This week’s BEACON Researchers at Work post is by MSU graduate student Randal Olson.
For well over a decade, I have been fascinated with the idea that computers could achieve the same level of intelligence as humans. I would often ask my friends, “How cool would it be to combine human-level intelligence with the massive computing power of computers? A machine that could think like us, but infinitely faster… imagine the possibilities!” Usually my friends responded by rolling their eyes, or started talking about how that kind of Artificial Intelligence (AI) would kill us all. (Thanks to Terminator, and about every other popular movie that features an evil AI!) Since I have never been content to just think about what an Artificial Intelligence would be capable of, I set out on a long journey to figure out how I could help make a (not-so-evil) AI a reality in my lifetime.
Any good story begins with a prologue, so let’s start there.
Ever since computers were created, scientists have tried to figure out how a computer could achieve human-level intelligence. By the 1960′s, AI researchers had already discovered that the computer could perform many intelligent tasks that humans were capable of, at speeds that were previously considered impossible.
Indeed, the leading AI researchers at the time claimed that,
“Machines will be capable, within twenty years, of doing any work a man can do.”
“Within a generation … the problem of creating ‘artificial intelligence’ will substantially be solved.”
Unfortunately, creating an AI with human-level intelligence (called “strong AI”) inside a computer turned out to be a much harder problem than everyone thought, but that hasn’t deterred AI researchers. Since the 1960′s, AI researchers have made tremendous strides in creating intelligent machines whose abilities rival (and oftentimes outperform) those of humans. Today, we have AIs that defeat world chess champions in chess, act as electronic personal assistants (with an attitude!), drive cars without human intervention, and even help diagnose and predict health issues in patients.
Despite this remarkable progress, an AI with human-level intelligence has eluded AI researchers for over half a century. A growing number of researchers, particularly those with a neuroscience background, claim that strong AI will continue to elude AI researchers as long as they try to “hard-code” AI; to create a true general Artificial Intelligence, AI researchers must first understand the biological mechanisms that ubderly human intelligence.
This is where my story comes in.
Along with my colleagues in Dr. Chris Adami’s lab and the Evolved Intelligence research group at the BEACON HQ, I have been working on understanding how human-level intelligence evolved in nature using Artificial Life as a research platform. We are trying to learn about the problems that early creatures faced that made intelligent behavior a favorable trait to evolve, with the goal of eventually working our way up to understanding how early forms of intelligence evolved into the complex forms of adaptive, social, and predictive intelligence that humans are capable of. In essence, we are attempting to reproduce the evolutionary path to human-level intelligence in silico.
Lately, I have been studying the evolution of swarming behavior in animals. Swarming behavior is a particularly interesting intelligent behavior because it enables groups of animals to accomplish tasks that would otherwise be difficult or even impossible to accomplish alone. Take this group of European starlings under attack by a Peregrine falcon, for example.
If the starlings didn’t work together, they would make easy prey for the falcon. However, by swarming together, they are better able to defend themselves against the falcon’s attacks. In a recent paper, my colleagues and I addressed one of many possible reasons why the starlings evolved this swarming behavior.
Understanding grouping behavior and group decision making in animals is an important part of understanding the evolution of human-level intelligence. Humans are social creatures that evolved to work in groups, so understanding how and why animals work and live in groups will give us a better grasp of the interplay between grouping behavior and human-level intelligence. It could very well be that grouping behavior is a trait that creatures must have before they can evolve higher-level intelligence.
By following this path to strong AI, I hope to one day turn from SciFi into reality the intelligent machines that will transform our society for the better.
For more information about Randy’s work, you can contact him at rso at randalolson dot com.