Artificial intelligence (AI) is all around us, helping us live our daily lives. It involves numerous technologies and algorithms that are used to train systems to emulate human intelligence by perceiving, learning, reasoning, planning and solving problems. As the systems learn, they become more able to perform intellectual tasks by themselves, and appear to assume the human quality of intelligence.
For example, our choices of streaming music, films, and news are soon learned by the internet algorithms, which then suggestions what we might like to view or listen to next. Our social media streams are also filtered according to AI learning principles that have been used to get to know us better. Sound complicated? Well, researchers from the University of Iowa recently set out to demystify the way in which AI works, by testing how simple pigeons learn, and comparing this to the way in which AI learning takes place.
The researchers identified two types of learning: declarative and associative. Declarative learning is based on exercising reason according to a set of rules or strategies – a so-called higher order of learning that is typical in humans. In contrast, associative learning centers on recognizing and making connections between objects or patterns, such as in “sky-blue” and “water-wet.” It is considered a lower-order way of thinking. Numerous animal species use associative learning, but only a few – dolphins, chimpanzees and humans among them – are capable of declarative learning.
In order to tease out the way in which pigeons learn, the researchers set a “diabolically difficult” task for them. They gave the pigeons complex categorization tests that higher-level thinking, such as using logic or reasoning, would not aid in solving. Instead, the pigeons, had to use trial and error to memorize enough of the scenarios in the test in order to find the correct solution.
Each test pigeon was shown a stimulus and had to decide, by pecking a button on the right or on the left, to which category that stimulus belonged. They had to consider variables such as line width, line angle, concentric rings, and sectioned rings, in matching stimuli and solutions. A correct answer yielded a tasty pellet; an incorrect response yielded nothing. What made the test so demanding was its arbitrariness: no rules or logic would help the pigeons decipher the solutions.
“These stimuli are special. They don’t look like one another, and they’re never repeated,” says study co-author Ed Wasserman. “You have to memorize the individual stimuli or regions from where the stimuli occur in order to do the task,” said Wasserman, who has studied pigeon intelligence for five decades.
Each of the four test pigeons began by correctly answering about half the time. But over hundreds of tests, the birds eventually upped their score to an average of 68 percent correct. They had clearly learned how to respond to the tasks using an associative learning process.
The researchers equate the pigeons’ repetitive, trial-and-error approach to the learning that takes place by machines. Computers employ the same basic methodology, the researchers contend, and are “taught” how to identify the patterns and objects easily recognized by humans. Granted, computers, because of their enormous memory and storage power, far surpass anything the pigeon brain can conjure. However, the basic process of making associations is the same between the test-taking pigeons and the latest AI advances.
“You hear all the time about the wonders of AI, all the amazing things that it can do,” says Wasserman. “It can beat the pants off people playing chess, or at any video game, for that matter. It can beat us at all kinds of things. How does it do it? Is it smart? No, it’s using the same system or an equivalent system to what the pigeon is using here.”
“The pigeons are like AI masters,” Wasserman says. “They’re using a biological algorithm, the one that nature has given them, whereas the computer is using an artificial algorithm that humans gave them.”
The common denominator is that AI and pigeons both employ associative learning, and it is that base-level thinking that allowed the pigeons to ultimately score successfully. If people were to take the same test, Wasserman says, they’d score poorly and would probably give up because they are more used to using declarative learning in order to solve problems.
“The goal was to see to what extent a simple associative mechanism was capable of solving a task that would trouble us because people rely so heavily on rules or strategies,” Wasserman adds. “In this case, those rules would get in the way of learning. The pigeon never goes through that process. It doesn’t have that high-level thinking process. But it doesn’t get in the way of their learning. In fact, in some ways it facilitates it.”
Wasserman sees a paradox in how associative learning is viewed. “People are wowed by AI doing amazing things, using a learning algorithm much like the pigeon,” he says, “yet when people talk about associative learning in humans and animals, it is discounted as rigid and unsophisticated.”
Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.