Tag Archives: Brent Harrison

What robots and humans?

I have two robot news bits for this posting. The first probes the unease currently being expressed (pop culture movies, Stephen Hawking, the Cambridge Centre for Existential Risk, etc.) about robots and their increasing intelligence and increased use in all types of labour formerly and currently performed by humans. The second item is about a research project where ‘artificial agents’ (robots) are being taught human values with stories.

Human labour obsolete?

‘When machines can do any job, what will humans do?’ is the question being asked in a presentation by Rice University computer scientist, Moshe Vardi, for the American Association for the Advancement of Science (AAAS) annual meeting held in Washington, D.C. from Feb. 11 – 15, 2016.

Here’s more about Dr. Vardi’s provocative question from a Feb. 14, 2016 Rice University news release (also on EurekAlert),

Rice University computer scientist Moshe Vardi expects that within 30 years, machines will be capable of doing almost any job that a human can. In anticipation, he is asking his colleagues to consider the societal implications. Can the global economy adapt to greater than 50 percent unemployment? Will those out of work be content to live a life of leisure?

“We are approaching a time when machines will be able to outperform humans at almost any task,” Vardi said. “I believe that society needs to confront this question before it is upon us: If machines are capable of doing almost any work humans can do, what will humans do?”

Vardi addressed this issue Sunday [Feb. 14, 2016] in a presentation titled “Smart Robots and Their Impact on Society” at one of the world’s largest and most prestigious scientific meetings — the annual meeting of the American Association for the Advancement of Science in Washington, D.C.

“The question I want to put forward is, Does the technology we are developing ultimately benefit mankind?” Vardi said. He asked the question after presenting a body of evidence suggesting that the pace of advancement in the field of artificial intelligence (AI) is increasing, even as existing robotic and AI technologies are eliminating a growing number of middle-class jobs and thereby driving up income inequality.

Vardi, a member of both the National Academy of Engineering and the National Academy of Science, is a Distinguished Service Professor and the Karen Ostrum George Professor of Computational Engineering at Rice, where he also directs Rice’s Ken Kennedy Institute for Information Technology. Since 2008 he has served as the editor-in-chief of Communications of the ACM, the flagship publication of the Association for Computing Machinery (ACM), one of the world’s largest computational professional societies.

Vardi said some people believe that future advances in automation will ultimately benefit humans, just as automation has benefited society since the dawn of the industrial age.

“A typical answer is that if machines will do all our work, we will be free to pursue leisure activities,” Vardi said. But even if the world economic system could be restructured to enable billions of people to live lives of leisure, Vardi questioned whether it would benefit humanity.

“I do not find this a promising future, as I do not find the prospect of leisure-only life appealing. I believe that work is essential to human well-being,” he said.

“Humanity is about to face perhaps its greatest challenge ever, which is finding meaning in life after the end of ‘In the sweat of thy face shalt thou eat bread,’” Vardi said. “We need to rise to the occasion and meet this challenge” before human labor becomes obsolete, he said.

In addition to dual membership in the National Academies, Vardi is a Guggenheim fellow and a member of the American Academy of Arts and Sciences, the European Academy of Sciences and the Academia Europa. He is a fellow of the ACM, the American Association for Artificial Intelligence and the Institute for Electrical and Electronics Engineers (IEEE). His numerous honors include the Southeastern Universities Research Association’s 2013 Distinguished Scientist Award, the 2011 IEEE Computer Society Harry H. Goode Award, the 2008 ACM Presidential Award, the 2008 Blaise Pascal Medal for Computer Science by the European Academy of Sciences and the 2000 Goedel Prize for outstanding papers in the area of theoretical computer science.

Vardi joined Rice’s faculty in 1993. His research centers upon the application of logic to computer science, database systems, complexity theory, multi-agent systems and specification and verification of hardware and software. He is the author or co-author of more than 500 technical articles and of two books, “Reasoning About Knowledge” and “Finite Model Theory and Its Applications.”

In a Feb. 5, 2015 post, I rounded up a number of articles about our robot future. It provides a still useful overview of the thinking on the topic.

Teaching human values with stories

A Feb. 12, 2016 Georgia (US) Institute of Technology (Georgia Tech) news release (also on EurekAlert) describes the research,

The rapid pace of artificial intelligence (AI) has raised fears about whether robots could act unethically or soon choose to harm humans. Some are calling for bans on robotics research; others are calling for more research to understand how AI might be constrained. But how can robots learn ethical behavior if there is no “user manual” for being human?

Researchers Mark Riedl and Brent Harrison from the School of Interactive Computing at the Georgia Institute of Technology believe the answer lies in “Quixote” — to be unveiled at the AAAI [Association for the Advancement of Artificial Intelligence]-16 Conference in Phoenix, Ariz. (Feb. 12 – 17, 2016). Quixote teaches “value alignment” to robots by training them to read stories, learn acceptable sequences of events and understand successful ways to behave in human societies.

“The collected stories of different cultures teach children how to behave in socially acceptable ways with examples of proper and improper behavior in fables, novels and other literature,” says Riedl, associate professor and director of the Entertainment Intelligence Lab. “We believe story comprehension in robots can eliminate psychotic-appearing behavior and reinforce choices that won’t harm humans and still achieve the intended purpose.”

Quixote is a technique for aligning an AI’s goals with human values by placing rewards on socially appropriate behavior. It builds upon Riedl’s prior research — the Scheherazade system — which demonstrated how artificial intelligence can gather a correct sequence of actions by crowdsourcing story plots from the Internet.

Scheherazade learns what is a normal or “correct” plot graph. It then passes that data structure along to Quixote, which converts it into a “reward signal” that reinforces certain behaviors and punishes other behaviors during trial-and-error learning. In essence, Quixote learns that it will be rewarded whenever it acts like the protagonist in a story instead of randomly or like the antagonist.

For example, if a robot is tasked with picking up a prescription for a human as quickly as possible, the robot could a) rob the pharmacy, take the medicine, and run; b) interact politely with the pharmacists, or c) wait in line. Without value alignment and positive reinforcement, the robot would learn that robbing is the fastest and cheapest way to accomplish its task. With value alignment from Quixote, the robot would be rewarded for waiting patiently in line and paying for the prescription.

Riedl and Harrison demonstrate in their research how a value-aligned reward signal can be produced to uncover all possible steps in a given scenario, map them into a plot trajectory tree, which is then used by the robotic agent to make “plot choices” (akin to what humans might remember as a Choose-Your-Own-Adventure novel) and receive rewards or punishments based on its choice.

The Quixote technique is best for robots that have a limited purpose but need to interact with humans to achieve it, and it is a primitive first step toward general moral reasoning in AI, Riedl says.

“We believe that AI has to be enculturated to adopt the values of a particular society, and in doing so, it will strive to avoid unacceptable behavior,” he adds. “Giving robots the ability to read and understand our stories may be the most expedient means in the absence of a human user manual.”

So there you have it, some food for thought.