Twitter has acquired a stealthy computer vision startup called Madbits, which was founded by former New York University researchers. Clément Farabet and Louis-Alexandre Etezad-Heydari. Farabet is a protégé of Facebook AI Lab director and New York University professor Yann LeCun, while Etezad-Heydari was advised by Larry Maloney and Eero Simoncelli.
The notice on the Madbits website, which is dated Monday, is light on detail about the company’s technology, but includes this brief explanation of what it was doing and why it decided to accept Twitter’s offer:
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.
We prototyped and tested about ten different applications, and as we’ve prepared to launch publicly, we’ve decided to bring the technology to Twitter, a company that shares our ambitions and vision and will help us scale this technology.
It’s not at all surprising that Twitter would want to acquire the company, given the tremendous amount of images published on Twitter every day. If Twitter wants to roll out functions such as image search, improve its search rankings based on image content, or perhaps even analyze images to get a better sense of what people are tweeting about, it will need people who understand how to do it. Yahoo, Dropbox and Pinterest have also made acquisitions in the computer vision space likely to glean the same types of capabilities.
Deep learning is a form of artificial intelligence that has been around for decades but has made major strides in the past several years with regard to accuracy in areas such as computer vision, text analysis and speech recognition. Many of the researchers responsible for those advances are currently employed, at least part-time, by larger web companies such as Google, Facebook, Microsoft and Baidu.
Details of the acquisition weren’t announced, but one can assume Madbits was able to demand a not-insignificant amount of money. Deep learning is such a hot technology, and good talent is so rare, that Microsoft Research director Peter Lee told Businessweek, “Last year, the cost of a top, world-class deep learning expert was about the same as a top NFL quarterback prospect.” Earlier this year, Google paid $400 million for another deep learning startup called DeepMind.
If they took a stock deal, that might work out well: Twitter’s earnings Tuesday were a surprise hit, and investors are likely to send its stock price higher tomorrow morning when trading opens.
Update: This post was updated at 9:45 a.m. on July 30 to correctly identify Larry Maloney and Eero Simoncelli as Etezad-Heydari’s advisers.