attendseg on-device semantic segmentation
AttendSeg, a neural network developed by DarwinAI and the University of Waterloo, can perform semantic segmentation on edge devices.
end user
Employees, not the employers, are now running the show. The age of end-user first SaaS is here, it’s either adapt or die.
reinforcement learning real-world
UC Berkeley's Sergey Levine “self-supervised offline reinforcement learning” as a promising direction of research for AI.
microsoft hugging face collaboration
Hugging Face launched Endpoints on Azure in collaboration with Microsoft. But creating a business around transformers presents challenges that favor large tech companies and put companies like Hugging Face at a disadvantage.
transformer model
Switching transformer models from deep to wide architecture results in significant improvements in speed, memory, and interpretability.
Will we manage to regulate data ownership before digital dictatorships emerge? The future of life itself might depend on it.
The smart toys that parents give their children these days keep them busy. But what’s the cost of having less-fussy kids with help from tech?
Proponents of free speech dream of forming alternative digital content universes controlled by people instead of corporations. The invention of blockchain may be the missing technology that makes that possible.
Regulating facial recognition technology will help prevent abuse of the technology by law enforcement. But is the big tech's support for regulation sincere?
Human mind vs artificial intelligence
While scientists have pitted neural networks and symbolic AI against each other for decades, Researchers at IBM and MIT have proven that combining them creates something that is stronger than the sum of its parts.