Tag Archives: explained
A few years back, DeepMind’s Demis Hassabis famously prophesized that AI and neuroscience will positively feed into each other in a “virtuous circle.” If realized, this would fundamentally expand our insight into intelligence, both machine and human.
We’ve already seen some proofs of concept, at least in the brain-to-AI direction. For example, memory replay, a biological mechanism that fortifies our memories during sleep, also boosted AI learning when abstractly appropriated into deep learning models. Reinforcement learning, loosely based on our motivation circuits, is now behind some of AI’s most powerful tools.
Hassabis is about to be proven right again.
Last week, two studies independently tapped into the power of ANNs to solve a 70-year-old neuroscience mystery: how does our visual system perceive reality?
The first, published in Cell, used generative networks to evolve DeepDream-like images that hyper-activate complex visual neurons in monkeys. These machine artworks are pure nightmare fuel to the human eye; but together, they revealed a fundamental “visual hieroglyph” that may form a basic rule for how we piece together visual stimuli to process sight into perception.
In the second study, a team used a deep ANN model—one thought to mimic biological vision—to synthesize new patterns tailored to control certain networks of visual neurons in the monkey brain. When directly shown to monkeys, the team found that the machine-generated artworks could reliably activate predicted populations of neurons. Future improved ANN models could allow even better control, giving neuroscientists a powerful noninvasive tool to study the brain. The work was published in Science.
The individual results, though fascinating, aren’t necessarily the point. Rather, they illustrate how scientists are now striving to complete the virtuous circle: tapping AI to probe natural intelligence. Vision is only the beginning—the tools can potentially be expanded into other sensory domains. And the more we understand about natural brains, the better we can engineer artificial ones.
It’s a “great example of leveraging artificial intelligence to study organic intelligence,” commented Dr. Roman Sandler at Kernel.co on Twitter.
ANNs and biological vision have quite the history.
In the late 1950s, the legendary neuroscientist duo David Hubel and Torsten Wiesel became some of the first to use mathematical equations to understand how neurons in the brain work together.
In a series of experiments—many using cats—the team carefully dissected the structure and function of the visual cortex. Using myriads of images, they revealed that vision is processed in a hierarchy: neurons in “earlier” brain regions, those closer to the eyes, tend to activate when they “see” simple patterns such as lines. As we move deeper into the brain, from the early V1 to a nub located slightly behind our ears, the IT cortex, neurons increasingly respond to more complex or abstract patterns, including faces, animals, and objects. The discovery led some scientists to call certain IT neurons “Jennifer Aniston cells,” which fire in response to pictures of the actress regardless of lighting, angle, or haircut. That is, IT neurons somehow extract visual information into the “gist” of things.
That’s not trivial. The complex neural connections that lead to increasing abstraction of what we see into what we think we see—what we perceive—is a central question in machine vision: how can we teach machines to transform numbers encoding stimuli into dots, lines, and angles that eventually form “perceptions” and “gists”? The answer could transform self-driving cars, facial recognition, and other computer vision applications as they learn to better generalize.
Hubel and Wiesel’s Nobel-prize-winning studies heavily influenced the birth of ANNs and deep learning. Much of earlier ANN “feed-forward” model structures are based on our visual system; even today, the idea of increasing layers of abstraction—for perception or reasoning—guide computer scientists to build AI that can better generalize. The early romance between vision and deep learning is perhaps the bond that kicked off our current AI revolution.
It only seems fair that AI would feed back into vision neuroscience.
Hieroglyphs and Controllers
In the Cell study, a team led by Dr. Margaret Livingstone at Harvard Medical School tapped into generative networks to unravel IT neurons’ complex visual alphabet.
Scientists have long known that neurons in earlier visual regions (V1) tend to fire in response to “grating patches” oriented in certain ways. Using a limited set of these patches like letters, V1 neurons can “express a visual sentence” and represent any image, said Dr. Arash Afraz at the National Institute of Health, who was not involved in the study.
But how IT neurons operate remained a mystery. Here, the team used a combination of genetic algorithms and deep generative networks to “evolve” computer art for every studied neuron. In seven monkeys, the team implanted electrodes into various parts of the visual IT region so that they could monitor the activity of a single neuron.
The team showed each monkey an initial set of 40 images. They then picked the top 10 images that stimulated the highest neural activity, and married them to 30 new images to “evolve” the next generation of images. After 250 generations, the technique, XDREAM, generated a slew of images that mashed up contorted face-like shapes with lines, gratings, and abstract shapes.
This image shows the evolution of an optimum image for stimulating a visual neuron in a monkey. Image Credit: Ponce, Xiao, and Schade et al. – Cell.
“The evolved images look quite counter-intuitive,” explained Afraz. Some clearly show detailed structures that resemble natural images, while others show complex structures that can’t be characterized by our puny human brains.
This figure shows natural images (right) and images evolved by neurons in the inferotemporal cortex of a monkey (left). Image Credit: Ponce, Xiao, and Schade et al. – Cell.
“What started to emerge during each experiment were pictures that were reminiscent of shapes in the world but were not actual objects in the world,” said study author Carlos Ponce. “We were seeing something that was more like the language cells use with each other.”
This image was evolved by a neuron in the inferotemporal cortex of a monkey using AI. Image Credit: Ponce, Xiao, and Schade et al. – Cell.
Although IT neurons don’t seem to use a simple letter alphabet, it does rely on a vast array of characters like hieroglyphs or Chinese characters, “each loaded with more information,” said Afraz.
The adaptive nature of XDREAM turns it into a powerful tool to probe the inner workings of our brains—particularly for revealing discrepancies between biology and models.
The Science study, led by Dr. James DiCarlo at MIT, takes a similar approach. Using ANNs to generate new patterns and images, the team was able to selectively predict and independently control neuron populations in a high-level visual region called V4.
“So far, what has been done with these models is predicting what the neural responses would be to other stimuli that they have not seen before,” said study author Dr. Pouya Bashivan. “The main difference here is that we are going one step further and using the models to drive the neurons into desired states.”
It suggests that our current ANN models for visual computation “implicitly capture a great deal of visual knowledge” which we can’t really describe, but which the brain uses to turn vision information into perception, the authors said. By testing AI-generated images on biological vision, however, the team concluded that today’s ANNs have a degree of understanding and generalization. The results could potentially help engineer even more accurate ANN models of biological vision, which in turn could feed back into machine vision.
“One thing is clear already: Improved ANN models … have led to control of a high-level neural population that was previously out of reach,” the authors said. “The results presented here have likely only scratched the surface of what is possible with such implemented characterizations of the brain’s neural networks.”
To Afraz, the power of AI here is to find cracks in human perception—both our computational models of sensory processes, as well as our evolved biological software itself. AI can be used “as a perfect adversarial tool to discover design cracks” of IT, said Afraz, such as finding computer art that “fools” a neuron into thinking the object is something else.
“As artificial intelligence researchers develop models that work as well as the brain does—or even better—we will still need to understand which networks are more likely to behave safely and further human goals,” said Ponce. “More efficient AI can be grounded by knowledge of how the brain works.”
Image Credit: Sangoiri / Shutterstock.com Continue reading
Spider-Man is one of the most popular superheroes of all time. It’s a bit surprising given that one of the more common phobias is arachnophobia—a debilitating fear of spiders.
Perhaps more fantastical is that young Peter Parker, a brainy high school science nerd, seemingly developed overnight the famous web-shooters and the synthetic spider silk that he uses to swing across the cityscape like Tarzan through the jungle.
That’s because scientists have been trying for decades to replicate spider silk, a material that is five times stronger than steel, among its many superpowers. In recent years, researchers have been untangling the protein-based fiber’s structure down to the molecular level, leading to new insights and new potential for eventual commercial uses.
The applications for such a material seem near endless. There’s the more futuristic visions, like enabling robotic “muscles” for human-like movement or ensnaring real-life villains with a Spider-Man-like web. Near-term applications could include the biomedical industry, such as bandages and adhesives, and as a replacement textile for everything from rope to seat belts to parachutes.
Spinning Synthetic Spider Silk
Randy Lewis has been studying the properties of spider silk and developing methods for producing it synthetically for more than three decades. In the 1990s, his research team was behind cloning the first spider silk gene, as well as the first to identify and sequence the proteins that make up the six different silks that web slingers make. Each has different mechanical properties.
“So our thought process was that you could take that information and begin to to understand what made them strong and what makes them stretchy, and why some are are very stretchy and some are not stretchy at all, and some are stronger and some are weaker,” explained Lewis, a biology professor at Utah State University and director of the Synthetic Spider Silk Lab, in an interview with Singularity Hub.
Spiders are naturally territorial and cannibalistic, so any intention to farm silk naturally would likely end in an orgy of arachnid violence. Instead, Lewis and company have genetically modified different organisms to produce spider silk synthetically, including inserting a couple of web-making genes into the genetic code of goats. The goats’ milk contains spider silk proteins.
The lab also produces synthetic spider silk through a fermentation process not entirely dissimilar to brewing beer, but using genetically modified bacteria to make the desired spider silk proteins. A similar technique has been used for years to make a key enzyme in cheese production. More recently, companies are using transgenic bacteria to make meat and milk proteins, entirely bypassing animals in the process.
The same fermentation technology is used by a chic startup called Bolt Threads outside of San Francisco that has raised more than $200 million for fashionable fibers made out of synthetic spider silk it calls Microsilk. (The company is also developing a second leather-like material, Mylo, using the underground root structure of mushrooms known as mycelium.)
Lewis’ lab also uses transgenic silkworms to produce a kind of composite material made up of the domesticated insect’s own silk proteins and those of spider silk. “Those have some fairly impressive properties,” Lewis said.
The researchers are even experimenting with genetically modified alfalfa. One of the big advantages there is that once the spider silk protein has been extracted, the remaining protein could be sold as livestock feed. “That would bring the cost of spider silk protein production down significantly,” Lewis said.
Building a Better Web
Producing synthetic spider silk isn’t the problem, according to Lewis, but the ability to do it at scale commercially remains a sticking point.
Another challenge is “weaving” the synthetic spider silk into usable products that can take advantage of the material’s marvelous properties.
“It is possible to make silk proteins synthetically, but it is very hard to assemble the individual proteins into a fiber or other material forms,” said Markus Buehler, head of the Department of Civil and Environmental Engineering at MIT, in an email to Singularity Hub. “The spider has a complex spinning duct in which silk proteins are exposed to physical forces, chemical gradients, the combination of which generates the assembly of molecules that leads to silk fibers.”
Buehler recently co-authored a paper in the journal Science Advances that found dragline spider silk exhibits different properties in response to changes in humidity that could eventually have applications in robotics.
Specifically, spider silk suddenly contracts and twists above a certain level of relative humidity, exerting enough force to “potentially be competitive with other materials being explored as actuators—devices that move to perform some activity such as controlling a valve,” according to a press release.
Studying Spider Silk Up Close
Recent studies at the molecular level are helping scientists learn more about the unique properties of spider silk, which may help researchers develop materials with extraordinary capabilities.
For example, scientists at Arizona State University used magnetic resonance tools and other instruments to image the abdomen of a black widow spider. They produced what they called the first molecular-level model of spider silk protein fiber formation, providing insights on the nanoparticle structure. The research was published last October in Proceedings of the National Academy of Sciences.
A cross section of the abdomen of a black widow (Latrodectus Hesperus) spider used in this study at Arizona State University. Image Credit: Samrat Amin.
Also in 2018, a study presented in Nature Communications described a sort of molecular clamp that binds the silk protein building blocks, which are called spidroins. The researchers observed for the first time that the clamp self-assembles in a two-step process, contributing to the extensibility, or stretchiness, of spider silk.
Another team put the spider silk of a brown recluse under an atomic force microscope, discovering that each strand, already 1,000 times thinner than a human hair, is made up of thousands of nanostrands. That helps explain its extraordinary tensile strength, though technique is also a factor, as the brown recluse uses a special looping method to reinforce its silk strands. The study also appeared last year in the journal ACS Macro Letters.
Making Spider Silk Stick
Buehler said his team is now trying to develop better and faster predictive methods to design silk proteins using artificial intelligence.
“These new methods allow us to generate new protein designs that do not naturally exist and which can be explored to optimize certain desirable properties like torsional actuation, strength, bioactivity—for example, tissue engineering—and others,” he said.
Meanwhile, Lewis’ lab has discovered a method that allows it to solubilize spider silk protein in what is essentially a water-based solution, eschewing acids or other toxic compounds that are normally used in the process.
That enables the researchers to develop materials beyond fiber, including adhesives that “are better than an awful lot of the current commercial adhesives,” Lewis said, as well as coatings that could be used to dampen vibrations, for example.
“We’re making gels for various kinds of of tissue regeneration, as well as drug delivery, and things like that,” he added. “So we’ve expanded the use profile from something beyond fibers to something that is a much more extensive portfolio of possible kinds of materials.”
And, yes, there’s even designs at the Synthetic Spider Silk Lab for developing a Spider-Man web-slinger material. The US Navy is interested in non-destructive ways of disabling an enemy vessel, such as fouling its propeller. The project also includes producing synthetic proteins from the hagfish, an eel-like critter that exudes a gelatinous slime when threatened.
Lewis said that while the potential for spider silk is certainly headline-grabbing, he cautioned that much of the hype is not focused on the unique mechanical properties that could lead to advances in healthcare and other industries.
“We want to see spider silk out there because it’s a unique material, not because it’s got marketing appeal,” he said.
Image Credit: mycteria / Shutterstock.com Continue reading
Dr. Been Kim wants to rip open the black box of deep learning.
A senior researcher at Google Brain, Kim specializes in a sort of AI psychology. Like cognitive psychologists before her, she develops various ways to probe the alien minds of artificial neural networks (ANNs), digging into their gory details to better understand the models and their responses to inputs.
The more interpretable ANNs are, the reasoning goes, the easier it is to reveal potential flaws in their reasoning. And if we understand when or why our systems choke, we’ll know when not to use them—a foundation for building responsible AI.
There are already several ways to tap into ANN reasoning, but Kim’s inspiration for unraveling the AI black box came from an entirely different field: cognitive psychology. The field aims to discover fundamental rules of how the human mind—essentially also a tantalizing black box—operates, Kim wrote with her colleagues.
In a new paper uploaded to the pre-publication server arXiv, the team described a way to essentially perform a human cognitive test on ANNs. The test probes how we automatically complete gaps in what we see, so that they form entire objects—for example, perceiving a circle from a bunch of loose dots arranged along a clock face. Psychologist dub this the “law of completion,” a highly influential idea that led to explanations of how our minds generalize data into concepts.
Because deep neural networks in machine vision loosely mimic the structure and connections of the visual cortex, the authors naturally asked: do ANNs also exhibit the law of completion? And what does that tell us about how an AI thinks?
Enter the Germans
The law of completion is part of a series of ideas from Gestalt psychology. Back in the 1920s, long before the advent of modern neuroscience, a group of German experimental psychologists asked: in this chaotic, flashy, unpredictable world, how do we piece together input in a way that leads to meaningful perceptions?
The result is a group of principles known together as the Gestalt effect: that the mind self-organizes to form a global whole. In the more famous words of Gestalt psychologist Kurt Koffka, our perception forms a whole that’s “something else than the sum of its parts.” Not greater than; just different.
Although the theory has its critics, subsequent studies in humans and animals suggest that the law of completion happens on both the cognitive and neuroanatomical level.
Take a look at the drawing below. You immediately “see” a shape that’s actually the negative: a triangle or a square (A and B). Or you further perceive a 3D ball (C), or a snake-like squiggle (D). Your mind fills in blank spots, so that the final perception is more than just the black shapes you’re explicitly given.
Image Credit: Wikimedia Commons contributors, the free media repository.
Neuroscientists now think that the effect comes from how our visual system processes information. Arranged in multiple layers and columns, lower-level neurons—those first to wrangle the data—tend to extract simpler features such as lines or angles. In Gestalt speak, they “see” the parts.
Then, layer by layer, perception becomes more abstract, until higher levels of the visual system directly interpret faces or objects—or things that don’t really exist. That is, the “whole” emerges.
The Experiment Setup
Inspired by these classical experiments, Kim and team developed a protocol to test the Gestalt effect on feed-forward ANNs: one simple, the other, dubbed the “Inception V3,” far more complex and widely used in the machine vision community.
The main idea is similar to the triangle drawings above. First, the team generated three datasets: one set shows complete, ordinary triangles. The second—the “Illusory” set, shows triangles with the edges removed but the corners intact. Thanks to the Gestalt effect, to us humans these generally still look like triangles. The third set also only shows incomplete triangle corners. But here, the corners are randomly rotated so that we can no longer imagine a line connecting them—hence, no more triangle.
To generate a dataset large enough to tease out small effects, the authors changed the background color, image rotation, and other aspects of the dataset. In all, they produced nearly 1,000 images to test their ANNs on.
“At a high level, we compare an ANN’s activation similarities between the three sets of stimuli,” the authors explained. The process is two steps: first, train the AI on complete triangles. Second, test them on the datasets. If the response is more similar between the illusory set and the complete triangle—rather than the randomly rotated set—it should suggest a sort of Gestalt closure effect in the network.
Right off the bat, the team got their answer: yes, ANNs do seem to exhibit the law of closure.
When trained on natural images, the networks better classified the illusory set as triangles than those with randomized connection weights or networks trained on white noise.
When the team dug into the “why,” things got more interesting. The ability to complete an image correlated with the network’s ability to generalize.
Humans subconsciously do this constantly: anything with a handle made out of ceramic, regardless of shape, could easily be a mug. ANNs still struggle to grasp common features—clues that immediately tells us “hey, that’s a mug!” But when they do, it sometimes allows the networks to better generalize.
“What we observe here is that a network that is able to generalize exhibits…more of the closure effect [emphasis theirs], hinting that the closure effect reflects something beyond simply learning features,” the team wrote.
What’s more, remarkably similar to the visual cortex, “higher” levels of the ANNs showed more of the closure effect than lower layers, and—perhaps unsurprisingly—the more layers a network had, the more it exhibited the closure effect.
As the networks learned, their ability to map out objects from fragments also improved. When the team messed around with the brightness and contrast of the images, the AI still learned to see the forest from the trees.
“Our findings suggest that neural networks trained with natural images do exhibit closure,” the team concluded.
That’s not to say that ANNs recapitulate the human brain. As Google’s Deep Dream, an effort to coax AIs into spilling what they’re perceiving, clearly demonstrates, machine vision sees some truly weird stuff.
In contrast, because they’re modeled after the human visual cortex, perhaps it’s not all that surprising that these networks also exhibit higher-level properties inherent to how we process information.
But to Kim and her colleagues, that’s exactly the point.
“The field of psychology has developed useful tools and insights to study human brains– tools that we may be able to borrow to analyze artificial neural networks,” they wrote.
By tweaking these tools to better analyze machine minds, the authors were able to gain insight on how similarly or differently they see the world from us. And that’s the crux: the point isn’t to say that ANNs perceive the world sort of, kind of, maybe similar to humans. It’s to tap into a wealth of cognitive psychology tools, established over decades using human minds, to probe that of ANNs.
“The work here is just one step along a much longer path,” the authors conclude.
“Understanding where humans and neural networks differ will be helpful for research on interpretability by enlightening the fundamental differences between the two interesting species.”
Image Credit: Popova Alena / Shutterstock.com Continue reading