Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels
pdf: https://t.co/eb1qvqaO7I
abs: https://t.co/rw45sireIH
github: https://t.co/C6Po5bcKKt pic.twitter.com/KFKVBEbsfA

— AK (@ak92501) January 14, 2021

Encoder-Decoder models are going long-range in 🤗Transformers!

We just released 🤗Transformers v4.2.0 with Longformer Encoder-Decoder (LED) for long-range summarization from @i_beltagy

Summarize up to 16K tokens with 🤗's pipeline or our inference API: https://t.co/xeykBhHvMU pic.twitter.com/q2sV35V1fS

— Hugging Face (@huggingface) January 13, 2021

“Has anyone else lost interest in ML research?”

“My collaborators/advisors are mostly running after papers and don't seem to have interest in doing interesting off-the-track things. Ultimately, research has just become chasing one deadline after another.”https://t.co/oxul2zkmRg

— hardmaru (@hardmaru) January 13, 2021

This is really is a great tutorial/example for implementing a transformer from scratch, https://t.co/lU9EeIA15b https://t.co/KUNybhM8qz pic.twitter.com/NoSTauAN86

— Sebastian Raschka (@rasbt) January 13, 2021

Jeff actually links to *3* of @timnitGebru's papers in this post (after firing her last month & publicly claiming her latest paper didn't meet "Google standards", even though it was accepted to a top conference)

Model cards, Saving Face, and Closing the AI Accountability Gap https://t.co/huWK5iy9tZ

— Rachel Thomas (@math_rachel) January 13, 2021

Excited to announce that our Deep Bootstrap framework for understanding generalization in deep learning has been accepted @iclr_conf! #ICLR2021 https://t.co/LhcK6VL6zP

— Hanie Sedghi (@HanieSedghi) January 13, 2021

Overview of the amazing progress in deep learning and medical computer vision. Published in @nature #digitalmedicine with the great @AndreEsteva, @katherinechou, @syeung10, @nikhil_ai, @thisismadani, @samottaghi, @yun_liu, @EricTopol,@JeffDean https://t.co/7NYRIQlt5N pic.twitter.com/4GamYwRJOd

— Richard Socher (@RichardSocher) January 12, 2021

Folks are starting to poke around in that #Parler dataset. This map shows where the Parler users were posting from, which is roughly a population density map.

The dataset is linked in the source below. #DataScience #dataviz

Source: https://t.co/8rL6jIheOo pic.twitter.com/u0VfyXesAM

— Randy Olson (@randal_olson) January 12, 2021

Differentiable Vector Graphics Rasterization for Editing and Learning (SIGGRAPH Asia 2020)

Nice work that allows backpropagation through an image rasterizer, so we can apply the goodies that work on pixel images to vector graphicshttps://t.co/s6ooUYKsqnhttps://t.co/9jMZY7lPdS pic.twitter.com/TBEZYI3cVQ

— hardmaru (@hardmaru) January 12, 2021

Well this was unexpected: the FTC went after a privacy-violating company and required it to delete not just the data but also the models trained using the data. It reminded me why I work in tech policy—because it's only frustrating 90% of the time. https://t.co/rTCD02wj0N

— Arvind Narayanan (@random_walker) January 12, 2021

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity
pdf: https://t.co/0i6fcOuy4X
abs: https://t.co/AUKgennqZy
github: https://t.co/8QD4sJ2ckE pic.twitter.com/iDPDXj4bRR

— AK (@ak92501) January 12, 2021

We thank @svlevine for his excellent talk "Data-Driven Reinforcement Learning: Deriving Common Sense from Past Experience" last Friday, now available on our YouTube channel. https://t.co/RfFMxmLivj

— UCL CSML (@uclcsml) January 10, 2021