Facebook wants to use artificial intelligence for analysis and predictions

image: bgolub/Flickr

Madison Ruppert
Activist Post

Facebook is reportedly now conducting research on “deep learning” artificial intelligence technology to allow better analysis of the data captured from the social network’s 1 billion users, though details of the experiments remain secret.

This new project is just one of many ways Facebook is on the cutting edge of technology, including a massive facial recognition database. Facebook has also most recently come under fire for privacy violations and their role in NSA spying, especially in Europe.

Ultimately, the goal of the project is to predict your actions online, show you content that is more relevant to your interests and target advertisements more accurately, according to a report published by the MIT Technology Review.

Technology Review reports that the research has only recently begun and the actual experiments being conducted behind the scenes remain secret.

The project will reportedly run by an eight-person group known only as the “AI team,” and will attempt to use deep learning to uncover the true meaning behind what people say and share on Facebook.

In other contexts, deep learning has shown potential in giving software the ability to recognize emotions behind text or events described in text, even if they’re not referred to explicitly.

Deep learning has also shown the ability to analyze photos and recognize objects and make complex predictions about the likely future behavior of individuals.

One of the likely applications of the technology in the case of Facebook will be in improving their news feed by only showing what the social network deems most relevant to the users. The technology could also help people organize photographs or help choose the best one to post online.

Google has also been delving into deep learning with Ray Kurzweil and has used the technology to recognize cats in videos without the software being told that cats exist.

The Verge points out that deep learning relies on using “a multi-layered approach to data, parsing information to build up a body of knowledge that can then be used to figure out concepts, or understand what objects sound and look like.”

Machine learning – though not nearly as complex as the type being researched by Facebook and others – is also used in facial recognition.

However, applying deep learning can offer significant improvements over the more established techniques.

“In tasks like vision or speech, we’re seeing 30 percent-plus improvements with deep learning,” Elliot Turner, founder and CEO of AlchemyAPI, said to MIT Technology Review. (See here for follow up.)

The team is reportedly going to involve deep learning expert MarcAurelio Ranzato, who they hired away from Google, and Yaniv Taigman, the co-founder of facial recognition company Face.com. Others include vision expert Lubomir Bourdev and long-time Facebook engineer Keith Adams, according to the Technology Review.

Facebook will reportedly use some of their research, in more general machine learning areas, public. For now, all of their experiments remain secret.

I’d love to hear your opinion, take a look at your story tips and even your original writing if you would like to get it published. I am also available for interviews on radio, television or any other format. Please email me at [email protected]

Please support our work and help us start to pay contributors by doing your shopping through our Amazon link or check out some must-have products at our store.

This article first appeared at End the Lie.