ArtEmis: A dataset of 439K emotion attributions and explanations from humans, on artworks from WikiArt
— hardmaru (@hardmaru) February 1, 2021
Paper: https://t.co/sRMsH8C4M8
Project: https://t.co/3471NVlKer
They train a visual captioning model using data, and below are some art critiques generated on unseen artworks: pic.twitter.com/4VyJE2Q3kY
Putting together material for my lectures next week, I stumbled upon this great figure summarizing the different design decisions / inductive biases of common neural net architectures (from: https://t.co/iHDYK4V98w) pic.twitter.com/JAcqZ3Bz6O
— Sebastian Raschka (@rasbt) January 31, 2021
Every founder loves to complain that hiring is hard, but my experience is that hiring is work but isn't hard. Throw garbage like this out the window and think about what your team needs to accomplish and what it'll take to get you there. https://t.co/3S1eRgK0OU
— Hilary Mason (@hmason) January 31, 2021
Targeted advertising for birds 🌀
— hardmaru (@hardmaru) January 31, 2021
We’re not so different, you and I 🐤pic.twitter.com/85xkwvJMVX
I'm a huge sucker for "small multiple" charts and NBA teams make a great subject for them because you can make a tidy grid with 5 rows and 6 columns. Here's a thread of some of my favorite 30-team small multiples, starting with one of my own that shows 20 years of franchise wins: pic.twitter.com/kjeFbXH91w
— Todd Whitehead (@CrumpledJumper) January 31, 2021
The majority of U.S. adults prefer to do away with the #ElectoralCollege, with conservatives being the lone holdout. #dataviz
— Randy Olson (@randal_olson) January 30, 2021
Source: https://t.co/HarvHl73Mb pic.twitter.com/V5hLSZhxol
Visualization of the two latest US Inauguration Addresses shows the most frequently used words. No reason to worry, all is GREAT. Source: https://t.co/g0uBahW5fu pic.twitter.com/nf0zeiEKjf
— Simon Kuestenmacher (@simongerman600) January 30, 2021
✍️ Text2Image v.3.0 🎨
— Bomze (@tg_bomze) January 30, 2021
This neural network utilizes CLIP (with CMA-ES optimizer) to steer BigGAN towards images that match a textual prompt in any language.
👍Special thanks to @eyaler
📙Try it online: https://t.co/iFRPXf2t0v pic.twitter.com/pNYbmz99D6
Good to see the climbing in US vaccinations 👍
— Eric Topol (@EricTopol) January 30, 2021
Peak so far of 1.7 M yesterday and another 1.6 M today pic.twitter.com/kMk7G9ZQL6
Introducing TT-Rec, a new method to dramatically shrink memory-intensive Deep Learning Recommendation Models and make them easier to deploy at scale. https://t.co/IiGmIakkZZ pic.twitter.com/pjtGhZhOG3
— Facebook AI (@facebookai) January 29, 2021
🆕 We've released Stanza v1.2—our Python neural NLP toolkit for PoS, parsing, NER for dozens of human languages. UD 2.7 models, multi-document support, faster tokenization, fix race conditions, fixes “data gap bugs” with tokenization & deps in many las https://t.co/Lm2WdAddTz
— Stanford NLP Group (@stanfordnlp) January 29, 2021
Check out a new strategy for intermediate pre-training and data filtering that enables table parsing models to learn better, faster & from less data, achieving 4x gains in speed & memory utilization without a significant drop in performance. Read more ↓ https://t.co/tjY4GCYtRW
— Google AI (@GoogleAI) January 29, 2021
