Our first challenge for this year is the Open Cities AI Challenge by GFDRR WorldBank DrivenData Labs za. We will build a deep learning model to identify building footprints from aerial imagery across diverse African cities. Hasan Nasrallah

Google Dataset Search is now officially out of beta. "Dataset Search has indexed almost 25 million of these datasets, giving you a single place to search for datasets & find links to where the data is." Nice work, Natasha Noy and everyone else involved! blog.google/products/searc…

Today, we moved our NVIDIA TITAN Xp GPU to a new workstation with enhanced cooling and storage capabilities, 500 GB SSD and 64 GB of RAM. Hasan Nasrallah

Bored Yann LeCun CDC = Center for Deep Convolutions

Our team at CNRS Lebanon ranked 26 out of 1106 at World Bank OpenCitiesAI Challenge where we built a scale invariant deep learning model to identify buildings from aerial imagery across diverse cities. Thanks Hasan Nasrallah GFDRR WorldBank DrivenData Labs za

Building segmentation results using our classifier from the World Bank OpenCitiesAI Challenge. Violet Blue pixels correspond to detected building mask. Hasan Nasrallah GFDRR WorldBank DrivenData Labs za Dave Luo

For students and others interested in expanding their knowledge of AI during this period, we thought it might be helpful to ask our researchers what they consider to be the most impactful and insightful resources available to use #AtHomeWithAI (1/9)

#biking especially in those critical economical conditions in Lebanon

ولمّا العالم يسبولها بتطلع بتبكي من التنمر! هلا هيلا هو .. الِك ولعيلتِك❤

Beirut!

We turn 4 this week! It seems like yesterday that we kicked off SpaceNet 1. Our GM, IQT CosmiQ Works' Ryan Lewis, reflects on lessons learned, our strategic focus & next steps. Thank you to all of our Partners for making this possible! #SpaceNetTurns4 ow.ly/CGHp50AVDWp

In red, fully damaged units following #Beirut blast using our #DeepLearning models for buildings' segmentation and change detection from aerial images provided by #KhalifaSAT. Thanks to MBR Space Centre وكالة الإمارات للفضاء Ghaleb Faour. Results by my graduate student Hasan Nasrallah

blog.maxar.com/earth-intellig…

🎉 Introducing...Papers with Libraries! We are now indexing ML libraries from the community. We start with the amazing TIMM by Ross Wightman. Explore 330+ models, visualize architectures, compare results and hyperparameters, and so much more! paperswithcode.com/lib/timm

What makes radar satellite imaging unique? It can see at night when others can’t. Capella's nighttime imagery shows the #EverGiven at 9:24 PM Egyptian time still blocking the #SuezCanal. No optical satellite will be able to report on this for hours. #SAR #allweather #nightvision

EfficientNetV2: Smaller Models and Faster Training With progressive learning, EfficientNetV2 significantly outperforms previous models on ImageNet, including ViT by 2.0% acc. while training 5x-11x faster. abs: arxiv.org/abs/2104.00298 code: github.com/google/automl/…

Check out our pre-print recently submitted to JSTARS that features the first Lebanese Solar rooftop potential map... love to hear your thoughts: arxiv.org/abs/2111.11397

our article mentioned in the newsletter this morning Hasan Nasrallah جريدة الأخبار - Al-Akhbar Maher Salameh al-akhbar.com/Issues/339455/…

Our paper entitled "Lebanon Solar Rooftop Potential Assessment using Buildings Segmentation from Aerial Images" has has just been published on IEEE Xplore: ieeexplore.ieee.org/document/97936… IEEE Geoscience and Remote Sensing Society Hasan Nasrallah Xiaoxiang ZHU @zhu_xlab International AI Future Lab on AI4EO IEEE Xplore

was just informed by Radwan Chouaib that our AI urban detection model is performing extremely well😍🥳 Maps committed 5 false positive and 1 false negative for this tile whereas ours had 100% accuracy: geoai.cnrs.edu.lb/urbanmodels/ cc Hasan Nasrallah CNRS Lebanon Ghaleb Faour Tamara ELZEIN 🇱🇧