demo: https://
huggingface.co/spaces/carolin
eec/informativedrawings
…
github: https://
github.com/carolineec/inf
ormative-drawings
…
paper: https://
arxiv.org/abs/2203.12691
RESEARCH
-
AI Research Paper and Demo on Informative Drawings
By
–
-
2022 AI Progress Foundation for Greater 2023 Advances
By
–
2022 has been a great year for AI progress. Excitingly, this has laid a foundation for even more progress in 2023. Happy New Year, friends!
-
2023 Tech Wishes: AI, Learning, Video, Open Models, Robotics
By
–
In 2023, may your: – Generative AI produce beautiful images & text
– Active Learning framework ask the right questions – generations be as beautiful as image
– Model be OSS without censors
– Robotics simulation be faithful to real world Happy New Year everyone! -
Advances in Surgical AI: Skill Assessment and Patient Outcome Prediction
By
–
We made strides in surgical #AI which involves assessing the skill of surgeons, predicting patient outcomes, and discovering novel surgeon biomarkers based on multi-modal data and deep learning algorithms. @AjhungMD gives an excellent overview here
-
ZerO: Deterministic Weight Initialization with Binary Values
By
–
We developed the first fully deterministic weight initialization procedure for training any neural network. Our scheme, ZerO, initializes networks’ weights with only zeros and ones based on identity, and Hadamard transforms. ZerO achieves SOTA https://
arxiv.org/abs/2110.12661 -
Robust Vision Transformer Architecture Wins Semantic Segmentation Challenge
By
–
We also developed robust vision transformer architecture, fully attention networks (FAN), with channel-based attention for robustness. We won the Semantic Segmentation Tracking of Robust Vision Challenge at ECCV. https://
arxiv.org/abs/2210.12852 -
Neural Operators Enable Single-Step Diffusion Model Image Generation
By
–
2022 saw diffusion models for generative AI. But one downside is very slow sampling. We used neural operators to replace ODE solvers and showed that just one function evaluation is sufficient to achieve the best image generation quality. https://
arxiv.org/abs/2211.13449 -
MineDojo: Generalist AI Agents with Text Prompts in Minecraft
By
–
Decision-Making with Generalist AI Agents: We built Minedojo framework in Minecraft with text prompts for open-ended tasks using internet-scale knowledge of videos, wiki, and text information. It was an outstanding paper at NeurIPS 2022.
-
Digital Biology Foundation Model Predicts Coronavirus Variants, Wins ACM Prize
By
–
Digital biology took off. We trained largest biological foundation model for genome-scale modeling of #coronavirus to predict new variants of concern. It won the 2022 ACM Gordon Bell Special Prize for HPC-Based COVID-19 Research.
-
Neural Operators Accelerate Scientific Computing with Massive Speedups
By
–
Neural operators accelerated science. Carbon capture and storage modeling and weather forecasting achieved 4-5 orders of magnitude speedups. We improved mask optimization in lithography and inverse design of airfoils. They learned chaotic systems.