The Center for Data Innovation spoke with Derik Pridmore, president of Osaro, a machine learning startup based in San Francisco. Pridmore discussed how machines can use deep reinforcement learning to solve new problems and what a factory using learning machines can accomplish.
Osaro's machine intelligence software combines perception (which we've seen plenty of in the past, including with image identification) with decision-making abilities that will ostensibly help computer and robotic systems teach themselves to act more efficiently through trial and error (which we've seen more infrequently).
In an echo of DeepMind, Osaro has built an AI engine that can play classic games. But the company’s ultimate aim is to offer this technology as a way of driving the next generation of robots used in warehouses and factories. Much like humans, it gets better through practice. “Think about kids. They learn a lot through trial and error,” says Osaro founder and CEO Itamar Arel. “They come to understand what maximizes pleasure and minimizes pain.”
In the small world of AI startups, an approach called deep learning is currently in vogue. The method involves training neural nets on large quantities of data, such as photos, and then getting them to make inferences on new data. Osaro blends deep learning with reinforcement learning, a technique that generally entails teaching machines through trial and error. The startup’s technology “allows human mentors to teach machines how to autonomously take actions to achieve high-level goals,” according to a statement.
SAN FRANCISCO -- December 2, 2015 -- Osaro, a startup developing advanced machine learning technology known as deep reinforcement learning, announced a $3.3M seed round of funding to bring next-generation artificial intelligence (AI) products to the marketplace. The funding comes from an impressive slate of technology investors, including Peter Thiel, Scott Banister and Jerry Yang’s AME Cloud Ventures.
The Osaro team will attend NIPS 2015 in Montreal. We're sponsoring a lunch during the deep reinforcement learning workshop and will give an overview of the company. Details can be found here: http://rll.berkeley.edu/deeprlworkshop/