Facebook launches what it calls a “universal production recognition model”. It uses artificial intelligence for identifying consumer goods, from fast cars and fast fashion to furniture.
This is the first step towards the future. Moreover, in that future, products in every image on its site can be identified and potentially be bought. Manohar Paluri is head of Applied Computer Vision at Facebook. He told the Verge that they want to make everything and anything on the platform shoppable. Thus, this will become available whenever the experience feels right. They appear to have quite is a grand vision.
Production recognition is the first in a slew of artificial intelligence-powered updates coming soon to its e-commerce platforms. That is what the company says. Eventually, these will combine augmented reality, even digital assistants, and augmented reality for creating what it calls a “social-first” shopping experience. Moreover, it launched a feature today called shops. It lets small businesses set up free storefronts on Instagram and Facebook.
Fashion will be a central part of this innovation. The company suggests that a future Facebook “artificial intelligence fashion stylist” can offer users personalized shopping recommendations.
That bases on their daily suggestions and a wardrobe for outfits tailored to their schedule and the weather.
Tamara Berg is a Facebook research scientist. She told The Verge that this is something she has wanted to build ever since she watched the movie Clueless. Thus, back in 1995, they had already imagined this future. Nevertheless, the technology she thinks is now finally ready to make that imagination come true.
So, as the reference to Clueless suggests, those features are not exactly new ideas. They have been tested and tried for years, even outside of Hollywood, often with varying results.
Amazon already built its artificial intelligence-powered assistant with the Echo Look. Nevertheless, it is not particularly popular as of now. Moreover, since at least the Amazon Fire Phone, using machine vision for identifying and shopping products has become a reality. Meanwhile, online shopping platforms, like eBay, are already using artificial intelligence to speed up the process of lists items for sale. Moreover, Amazon is one of several firms that has launched its own “Shazam for Clothes”, thanks to machine learning.
Facebook is saying what makes its tools different are its accuracy and scope. Thus, the company now has a new production recognition tool. Its name is GrokNet.
GrokNet is identifying tens of thousands of different attributes in an image. Those range from things like size and colour to specific brands.
Moreover, Facebook has already GrokNet on the Facebook Marketplace. There, by identifying what is in them and generating short descriptions, GrokNet helps users quickly list items for sale. For instance, you can upload your couch’s photo, and the marketplace will suggest listing it as sectional, leather, back sofa.
The company is also testing a version of that toll for businesses. Thus, the artificial intelligence system can automatically tag them and make links to shopping pages, once they upload photos to their page containing their products.
Facebook’s access to users’ photos on the marketplace helps the social network build those tools further.
