Boosting Contrastive Self-Supervised Learning with False Negative Cancellation
— AK (@ak92501) October 19, 2021
abs: https://t.co/hrLNic6CRL
github: https://t.co/FwpPAPxAhp pic.twitter.com/BNLRIZImul
Boosting Contrastive Self-Supervised Learning with False Negative Cancellation
— AK (@ak92501) October 19, 2021
abs: https://t.co/hrLNic6CRL
github: https://t.co/FwpPAPxAhp pic.twitter.com/BNLRIZImul
Breathing K-Means is an interesting extension to K-Means to make it significantly more likely to converge to lower cost minima with a single kmeans++ random init (the manuscript linked in the README has many interesting experimental results): https://t.co/4iJndEGXhH
— Olivier Grisel (@ogrisel) October 19, 2021
T0 outperforms GPT3 on 9 out of 11 benchmarks despite being 16x smaller https://t.co/llLydO6euk pic.twitter.com/9cOg4zoGIZ
— Mark Riedl is a Metaverse Company (@mark_riedl) October 19, 2021
HRFormer: High-Resolution Transformer for Dense Prediction
— AK (@ak92501) October 19, 2021
abs: https://t.co/WuCrhSHWU3
github: https://t.co/tNfI7Ba1Go pic.twitter.com/Fa1n1k4eNt
Understanding and Improving Robustness of Vision
— AK (@ak92501) October 18, 2021
Transformers through Patch-based Negative
Augmentation
abs: https://t.co/mFJURTVAHN
show that patch-based negative augmentation consistently improves robustness of ViTs across a wide set of ImageNet based robustness benchmarks pic.twitter.com/CUQrtfdzxe
1/ After a year of work, our paper on mRNA Degradation is finally out!
— Bojan Tunguz (@tunguz) October 16, 2021
paper: https://t.co/s63ik0c3Ey
code: https://t.co/UWIPSbOvHH pic.twitter.com/ooT2wvvuah
Introducing a minimalist and effective approach for vision language model pre-training that learns a single representation from both visual and language inputs and efficiently leverages scaled datasets to achieve state-of-the-art performance. Learn more ↓ https://t.co/U9DY2CZbqR
— Google AI (@GoogleAI) October 15, 2021
Hey! Got 40 seconds? ⏱️ Learn how we achieve photorealistic reposing and virtual try-on in the upcoming SIGGRAPH Asia paper *Pose with Style*. 🤩
— Jia-Bin Huang (@jbhuang0604) October 15, 2021
Paper: https://t.co/UtVyBn8eA3
Web: https://t.co/b55nH3SSDB
Brought to you by the amazing @BadourAlBahar! pic.twitter.com/Xf9jYFaOY5
Tesla officially launches its insurance using ‘real-time driving behavior’
— hardmaru (@hardmaru) October 15, 2021
Will be interesting to see long-term effects of this "machine teaching" experiment on society; whether driving habits of masses can be altered by real-time economic reward signal.https://t.co/ujEelRXG4y pic.twitter.com/Nv68Fc8SeV
bert2BERT: Towards Reusable Pretrained Language Models
— AK (@ak92501) October 15, 2021
abs: https://t.co/x7Six076zh pic.twitter.com/9ZJNvpjf8k
Symbolic Knowledge Distillation: from General Language Models to Commonsense Models
— AK (@ak92501) October 15, 2021
abs: https://t.co/tvnpkIUywh
symbolic knowledge distillation, model-to-corpus-to-model pipeline for commonsense that does not require human-authored knowledge–instead, using machine generation pic.twitter.com/XDjvNABTUF
Active Learning for Deep Object Detection via Probabilistic Modeling
— AK (@ak92501) October 14, 2021
abs: https://t.co/u5x9EA2tgZ
github: https://t.co/SRylg7UWOY pic.twitter.com/MjLgaDsK4E