September 19, 2016

Need Some AI? Yeah, There’s a Marketplace for That

Diego Oppenheimer is worried that the Googles and the Facebooks will dominate the world of artificial intelligence.

It’s a legitimate worry. Elon Musk and Sam Altman are worried about the same thing. That’s why they created a startup called OpenAI. In recent years, Google and Facebook have snapped up so many researchers at the heart of the deep learning movement, an AI movement that’s rapidly reinventing everything from speech recognition to security. So, Musk and Altman grabbed several top AI researchers from Google and Facebook and vowed to share their work with the world at large.

Now, Oppenheimer and his startup, Algorithmia, are doing their part in the battle against AI hegemony. Algorithmia is what Oppenheimer calls an open marketplace for algorithms—code that companies and developers can use to beef up their websites and apps—and this marketplace now includes deep learning algorithms that handle tasks like face recognition and character recognition. Whereas OpenAI shares raw AI research, Algorithmia offers working algorithms designed to slot right into new services. “Maybe the future was already invented,” Oppenheimer says. “It’s just stuck in academic papers.”

It’s part of a much larger effort to democratize AI. Startups like Clarifai, Nara Logics, and MetaMind (now owned by also offer tools for building deep learning into any application. And giants like Google, Amazon, and Microsoft are building cloud computing services that work in similar ways.

Algorithmia uses the same marketplace model that startups have applied to so many other goods and services, including everything from artisanal crafts and graphic design to real estate and good ol’ retail. Anyone can upload an algorithm to the market, and then anyone can pay to use it. The author of the algorithm chooses the price of each API call (free to one dollar), and then 70 percent of the proceeds go to the author, with Algorithmia keeping the rest. When an algorithm is completely open source, Algorithmia charges only for the computing power used.

What’s new are the deep learning algorithms. Today, Algorithmia offers just over 2,500 algorithms, sixteen of which are deep learning models added by coders in July, at the urging of Oppenheimer and company. Loosely based on the networks of neurons in the human brain, deep neural nets can learn discrete tasks by analyzing enormous amounts of data. Oppenheimer points to the Colorize algorithm—which uses deep learning to automatically colorize black and white photos—as a prime example. Colorize was built the UC Berkeley Vision Lab, as part of a graduate student’s academic paper, and the Algorithmia team urged him to upload the algorithm to their marketplace.

Oren Etzioni, CEO of the Allen Institute for Artificial Intelligence and an investor in Algorithmia, believes these algorithm can help transform deep learning from a “dark art” practiced only by the giants of the Internet into a technology used by “the 99 percent.” Dennis R. Mortensen, the CEO and founder of the artificial intelligence startup, agrees—up to a point. He adds that these algorithms serve only some needs. As he points out, other companies are providing more complex tools for building deep learning systems, including Google, which recently open sourced a deep learning software engine called TensorFlow. He also points out that deep learning requires enormous amounts of data for training. Just because you have the algorithm, he says, that doesn’t mean you’ll get useful results. Algorithmia won’t help build the most advanced AI, he says, but it will help you layer a little AI into your app.