RTMV: A Ray-Traced Multi-View Synthetic Dataset for Novel View Synthesis
— AK (@ak92501) May 17, 2022
abs: https://t.co/4kdIL9g29r pic.twitter.com/JmXpCqFPXx
RTMV: A Ray-Traced Multi-View Synthetic Dataset for Novel View Synthesis
— AK (@ak92501) May 17, 2022
abs: https://t.co/4kdIL9g29r pic.twitter.com/JmXpCqFPXx
A short thread about DeepMind's recent GATO paper. It trains a basic transformer on an impressive number of datasets pic.twitter.com/ncpP8aLFgs
— Eric Jang (@ericjang11) May 13, 2022
Simple Open-Vocabulary Object Detection with Vision Transformers
— AK (@ak92501) May 13, 2022
abs: https://t.co/ytb2Tvliu1 pic.twitter.com/0xUjokjLcB
"Whether to go with a decoder-only or encoder-decoder transformer?"
— Mostafa Dehghani (@m__dehghani) May 12, 2022
It turned out that this question on the architecture of the model is not actually that important!
You just need the right objective function and a simple prompting to switch mode during pretraining/finetuning. pic.twitter.com/lanaYmHynW
Our 50+-year review of #forecast combinations is now out. @YanfeiKang @f3ngli @Xia0qianWang https://t.co/MTBhI2trNe
— Rob J Hyndman (@robjhyndman) May 10, 2022
Presenting: The AI Economist
— Richard Socher (@RichardSocher) May 5, 2022
This is one of the most impactful lines of AI research I've ever worked on. Its implications span from immediately impactful to highly philosophical.
Blog: https://t.co/vmkYroUpUY
Paper: https://t.co/xjZXMAyQMn
A 🧵 with high level take-aways: pic.twitter.com/xjBl37gzUt
Use All The Labels: A Hierarchical Multi-Label Contrastive Learning Framework
— AK (@ak92501) April 29, 2022
abs: https://t.co/N4PZEKFIey
github: https://t.co/CsRcp667Ge pic.twitter.com/z4IfiuTfpj
Interesting work on making the stride parameter of convolutional layers learnable. No, it doesn't require discrete optimization as in conventional NAS but is based on gradient descent. Code curently only for TF, but should be portable to PyTorch. https://t.co/QtRrY4KpSZ pic.twitter.com/5wFOtTPpn9
— Sebastian Raschka (@rasbt) April 27, 2022
Understanding The Robustness in Vision Transformers
— AK (@ak92501) April 27, 2022
abs: https://t.co/reQ35twd49
model achieves a state-of-the-art 87.1% accuracy and 35.8% mCE on ImageNet-1k and ImageNet-C with 76.8M parameters pic.twitter.com/H5pTsUpEE0
High Quality Segmentation for Ultra High-resolution Images
— AK (@ak92501) April 25, 2022
abs: https://t.co/ljV7jl3olc
github: https://t.co/57fX0G15IM pic.twitter.com/SEZAHEQe8K
Autoregressive Search Engines: Generating Substrings as Document Identifiers
— AK (@ak92501) April 25, 2022
abs: https://t.co/eIR7VVbfQ7
github: https://t.co/PwAidm4h7g pic.twitter.com/KoRfty8s35
I missed that SalesForce has released a collection of code models (including weights!) ranging from 350M params all the way up to 16B params! The largest model outperforms Codex on the HumanEval dataset https://t.co/hIpg7aHTXN
— Brendan Dolan-Gavitt (@moyix) April 20, 2022