Eyes Tell All: Irregular Pupil Shapes Reveal GAN-generated Faces
pdf: https://t.co/lgcfsZO7dp
abs: https://t.co/ZV5TJyVAp4 pic.twitter.com/9BYyrVpxAC

— AK (@ak92501) September 2, 2021

∞-former: Infinite Memory Transformer
pdf: https://t.co/4B4sxwGEM5
abs: https://t.co/sBtFRIv7rc

propose the ∞-former, which extends the vanilla transformer with an unbounded long-term memory pic.twitter.com/0oFWpixF3b

— AK (@ak92501) September 2, 2021

Article in @petapixel about @GoogleAI's SR3 Image Super-resolution method.

Neural network-based photo upscaling will likely be commonplace in most smartphones in the near future.https://t.co/7TKE3jeTv2

— hardmaru (@hardmaru) September 1, 2021

WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU
pdf: https://t.co/6z9WZacJyk
abs: https://t.co/wUX0MThFGc pic.twitter.com/vR6V68Zdgv

— AK (@ak92501) September 1, 2021

It's finally ready:

Prodfiler, a continuous profiler that "just works" -- for C/C++/Rust/Go/JVM/Python/Perl/PHP -- no code change required, no symbols on the machine required, no service restart required.

Check out: https://t.co/EL2DHpoLkl or the blog post below. https://t.co/3HRamFZyhd

— Halvar Flake (@halvarflake) August 31, 2021

AI is influencing the world and right now most of the actors that have power over AI are in the private sector. This is probably not optimal. Here's some research from @jesswhittles and I on how to change that. https://t.co/vljhYFXyGC

— Jack Clark (@jackclarkSF) August 31, 2021

SummerTime: Text Summarization Toolkit for Non-experts
pdf: https://t.co/RCgAVCFPLx
abs: https://t.co/xuGph3jsdB
github: https://t.co/OYTgNX6u0I pic.twitter.com/txTS9Lg0Cq

— AK (@ak92501) August 31, 2021

Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learners
pdf: https://t.co/CHAhM6eO0B
abs: https://t.co/Hav0uyFLpH

can convert small language models into better few-shot learners without any prompt engineering pic.twitter.com/qlTIbLvolc

— AK (@ak92501) August 31, 2021

Hire-MLP: Vision MLP via Hierarchical Rearrangement
pdf: https://t.co/4Vjf9gQZuv
abs: https://t.co/9umSzoO621

achieves an 83.4% top-1 accuracy on ImageNet, which surpasses previous Transformer-based and MLP-based models with better trade-off for accuracy and throughput pic.twitter.com/Ftsgaaipva

— AK (@ak92501) August 31, 2021

#PyTorch re-implementation of DeepMind's Perceiver IO: A General Architecture for Structured Inputs & Outputs https://t.co/c16ftYKgzJ pic.twitter.com/srwT1TiOaU

— Alexandr Kalinin (@alxndrkalinin) August 30, 2021

Helping juniors the right amount can feel like tricky business. Let me throw out some ideas that I've found helpful. And if you have ideas, HMU. I think we all have lots to learn here. A 🧵(in no particular order of importance) ... https://t.co/w9Pt0wzt4o

— JD Long (@CMastication) August 30, 2021

What’s more, if your training is in @PyTorch, you can rather easily add this behaviour with minimal changes to your codebase, using @higherpytorch.https://t.co/U5dFLBXTHZ

— Edward Grefenstette 🇪🇺 (@egrefen) August 28, 2021