Twitter joins deep learning trend, acquires Madbits

It looks like the deep learning trend has finally reached Twitter.

This past Monday, Twitter acquired Madbits, a deep learning startup that specializes in visual intelligence technology, according to the Madbits website.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":1516632,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"C"}']

Deep learning, a type of artificial intelligence that essentially allows computers to make inferences after being trained on lots of data, has become quite popular in the past few years. Deep learning technologies like text analysis, computer vision, and speech recognition can identify patterns and traits in music, images, text, and more — kind of like a person could. Major companies like Facebook, Google, Microsoft, and Baidu have already begun their own adventures in deep learning, employing a multitude of researchers responsible for advances in deep-learning technology. (See VentureBeat’s extensive interview with Baidu deep-learning head, Andrew Ng, here.)

The Madbits announcement doesn’t provide much detail, but it underscores Madbits’ focus on image-based learning and the company’s excitement to bring Twitter into the picture:

“Over this past year, we’ve built visual intelligence technology that automatically understands, organizes and extracts relevant information from raw media. Understanding the content of an image, whether or not there are tags associated with that image, is a complex challenge. We developed our technology based on deep learning, an approach to statistical machine learning that involves stacking simple projections to form powerful hierarchical models of a signal.”

It makes sense that Twitter would acquire a company like Madbits, given that Twitter users post thousands of images per day that it could analyze to get more information about what’s trending. Madbits could certainly help Twitter improve a variety of image search functions, such as the optimization of searches based on image content.