After an inspiring month-long hack, our overall winner of the 2023 #SASHackathon is Team JaWaRA! Using SAS AI, ML and IoT, the team aims better predict and prevent flooding as an early warning system for Jakarta. http://
2.sas.com/6017P4ozb #ExploreSAS
MACHINE LEARNING
-
SAS Hackathon Team Wins with AI Flood Prediction System
By
–
-
SAS Hackathon Unites Data Scientists Solving Real-World Problems
By
–
Einar Halvorsen, global lead #SASHackathon: "It brings together curious minds and is so much more than a competition… For 1-month teams of data scientists and tech enthusiasts… working together to solve real-world problems and save lives." #ExploreSAS http://
2.sas.com/6011P4WSB -
Academic perspective on fundamental limitations of large language models
By
–
@rao2z is an academic leader in AI and here presents a short coherent argument on the fundamental limitations of LLMs. [[Note that all technologies have limitations and it is magical thinking (in my seven deadly sins sense) to believe that LLMs will not.]] https://
cacm.acm.org/blogs/blog-cac
m/276268-can-llms-really-reason-and-plan/fulltext
… -
Training YOLOv8 for Real-Time Pothole Detection: Model Comparison
By
–
🚀YOLOv8 is one of the fastest and most accurate YOLO models out there.https://t.co/l9JUDdcJkD
— Satya Mallick (@LearnOpenCV) 14 septembre 2023
Why not leverage it for training a real-world dataset? Let's train YOLOv8 to detect potholes in real time and compare performance between three different YOLOv8 models.… pic.twitter.com/IocN7SXl5eYOLOv8 is one of the fastest and most accurate YOLO models out there. https://
learnopencv.com/train-yolov8-o
n-custom-dataset/
… Why not leverage it for training a real-world dataset? Let's train YOLOv8 to detect potholes in real time and compare performance between three different YOLOv8 models. -
AI Partnership for Maternal and Child Health Solutions
By
–
We are excited to partner with the @MinistryWCD to develop AI-powered solutions for improving maternal, newborn, and child health. We will work on problem identification, designing AI/ML tools, and organizing training workshops on practical applied AI/ML for the ministry.
-
Frame Generation Optimization: From DeepDream to Neural Art
By
–
Each frame took 15 minutes! The original technique was ~1000 optimization steps, similar to Deepdream. You could start from the previous frame and do fewer steps, but it didn't look great. https://
arxiv.org/abs/1508.06576 -
Worldwide Longitudinal Project Predicting Psychosis Using AI
By
–
One of the most exciting projects I learned about at @acii_conf from @JustinBakerMD was this ambitious, worldwide longitudinal project towards developing tools towards predicting psychosis before it occurs
-
NeurIPS Unlearning Competition Opens on Kaggle
By
–
The NeurIPS @unlearning_2023 Competition is open for submissions @kaggle
→ https://
goo.gle/45ONCc1! With the goal of developing algorithms that can unlearn a subset of training data, the competition is open to all, with prizes for the top performers. More from our earlier blog ↓ -
SAS Explore 2026: Generative AI Analytics Insights with theCUBE
By
–
We've had a blast with @theCUBE onsite at #ExploreSAS. Lots of great moments and interviews, that you can watch on demand: http://
2.sas.com/6014PNAZq Generative AI, analytics, data science…come for the topics, stay for the talent