Facebook used 1 billion photos from Instagram to help an artificial intelligence system learn more easily how to perceive what is visible within them, and we are talking about a truly incredible achievement for the American company.
Facebook used the photos of users on Instagram without informing them in advance about this, saying that the use of photos for research is a measure written in the terms and conditions of use for such a popular platform.
Facebook chose the name SEER for the system based on artificial intelligence, and this is because it "sees" what the pictures captured, just as people see everything that appears in them, only that here we are talking about a software that achieves this.
“Self-supervised learning has long been a focus for Facebook AI because it allows machines to learn directly from the vast amount of information available in the world, rather than just from training data created specifically for AI research.
Self-supervised learning has incredible ramifications for the future of computer vision, as well as other areas of research. Eliminating the need for human annotations and metadata allows the computer vision community to work with larger and more diverse datasets, learn from random public images, and potentially mitigate some of the biases that come into play with data cleaning.
Self-supervised learning can also help specialize models in areas where we have limited images or metadata, such as medical imaging. And without the labor required for tagging, models can be created and deployed faster, enabling faster and more accurate responses to rapidly evolving situations.”
