True Few-Shot Learning with Language Models
— AK (@ak92501) May 25, 2021
pdf: https://t.co/a0WvisxlJX
abs: https://t.co/7zb714iSYk
github: https://t.co/nPICjpoRuo
few-shot ability of LMs when held-out examples are unavailable, prior work significantly overestimated the true few-shot ability of LMs pic.twitter.com/k28CCi4iWk
Intriguing Properties of Vision Transformers
— AK (@ak92501) May 24, 2021
pdf: https://t.co/mMujDE3PQF
abs: https://t.co/3crBQpq9Qd
advantages of ViTs over CNNs for occlusion handling, robustness to distributional shifts and patch permutations, automatic segmentation with pixel supervision pic.twitter.com/3u4X0rl033
This is a summary of easily available datasets for generalized DALLE-pytorch training. https://t.co/yFP083YnBt #deeplearning #machinelearning #ml #ai #neuralnetworks #datascience #pytorch
— PyTorch Best Practices (@PyTorchPractice) May 23, 2021
I think every data scientist has a story about a time when someone from product or business with relative power demanded they find a specific pre-determined answer in the data. Data scientists generally push back. How this goes depends on the quality of the org and leadership.
— Hilary Mason (@hmason) May 23, 2021
An experience every data scientist either has had or will have is working for a product manager who has absolutely no concept of data but who insists their conclusions must be correct and it’s the fault of the data scientists for not being smart enough to prove them. https://t.co/oLElFHQEAf
— Emily G (not a newspaper) (@EmilyGorcenski) May 23, 2021
Average daily truck traffic across the US in 1998. 🚛🚛🚛🚛🚛 Source: https://t.co/bMp264WpT5 pic.twitter.com/kBiLi3hDH6
— Simon Kuestenmacher (@simongerman600) May 22, 2021
sometimes i think about how people in my field are building systems that directly harm people and how they either don't care to know or, worse, do know but don't care because it's making someone rich
— Rachael Tatman (@rctatman) May 21, 2021
and it just makes me very tired and very sad
Check out Know Your Data, a tool built by the PAIR team at Google to help understand datasets, improve data quality and mitigate bias issues. Try it for yourself on 71 datasets in the @TensorFlow Datasets catalog. https://t.co/b2mnaSovzk pic.twitter.com/DnRPt0TBsL
— Google AI (@GoogleAI) May 21, 2021
Today we are announcing our work on building speech recognition models without any labeled data! wav2vec-U rivals some of the best supervised systems from only two years ago.
— Michael Auli (@MichaelAuli) May 21, 2021
Paper: https://t.co/cYzF9MGu56
Blog: https://t.co/iiGmgdnCiV
Code: https://t.co/TQ56tT0unx
Check out KELM, a method of converting knowledge graphs to synthetic natural language sentences, leveraging their factual nature in order to augment existing #NLP pre-training corpora. Learn more and access the corpora at ↓ https://t.co/jbrDVHCgeF
— Google AI (@GoogleAI) May 20, 2021
I made a cookbook for taking absolute control over my @matplotlib ticks. Place them where you want, looking how you want, saying whatever you want. https://t.co/U9I3u0fqMv
— Brandon Rohrer (@_brohrer_) May 19, 2021
A Measure of Research Taste
— AK (@ak92501) May 19, 2021
pdf: https://t.co/XhtoezVihm
abs: https://t.co/R6eXxUhz27
project page: https://t.co/E8Nnk0VC0t
a citation-based measure that rewards both productivity and taste, balances impact of publications and their quantity pic.twitter.com/uRb4WWOE6P
