Google's DeepMind and Unity Tech Partner for 'Real-World' AI -- Pure AI

Google's DeepMind and Unity Tech Partner for 'Real-World' AI

By Richard Seeley
10/05/2018

Providing a more real-world testing environment for developing and testing machine learning (ML) applications is the goal of a recently announced partnership between Unity Technologies and DeepMind, Google's London-based artificial intelligence (AI) research company.

"The partnership will enable DeepMind to develop virtual environments and tasks in support of their fundamental AI research program," according to a recent blog written by Danny Lange, vice president of AI and machine learning at San Francisco-based Unity Technologies.

DeepMind proclaims that it is "on a scientific mission to push the boundaries of AI, developing programs that can learn to solve any complex problem without needing to be taught how."

The company, which is part of Google's Alphabet group, is setting ambitious goals including AI-powered breakthroughs ranging from climate change to healthcare. Working with the U.K. National Health Service, DeepMind is looking for ways to use AI and ML to speed analysis of patent test data to improve treatment in life threatening emergencies.

"DeepMind researchers are addressing huge AI problems," writes Lange, "and they have selected Unity as a primary research platform for creating complex virtual environments that will enable the development of algorithms capable of learning to solve complex tasks."

He notes that the Unity platform allows AI and ML researchers "to create and leverage simulation environments that are rich in sensory and physical complexity, provide compelling cognitive challenges, and support dynamic multi-agent interaction. These environments provide the foundation to accelerate AI research in areas such as computer vision, robotics, natural language instruction, autonomous vehicle development, and many other areas of science and technology."

A recently published reference paper details the importance of developing and testing AI and ML applications in simulated environments that reflect more accurately the real world where problems such as climate change await solutions.

"Recent advances in Deep Reinforcement Learning and Robotics have been driven by the presence of increasingly realistic and complex simulation environments," the paper states. "Many of the existing platforms, however, provide either unrealistic visuals, inaccurate physics, low task complexity, or a limited capacity for interaction among artificial agents."

The Unity ML-Agents Toolkit enables the development of "learning environments which are rich in sensory and physical complexity, provide compelling cognitive challenges, and support dynamic multi-agent interaction," according to Lange.