Paul Allen wants to make AI more sensible by teaching them common sense
Microsoft co-founder, Paul Allen, who had founded the Allen Institute for Artificial Intelligence (AI2) in 2014 on Wednesday pledged an additional $125 million into his Seattle-based nonprofit research lab to teach “artificial intelligence” some common sense.
For those unaware, AI2 is a research institute funded by Allen to achieve scientific breakthroughs by constructing AI (artificial intelligence) systems with reasoning, learning, and reading capabilities. It is led by Dr. Oren Etzioni, a former Professor at the University of Washington and CEO at Allen Institute.
The additional donation of $125 million from Allen is to help fund existing research as well as AI2’s new effort, Project Alexandria, whose goal is to create a system infused with good sense and judgement.
“When I founded AI2, I wanted to expand the capabilities of artificial intelligence through high-impact research,” said Allen. “Early in AI research, there was a great deal of focus on common sense, but that work stalled. AI still lacks what most 10-year-olds possess: ordinary common sense. We want to jump start that research to achieve major breakthroughs in the field.”
Currently, AI “recognizes objects, but can’t explain what it sees,” says Etzioni. “It can’t read a textbook and understand the questions in the back of the book. It is devoid of common sense.”
While AI has radically advanced in the past decade, there are still many things it is incapable of doing, such as unstructured problem-solving or managing unanticipated situations, which can be done by humans.
“This is an extremely complicated challenge,” said Allen. “If we want AI to approach human abilities and have the broadest possible impact in research, medicine and business, we need to fundamentally advance AI’s common sense abilities.”
Named after the ancient library in Alexandria, Project Alexandria aims to teach AI to answer questions such as, “How can you tell if a milk carton is full?” or “If I put my socks in the drawer, will they still be there tomorrow?”
Project Alexandria will help AI learn things like how to solve unstructured problems and manage unanticipated situations, which will provide a foundation for future AI systems to build on. It will use current artificial intelligence methods to aggregate data from images and text, via machine learning, and develop novel crowd-sourcing methods to “acquire common sense knowledge from people at an unprecedented scale”. It will develop applications that will make use of common sense to improve performance for a broad range of practical AI challenges, from machine reading to robotic vision. The goal is to create a “massive repository of knowledge”, as stated by Dr. Etzioni.
“Common sense is the precondition for general intelligence; until we get there we will be stuck with narrow AI that is rarely robust and never as flexible as human reasoners,” said Gary Marcus, founder of Geometric Intelligence (acquired by Uber) and professor of psychology and neural science at NYU. “I am hugely excited about Project Alexandria. There’s only been one serious large-scale effort to endow machines with common sense, and it was launched over three decades ago. The time is right for a fresh approach to the problem.”
For more information on AI2’s Project Alexandria, please see: https://allenai.org/alexandria/