeddings: rajatvd.github.io/Generating-Wor… I've uploaded the code for the project on my github here: github.com/rajatvd/WordGe… Go ahead and sample some words and share any cool looking ones you find!

Check out my new blog post on Neural ODEs and Adversarial Attacks: rajatvd.github.io/Neural-ODE-Adv… … My experiments lead to some interesting parallels between adversarial robustness and equilibria of dynamical systems. Read the post for all the details!

I finally got around to making a colab notebook for my post on Generating Words from Embeddings. You can get straight into sampling new words from your browser in a matter of minutes! colab.research.google.com/drive/1f_vW0k8… The blog post: rajatvd.github.io/Generating-Wor…

I gave a talk titled "Generators Explained" PyCon India. You can find the slides here: rajatvd.github.io/PyCon/ I had a really great time meeting new people there, and I particularly enjoyed dabeaz's mind blowing keynote talk on stack machines, python, and WebAssembly!

New blog post on Visualizing Tensor Operations with Factor Graphs: rajatvd.github.io/Factor-Graphs/ This first post covers basic intuitions and features a neat visual proof! I plan to cover message passing and more in future posts. Animations made using Grant Sanderson's manim.

Looks like the new GPT-2 model prefers torch over tensorflow so much that it will overwrite your tf import. I don't think that's a good idea for a tutorial. #talktotransformer

# on technical accessibility One interesting observation I think back to often: - when I first published the micrograd repo, it got some traction on GitHub but then somewhat stagnated and it didn't seem that people cared much. - then I made the video building it from scratch,

In 2017, I did something dumb. There was a total solar eclipse passing through the US, but I was in NY, far from the path. Was I really gonna get on a plane to see a cool thing for two minutes? Nah. I had shit to do. The day came. I put my stupid glasses on and saw the sun

just finally grokked how NNs are basically gaussian processes

New blog post on Floating Point numbers: eigentales.com/Floating-Point/ I highlight a perspective and intuition about floating point that I haven't seen emphasized before. TL;DR: Floating point is fixed point in log space!

most foundational concept in deep learning that no one understands is probably the Neural Tangent Kernel (NTK) this line of work studies neural networks of *infinite width*, which explain a lot about normal finite-width NNs and there is exactly one Very Good blog post on them:

five years ago, when I was struggling to get a grasp of NTK, I remember coming across Rajat V D 's blog which is when it all properly clicked in my head.

1/10 ML can solve PDEs – but precision🔬is still a challenge. Towards high-precision methods for scientific problems, we introduce BWLer 🎳, a new architecture for physics-informed learning achieving (near-)machine-precision (up to 10⁻¹² RMSE) on benchmark PDEs. 🧵How it works: