Image-processing system learns largely on its own, much like a human baby
Neuroscience and artificial intelligence experts from Rice University and Baylor College of Medicine have taken inspiration from the human brain in creating a new “deep learning” method that enables computers to learn about the visual world largely on their own, much as human babies do.
In tests, the group’s “deep rendering mixture model” largely taught itself how to distinguish handwritten digits using a standard dataset of 10,000 digits written by federal employees and high school students. In results presented this month at the Neural Information Processing Systems (NIPS) conference in Barcelona, Spain, the researchers described how they trained their algorithm by giving it just 10 correct examples of each handwritten digit between zero and nine and then presenting it with several thousand more examples that it used to further teach itself. In tests, the algorithm was more accurate at correctly distinguishing handwritten digits than almost all previous algorithms that were trained with thousands of correct examples of each digit.
“In deep-learning parlance, our system uses a method known as semisupervised learning,” said lead researcher Ankit Patel, an assistant professor with joint appointments in neuroscience at Baylor and electrical and computer engineering at Rice. “The most successful efforts in this area have used a different technique called supervised learning, where the machine is trained with thousands of examples: This is a one. This is a two.
“Humans don’t learn that way,” Patel said. “When babies learn to see during their first year, they get very little input about what things are. Parents may label a few things: ‘Bottle. Chair. Momma.’ But the baby can’t even understand spoken words at that point. It’s learning mostly unsupervised via some interaction with the world.”
Patel said he and graduate student Tan Nguyen, a co-author on the new study, set out to design a semisupervised learning system for visual data that didn’t require much “hand-holding” in the form of training examples. For instance, neural networks that use supervised learning would typically be given hundreds or even thousands of training examples of handwritten digits before they would be tested on the database of 10,000 handwritten digits in the Mixed National Institute of Standards and Technology (MNIST) database.
The semisupervised Rice-Baylor algorithm is a “convolutional neural network,” a piece of software made up of layers of artificial neurons whose design was inspired by biological neurons. These artificial neurons, or processing units, are organized in layers, and the first layer scans an image and does simple tasks like searching for edges and color changes. The second layer examines the output from the first layer and searches for more complex patterns. Mathematically, this nested method of looking for patterns within patterns within patterns is referred to as a nonlinear process.
“It’s essentially a very simple visual cortex,” Patel said of the convolutional neural net. “You give it an image, and each layer processes the image a little bit more and understands it in a deeper way, and by the last layer, you’ve got a really deep and abstract understanding of the image. Every self-driving car right now has convolutional neural nets in it because they are currently the best for vision.”
Like human brains, neural networks start out as blank slates and become fully formed as they interact with the world. For example, each processing unit in a convolutional net starts the same and becomes specialized over time as they are exposed to visual stimuli.
“Edges are very important,” Nguyen said. “Many of the lower layer neurons tend to become edge detectors. They’re looking for patterns that are both very common and very important for visual interpretation, and each one trains itself to look for a specific pattern, like a 45-degree edge or a 30-degree red-to-blue transition.
“When they detect their particular pattern, they become excited and pass that on to the next layer up, which looks for patterns in their patterns, and so on,” he said. “The number of times you do a nonlinear transformation is essentially the depth of the network, and depth governs power. The deeper a network is, the more stuff it’s able to disentangle. At the deeper layers, units are looking for very abstract things like eyeballs or vertical grating patterns or a school bus.”
Nguyen began working with Patel in January as the latter began his tenure-track academic career at Rice and Baylor. Patel had already spent more than a decade studying and applying machine learning in jobs ranging from high-volume commodities training to strategic missile defense, and he’d just wrapped up a four-year postdoctoral stint in the lab of Rice’s Richard Baraniuk, another co-author on the new study. In late 2015, Baraniuk, Patel and Nguyen published the first theoretical framework that could both derive the exact structure of convolutional neural networks and provide principled solutions to alleviate some of their limitations.
Baraniuk said a solid theoretical understanding is vital for designing convolutional nets that go beyond today’s state-of-the-art.
“Understanding video images is a great example,” Baraniuk said. “If I am looking at a video, frame by frame by frame, and I want to understand all the objects and how they’re moving and so on, that is a huge challenge. Imagine how long it would take to label every object in every frame of a video. No one has time for that. And in order for a machine to understand what it’s seeing in a video, it has to understand what objects are, the concept of three-dimensional space and a whole bunch of other really complicated stuff. We humans learn those things on our own and take them for granted, but they are totally missing in today’s artificial neural networks.”
Patel said the theory of artificial neural networks, which was refined in the NIPS paper, could ultimately help neuroscientists better understand the workings of the human brain.
“There seem to be some similarities about how the visual cortex represents the world and how convolutional nets represent the world, but they also differ greatly,” Patel said. “What the brain is doing may be related, but it’s still very different. And the key thing we know about the brain is that it mostly learns unsupervised.
“What I and my neuroscientist colleagues are trying to figure out is, What is the semisupervised learning algorithm that’s being implemented by the neural circuits in the visual cortex? and How is that related to our theory of deep learning?” he said. “Can we use our theory to help elucidate what the brain is doing? Because the way the brain is doing it is far superior to any neural network that we’ve designed.”
Chartered in 1845 by the Republic of Texas, Baylor is the oldest continuously-operating university in Texas and was one of the first educational institutions west of the Mississippi River. The university’s 1,000-acre campus is located on the banks of the Brazos River next to freeway I-35, between the Dallas-Fort Worth Metroplex and Austin. Baylor University is accredited by the Southern Association of Colleges and Schools. Baylor is notable for its law, business, science, music and English programs.
Baylor University research articles from Innovation Toronto
When individuals engage in risky business transactions with each other, they may end up being disappointed.
This is why they’d rather leave the decision on how to divvy up jointly-owned monies to a computer than to their business partner. This subconscious strategy seems to help them avoid the negative emotions associated with any breaches of trust. This is the result of a study by scientists from the University of Bonn and US peers. They are presenting their findings in the scientific journal “Proceedings of the Royal Society B.”
Trust is an essential basis for business relationships. However, this basis can be shaken if one business partner exhibits dishonest behavior. “Everyone knows that trust can be shattered in risky businesses,” explained Prof. Dr. Bernd Weber from the Center for Economics and Neuroscience (CENs) at the University of Bonn. “As a result, people are not all that eager to put their trust in others.” Scientists call this attitude “betrayal aversion” – people try to avoid being disappointed by potential breaches of trust.
In a current study, Prof. Weber and his US colleagues, Prof. Dr. Jason A. Aimone from Baylor University and Prof. Dr. Daniel Houser from George Mason University examined in experiments the effects betrayal aversion has on simple financial decisions. A total of 30 subjects played a computer game at George Mason University in Arlington, VA (USA) that promised real monies to the winners. At the Life & Brain Zentrum of the University of Bonn, the same number of subjects then made their decisions based on the results of the earlier experiment. And while the Bonn subjects were responding to their gaming partners’ decisions made earlier in Arlington, their brain activity was measured by means of MRI scans.
Sharing fairly or making a profit at the other person’s expense?
In this experiment, the test subjects in Bonn were able to select whether they and their US partners would get one euro each only, or whether they wanted to have a higher amount – i.e., 6 euros – divided up. However, the latter variant came with a risk. So, for example, the other player might get as much as 5.60 euros while the Bonn player would be left with only 40 Cents. The actual dividing of the amount, which came in a second step, could be left either to one’s partner or to the computer. However, the computer gave out exactly the same decisions as the real test subjects. “So, from the point of view of winnings, there was no difference whether the other player or the machine divided the amount,” explained Prof. Weber. “And the subjects had explicitly been told so from the very start.”
Even though the winnings were exactly the same in the end, more subjects put their trust into the computer. When the money was divided by the computer, 63 percent of subjects trusted the process and only 37 percent preferred taking just the one euro. But if the arrangement was that the human partners would make the decision, only 49 percent of test subjects trusted them – 51 percent would rather take the more secure, small amount. “These results show that more subjects prefer to leave risky decisions in which they may be betrayed to an impersonal device, thus avoiding the negative feeling that comes from having wrongly trusted a human,” said Prof. Weber, adding that obviously a breach of trust committed by an impersonal computer was less emotionally stressful than if had been a private business partner.
The brain’s frontal insula was especially active
The University of Bonn’s subjects also showed interesting brain activities as measured in MRI scans. In the process of making financial decisions, the frontal insula was especially active when it was another player who made the decision on how to divide the amount. “This area of the brain is always involved when negative emotions such as pain, disappointment or fear are activated,” explained Prof. Weber. He added that the fact that the frontal insula was activated is a clear indication that negative emotions played an important role in these situations.
Financial decisions are very complex. “This is a very contrary phenomenon. Many studies show that the anonymity of business partners on the Internet results in a loss of trust,” said Prof. Weber. “But our results indicate that this anonymity can also help avoid negative feelings.” He added that these decision processes in financial transactions would yet have to be studied in more detail.