Make no mistake, Google is not the only web service collecting and analyzing big data with the goal of “breaking” artificial intelligence. MIT’s Technology Review reported today that Facebook is working on what it calls “deep learning” (deep learning from big data).
The company has created a small research team for this purpose. The project and its work are secret, but it includes experimentation with deep learning, a technology that aims to enable machines to think, understand, and educate themselves, an education far beyond that of humans.
Deep learning replicates the layers of neurons in the neocortex—the part of the human brain that thinks. It’s the idea behind Siri, semantic search, IBM’s Watson, new medical breakthroughs, natural language recognition, self-driving cars, and robots as a workforce. Right now, it’s the best chance we have at making the scientific dream of artificial intelligence a modern reality.
For the past two years, Google has dominated the deep learning space, but it makes sense that Facebook is a worthy opponent in this game. The social network has a very important ace up its sleeve: The data from 1 billion users (300 million active) and the personal information from their lives that flows into its databases every day.
Big data and machine learning go hand in hand, since a “neural network” can teach itself by drawing connections and ideas from the information it has access to.
Our minds may be much more complex than a machine trying to imitate, but Google and Facebook have so much information that it is impossible for a human brain to capture and retain it.
[one_third]Vicarious is developing machine learning software based on the computational principles of the human brain. Our first technology is a visual perception system that interprets the contents of photographs and videos in a manner similar to humans. Powering this technology is a new computational paradigm we call the Recursive Cortical Network ™.[/one_third] A Russian venture capital firm relied on this fact when it decided to invest hundreds of millions of dollars in Facebook two years ago, believing that the social network is the future of artificial intelligence. Facebook when it invested in Dustin Moskovitz, Vicarious, a company that aims to replicate the human brain in order to create intelligent, thinking applications and, by extension, intelligent thinking machines.
What's more disturbing is that Facebook's virtual brain, built from users' personal data, could help the company see even deeper into your life. Technology Review reported that deep learning technology could understand the emotions of a user's posts or events, recognize objects in photos, and even predict how people are likely to behave in the future. That thought alone is scary enough.
To some extent, this is already happening, simply with traditional machine learning Facebook uses its data today to improve its user experience and develop new features. The news that has been “heard” recently is that Facebook knows characteristics and details about you that you have never publicly stated – like whether someone is gay or a Democrat.
The question is: Do we care that our personal data is being used to create an artificially intelligent brain that can learn even more about us?
