Tag Archives: university

#434643 Sensors and Machine Learning Are Giving ...

According to some scientists, humans really do have a sixth sense. There’s nothing supernatural about it: the sense of proprioception tells you about the relative positions of your limbs and the rest of your body. Close your eyes, block out all sound, and you can still use this internal “map” of your external body to locate your muscles and body parts – you have an innate sense of the distances between them, and the perception of how they’re moving, above and beyond your sense of touch.

This sense is invaluable for allowing us to coordinate our movements. In humans, the brain integrates senses including touch, heat, and the tension in muscle spindles to allow us to build up this map.

Replicating this complex sense has posed a great challenge for roboticists. We can imagine simulating the sense of sight with cameras, sound with microphones, or touch with pressure-pads. Robots with chemical sensors could be far more accurate than us in smell and taste, but building in proprioception, the robot’s sense of itself and its body, is far more difficult, and is a large part of why humanoid robots are so tricky to get right.

Simultaneous localization and mapping (SLAM) software allows robots to use their own senses to build up a picture of their surroundings and environment, but they’d need a keen sense of the position of their own bodies to interact with it. If something unexpected happens, or in dark environments where primary senses are not available, robots can struggle to keep track of their own position and orientation. For human-robot interaction, wearable robotics, and delicate applications like surgery, tiny differences can be extremely important.

Piecemeal Solutions
In the case of hard robotics, this is generally solved by using a series of strain and pressure sensors in each joint, which allow the robot to determine how its limbs are positioned. That works fine for rigid robots with a limited number of joints, but for softer, more flexible robots, this information is limited. Roboticists are faced with a dilemma: a vast, complex array of sensors for every degree of freedom in the robot’s movement, or limited skill in proprioception?

New techniques, often involving new arrays of sensory material and machine-learning algorithms to fill in the gaps, are starting to tackle this problem. Take the work of Thomas George Thuruthel and colleagues in Pisa and San Diego, who draw inspiration from the proprioception of humans. In a new paper in Science Robotics, they describe the use of soft sensors distributed through a robotic finger at random. This placement is much like the constant adaptation of sensors in humans and animals, rather than relying on feedback from a limited number of positions.

The sensors allow the soft robot to react to touch and pressure in many different locations, forming a map of itself as it contorts into complicated positions. The machine-learning algorithm serves to interpret the signals from the randomly-distributed sensors: as the finger moves around, it’s observed by a motion capture system. After training the robot’s neural network, it can associate the feedback from the sensors with the position of the finger detected in the motion-capture system, which can then be discarded. The robot observes its own motions to understand the shapes that its soft body can take, and translate them into the language of these soft sensors.

“The advantages of our approach are the ability to predict complex motions and forces that the soft robot experiences (which is difficult with traditional methods) and the fact that it can be applied to multiple types of actuators and sensors,” said Michael Tolley of the University of California San Diego. “Our method also includes redundant sensors, which improves the overall robustness of our predictions.”

The use of machine learning lets the roboticists come up with a reliable model for this complex, non-linear system of motions for the actuators, something difficult to do by directly calculating the expected motion of the soft-bot. It also resembles the human system of proprioception, built on redundant sensors that change and shift in position as we age.

In Search of a Perfect Arm
Another approach to training robots in using their bodies comes from Robert Kwiatkowski and Hod Lipson of Columbia University in New York. In their paper “Task-agnostic self-modeling machines,” also recently published in Science Robotics, they describe a new type of robotic arm.

Robotic arms and hands are getting increasingly dexterous, but training them to grasp a large array of objects and perform many different tasks can be an arduous process. It’s also an extremely valuable skill to get right: Amazon is highly interested in the perfect robot arm. Google hooked together an array of over a dozen robot arms so that they could share information about grasping new objects, in part to cut down on training time.

Individually training a robot arm to perform every individual task takes time and reduces the adaptability of your robot: either you need an ML algorithm with a huge dataset of experiences, or, even worse, you need to hard-code thousands of different motions. Kwiatkowski and Lipson attempt to overcome this by developing a robotic system that has a “strong sense of self”: a model of its own size, shape, and motions.

They do this using deep machine learning. The robot begins with no prior knowledge of its own shape or the underlying physics of its motion. It then repeats a series of a thousand random trajectories, recording the motion of its arm. Kwiatkowski and Lipson compare this to a baby in the first year of life observing the motions of its own hands and limbs, fascinated by picking up and manipulating objects.

Again, once the robot has trained itself to interpret these signals and build up a robust model of its own body, it’s ready for the next stage. Using that deep-learning algorithm, the researchers then ask the robot to design strategies to accomplish simple pick-up and place and handwriting tasks. Rather than laboriously and narrowly training itself for each individual task, limiting its abilities to a very narrow set of circumstances, the robot can now strategize how to use its arm for a much wider range of situations, with no additional task-specific training.

Damage Control
In a further experiment, the researchers replaced part of the arm with a “deformed” component, intended to simulate what might happen if the robot was damaged. The robot can then detect that something’s up and “reconfigure” itself, reconstructing its self-model by going through the training exercises once again; it was then able to perform the same tasks with only a small reduction in accuracy.

Machine learning techniques are opening up the field of robotics in ways we’ve never seen before. Combining them with our understanding of how humans and other animals are able to sense and interact with the world around us is bringing robotics closer and closer to becoming truly flexible and adaptable, and, eventually, omnipresent.

But before they can get out and shape the world, as these studies show, they will need to understand themselves.

Image Credit: jumbojan / Shutterstock.com Continue reading

Posted in Human Robots

#434616 What Games Are Humans Still Better at ...

Artificial intelligence (AI) systems’ rapid advances are continually crossing rows off the list of things humans do better than our computer compatriots.

AI has bested us at board games like chess and Go, and set astronomically high scores in classic computer games like Ms. Pacman. More complex games form part of AI’s next frontier.

While a team of AI bots developed by OpenAI, known as the OpenAI Five, ultimately lost to a team of professional players last year, they have since been running rampant against human opponents in Dota 2. Not to be outdone, Google’s DeepMind AI recently took on—and beat—several professional players at StarCraft II.

These victories beg the questions: what games are humans still better at than AI? And for how long?

The Making Of AlphaStar
DeepMind’s results provide a good starting point in a search for answers. The version of its AI for StarCraft II, dubbed AlphaStar, learned to play the games through supervised learning and reinforcement learning.

First, AI agents were trained by analyzing and copying human players, learning basic strategies. The initial agents then played each other in a sort of virtual death match where the strongest agents stayed on. New iterations of the agents were developed and entered the competition. Over time, the agents became better and better at the game, learning new strategies and tactics along the way.

One of the advantages of AI is that it can go through this kind of process at superspeed and quickly develop better agents. DeepMind researchers estimate that the AlphaStar agents went through the equivalent of roughly 200 years of game time in about 14 days.

Cheating or One Hand Behind the Back?
The AlphaStar AI agents faced off against human professional players in a series of games streamed on YouTube and Twitch. The AIs trounced their human opponents, winning ten games on the trot, before pro player Grzegorz “MaNa” Komincz managed to salvage some pride for humanity by winning the final game. Experts commenting on AlphaStar’s performance used words like “phenomenal” and “superhuman”—which was, to a degree, where things got a bit problematic.

AlphaStar proved particularly skilled at controlling and directing units in battle, known as micromanagement. One reason was that it viewed the whole game map at once—something a human player is not able to do—which made it seemingly able to control units in different areas at the same time. DeepMind researchers said the AIs only focused on a single part of the map at any given time, but interestingly, AlphaStar’s AI agent was limited to a more restricted camera view during the match “MaNA” won.

Potentially offsetting some of this advantage was the fact that AlphaStar was also restricted in certain ways. For example, it was prevented from performing more clicks per minute than a human player would be able to.

Where AIs Struggle
Games like StarCraft II and Dota 2 throw a lot of challenges at AIs. Complex game theory/ strategies, operating with imperfect/incomplete information, undertaking multi-variable and long-term planning, real-time decision-making, navigating a large action space, and making a multitude of possible decisions at every point in time are just the tip of the iceberg. The AIs’ performance in both games was impressive, but also highlighted some of the areas where they could be said to struggle.

In Dota 2 and StarCraft II, AI bots have seemed more vulnerable in longer games, or when confronted with surprising, unfamiliar strategies. They seem to struggle with complexity over time and improvisation/adapting to quick changes. This could be tied to how AIs learn. Even within the first few hours of performing a task, humans tend to gain a sense of familiarity and skill that takes an AI much longer. We are also better at transferring skill from one area to another. In other words, experience playing Dota 2 can help us become good at StarCraft II relatively quickly. This is not the case for AI—yet.

Dwindling Superiority
While the battle between AI and humans for absolute superiority is still on in Dota 2 and StarCraft II, it looks likely that AI will soon reign supreme. Similar things are happening to other types of games.

In 2017, a team from Carnegie Mellon University pitted its Libratus AI against four professionals. After 20 days of No Limit Texas Hold’em, Libratus was up by $1.7 million. Another likely candidate is the destroyer of family harmony at Christmas: Monopoly.

Poker involves bluffing, while Monopoly involves negotiation—skills you might not think AI would be particularly suited to handle. However, an AI experiment at Facebook showed that AI bots are more than capable of undertaking such tasks. The bots proved skilled negotiators, and developed negotiating strategies like pretending interest in one object while they were interested in another altogether—bluffing.

So, what games are we still better at than AI? There is no precise answer, but the list is getting shorter at a rapid pace.

The Aim Of the Game
While AI’s mastery of games might at first glance seem an odd area to focus research on, the belief is that the way AI learn to master a game is transferrable to other areas.

For example, the Libratus poker-playing AI employed strategies that could work in financial trading or political negotiations. The same applies to AlphaStar. As Oriol Vinyals, co-leader of the AlphaStar project, told The Verge:

“First and foremost, the mission at DeepMind is to build an artificial general intelligence. […] To do so, it’s important to benchmark how our agents perform on a wide variety of tasks.”

A 2017 survey of more than 350 AI researchers predicts AI could be a better driver than humans within ten years. By the middle of the century, AI will be able to write a best-selling novel, and a few years later, it will be better than humans at surgery. By the year 2060, AI may do everything better than us.

Whether you think this is a good or a bad thing, it’s worth noting that AI has an often overlooked ability to help us see things differently. When DeepMind’s AlphaGo beat human Go champion Lee Sedol, the Go community learned from it, too. Lee himself went on a win streak after the match with AlphaGo. The same is now happening within the Dota 2 and StarCraft II communities that are studying the human vs. AI games intensely.

More than anything, AI’s recent gaming triumphs illustrate how quickly artificial intelligence is developing. In 1997, Dr. Piet Hut, an astrophysicist at the Institute for Advanced Study at Princeton and a GO enthusiast, told the New York Times that:

”It may be a hundred years before a computer beats humans at Go—maybe even longer.”

Image Credit: Roman Kosolapov / Shutterstock.com Continue reading

Posted in Human Robots

#434580 How Genome Sequencing and Senolytics Can ...

The causes of aging are extremely complex and unclear. With the dramatic demonetization of genome reading and editing over the past decade, and Big Pharma, startups, and the FDA starting to face aging as a disease, we are starting to find practical ways to extend our healthspan.

Here, in Part 2 of a series of blogs on longevity and vitality, I explore how genome sequencing and editing, along with new classes of anti-aging drugs, are augmenting our biology to further extend our healthy lives.

In this blog I’ll cover two classes of emerging technologies:

Genome Sequencing and Editing;
Senolytics, Nutraceuticals & Pharmaceuticals.

Let’s dive in.

Genome Sequencing & Editing
Your genome is the software that runs your body.

A sequence of 3.2 billion letters makes you “you.” These base pairs of A’s, T’s, C’s, and G’s determine your hair color, your height, your personality, your propensity to disease, your lifespan, and so on.

Until recently, it’s been very difficult to rapidly and cheaply “read” these letters—and even more difficult to understand what they mean.

Since 2001, the cost to sequence a whole human genome has plummeted exponentially, outpacing Moore’s Law threefold. From an initial cost of $3.7 billion, it dropped to $10 million in 2006, and to $5,000 in 2012.

Today, the cost of genome sequencing has dropped below $500, and according to Illumina, the world’s leading sequencing company, the process will soon cost about $100 and take about an hour to complete.

This represents one of the most powerful and transformative technology revolutions in healthcare.

When we understand your genome, we’ll be able to understand how to optimize “you.”

We’ll know the perfect foods, the perfect drugs, the perfect exercise regimen, and the perfect supplements, just for you.
We’ll understand what microbiome types, or gut flora, are ideal for you (more on this in a later blog).
We’ll accurately predict how specific sedatives and medicines will impact you.
We’ll learn which diseases and illnesses you’re most likely to develop and, more importantly, how to best prevent them from developing in the first place (rather than trying to cure them after the fact).

CRISPR Gene Editing
In addition to reading the human genome, scientists can now edit a genome using a naturally-occurring biological system discovered in 1987 called CRISPR/Cas9.

Short for Clustered Regularly Interspaced Short Palindromic Repeats and CRISPR-associated protein 9, the editing system was adapted from a naturally-occurring defense system found in bacteria.

Here’s how it works:

The bacteria capture snippets of DNA from invading viruses (or bacteriophage) and use them to create DNA segments known as CRISPR arrays.
The CRISPR arrays allow the bacteria to “remember” the viruses (or closely related ones), and defend against future invasions.
If the viruses attack again, the bacteria produce RNA segments from the CRISPR arrays to target the viruses’ DNA. The bacteria then use Cas9 to cut the DNA apart, which disables the virus.

Most importantly, CRISPR is cheap, quick, easy to use, and more accurate than all previous gene editing methods. As a result, CRISPR/Cas9 has swept through labs around the world as the way to edit a genome.

A short search in the literature will show an exponential rise in the number of CRISPR-related publications and patents.

2018: Filled With CRISPR Breakthroughs
Early results are impressive. Researchers from the University of Chicago recently used CRISPR to genetically engineer cocaine resistance into mice.

Researchers at the University of Texas Southwestern Medical Center used CRISPR to reverse the gene defect causing Duchenne muscular dystrophy (DMD) in dogs (DMD is the most common fatal genetic disease in children).

With great power comes great responsibility, and moral and ethical dilemmas.

In 2015, Chinese scientists sparked global controversy when they first edited human embryo cells in the lab with the goal of modifying genes that would make the child resistant to smallpox, HIV, and cholera.

Three years later, in November 2018, researcher He Jiankui informed the world that the first set of CRISPR-engineered female twins had been delivered.

To accomplish his goal, Jiankui deleted a region of a receptor on the surface of white blood cells known as CCR5, introducing a rare, natural genetic variation that makes it more difficult for HIV to infect its favorite target, white blood cells.

Setting aside the significant ethical conversations, CRISPR will soon provide us the tools to eliminate diseases, create hardier offspring, produce new environmentally resistant crops, and even wipe out pathogens.

Senolytics, Nutraceuticals & Pharmaceuticals
Over the arc of your life, the cells in your body divide until they reach what is known as the Hayflick limit, or the number of times a normal human cell population will divide before cell division stops, which is typically about 50 divisions.

What normally follows next is programmed cell death or destruction by the immune system. A very small fraction of cells, however, become senescent cells and evade this fate to linger indefinitely.

These lingering cells secrete a potent mix of molecules that triggers chronic inflammation, damages the surrounding tissue structures, and changes the behavior of nearby cells for the worse.

Senescent cells appear to be one of the root causes of aging, causing everything from fibrosis and blood vessel calcification, to localized inflammatory conditions such as osteoarthritis, to diminished lung function.

Fortunately, both the scientific and entrepreneurial communities have begun to work on senolytic therapies, moving the technology for selectively destroying senescent cells out of the laboratory and into a half-dozen startup companies.

Prominent companies in the field include the following:

Unity Biotechnology is developing senolytic medicines to selectively eliminate senescent cells with an initial focus on delivering localized therapy in osteoarthritis, ophthalmology and pulmonary disease.
Oisin Biotechnologiesis pioneering a programmable gene therapy that can destroy cells based on their internal biochemistry.
SIWA Therapeuticsis working on an immunotherapy approach to the problem of senescent cells.

In recent years, researchers have identified or designed a handful of senolytic compounds that can curb aging by regulating senescent cells. Two of these drugs that have gained mainstay research traction are rapamycin and metformin.

Rapamycin
Originally extracted from bacteria found on Easter Island, Rapamycin acts on the m-TOR (mechanistic target of rapamycin) pathway to selectively block a key protein that facilitates cell division.

Currently, rapamycin derivatives are widely used as immunosuppression in organ and bone marrow transplants. Research now suggests that use results in prolonged lifespan and enhanced cognitive and immune function.

PureTech Health subsidiary resTORbio (which started 2018 by going public) is working on a rapamycin-based drug intended to enhance immunity and reduce infection. Their clinical-stage RTB101 drug works by inhibiting part of the mTOR pathway.

Results of the drug’s recent clinical trial include:

Decreased incidence of infection
Improved influenza vaccination response
A 30.6 percent decrease in respiratory tract infections

Impressive, to say the least.

Metformin
Metformin is a widely-used generic drug for mitigating liver sugar production in Type 2 diabetes patients.

Researchers have found that Metformin also reduces oxidative stress and inflammation, which otherwise increase as we age.

There is strong evidence that Metformin can augment cellular regeneration and dramatically mitigate cellular senescence by reducing both oxidative stress and inflammation.

Over 100 studies registered on ClinicalTrials.gov are currently following up on strong evidence of Metformin’s protective effect against cancer.

Nutraceuticals and NAD+
Beyond cellular senescence, certain critical nutrients and proteins tend to decline as a function of age. Nutraceuticals combat aging by supplementing and replenishing these declining nutrient levels.

NAD+ exists in every cell, participating in every process from DNA repair to creating the energy vital for cellular processes. It’s been shown that NAD+ levels decline as we age.

The Elysium Health Basis supplement aims to elevate NAD+ levels in the body to extend one’s lifespan. Elysium’s clinical study reports that Basis increases NAD+ levels consistently by a sustained 40 percent.

Conclusion
These are just a taste of the tremendous momentum that longevity and aging technology has right now. As artificial intelligence and quantum computing transform how we decode our DNA and how we discover drugs, genetics and pharmaceuticals will become truly personalized.

The next blog in this series will demonstrate how artificial intelligence is converging with genetics and pharmaceuticals to transform how we approach longevity, aging, and vitality.

We are edging closer to a dramatically extended healthspan—where 100 is the new 60. What will you create, where will you explore, and how will you spend your time if you are able to add an additional 40 healthy years to your life?

Join Me
Abundance Digital is my online educational portal and community of abundance-minded entrepreneurs. You’ll find weekly video updates from Peter, a curated newsfeed of exponential news, and a place to share your bold ideas. Click here to learn more and sign up.

Image Credit: ktsdesign / Shutterstock.com Continue reading

Posted in Human Robots

#434569 From Parkour to Surgery, Here Are the ...

The robot revolution may not be here quite yet, but our mechanical cousins have made some serious strides. And now some of the leading experts in the field have provided a rundown of what they see as the 10 most exciting recent developments.

Compiled by the editors of the journal Science Robotics, the list includes some of the most impressive original research and innovative commercial products to make a splash in 2018, as well as a couple from 2017 that really changed the game.

1. Boston Dynamics’ Atlas doing parkour

It seems like barely a few months go by without Boston Dynamics rewriting the book on what a robot can and can’t do. Last year they really outdid themselves when they got their Atlas humanoid robot to do parkour, leaping over logs and jumping between wooden crates.

Atlas’s creators have admitted that the videos we see are cherry-picked from multiple attempts, many of which don’t go so well. But they say they’re meant to be inspirational and aspirational rather than an accurate picture of where robotics is today. And combined with the company’s dog-like Spot robot, they are certainly pushing boundaries.

2. Intuitive Surgical’s da Vinci SP platform
Robotic surgery isn’t new, but the technology is improving rapidly. Market leader Intuitive’s da Vinci surgical robot was first cleared by the FDA in 2000, but since then it’s come a long way, with the company now producing three separate systems.

The latest addition is the da Vinci SP (single port) system, which is able to insert three instruments into the body through a single 2.5cm cannula (tube) bringing a whole new meaning to minimally invasive surgery. The system was granted FDA clearance for urological procedures last year, and the company has now started shipping the new system to customers.

3. Soft robot that navigates through growth

Roboticists have long borrowed principles from the animal kingdom, but a new robot design that mimics the way plant tendrils and fungi mycelium move by growing at the tip has really broken the mold on robot navigation.

The editors point out that this is the perfect example of bio-inspired design; the researchers didn’t simply copy nature, they took a general principle and expanded on it. The tube-like robot unfolds from the front as pneumatic pressure is applied, but unlike a plant, it can grow at the speed of an animal walking and can navigate using visual feedback from a camera.

4. 3D printed liquid crystal elastomers for soft robotics
Soft robotics is one of the fastest-growing sub-disciplines in the field, but powering these devices without rigid motors or pumps is an ongoing challenge. A variety of shape-shifting materials have been proposed as potential artificial muscles, including liquid crystal elastomeric actuators.

Harvard engineers have now demonstrated that these materials can be 3D printed using a special ink that allows the designer to easily program in all kinds of unusual shape-shifting abilities. What’s more, their technique produces actuators capable of lifting significantly more weight than previous approaches.

5. Muscle-mimetic, self-healing, and hydraulically amplified actuators
In another effort to find a way to power soft robots, last year researchers at the University of Colorado Boulder designed a series of super low-cost artificial muscles that can lift 200 times their own weight and even heal themselves.

The devices rely on pouches filled with a liquid that makes them contract with the force and speed of mammalian skeletal muscles when a voltage is applied. The most promising for robotics applications is the so-called Peano-HASEL, which features multiple rectangular pouches connected in series that contract linearly, just like real muscle.

6. Self-assembled nanoscale robot from DNA

While you may think of robots as hulking metallic machines, a substantial number of scientists are working on making nanoscale robots out of DNA. And last year German researchers built the first remote-controlled DNA robotic arm.

They created a length of tightly-bound DNA molecules to act as the arm and attached it to a DNA base plate via a flexible joint. Because DNA carries a charge, they were able to get the arm to swivel around like the hand of a clock by applying a voltage and switch direction by reversing that voltage. The hope is that this arm could eventually be used to build materials piece by piece at the nanoscale.

7. DelFly nimble bioinspired robotic flapper

Robotics doesn’t only borrow from biology—sometimes it gives back to it, too. And a new flapping-winged robot designed by Dutch engineers that mimics the humble fruit fly has done just that, by revealing how the animals that inspired it carry out predator-dodging maneuvers.

The lab has been building flapping robots for years, but this time they ditched the airplane-like tail used to control previous incarnations. Instead, they used insect-inspired adjustments to the motions of its twin pairs of flapping wings to hover, pitch, and roll with the agility of a fruit fly. That has provided a useful platform for investigating insect flight dynamics, as well as more practical applications.

8. Soft exosuit wearable robot

Exoskeletons could prevent workplace injuries, help people walk again, and even boost soldiers’ endurance. Strapping on bulky equipment isn’t ideal, though, so researchers at Harvard are working on a soft exoskeleton that combines specially-designed textiles, sensors, and lightweight actuators.

And last year the team made an important breakthrough by combining their novel exoskeleton with a machine-learning algorithm that automatically tunes the device to the user’s particular walking style. Using physiological data, it is able to adjust when and where the device needs to deliver a boost to the user’s natural movements to improve walking efficiency.

9. Universal Robots (UR) e-Series Cobots
Robots in factories are nothing new. The enormous mechanical arms you see in car factories normally have to be kept in cages to prevent them from accidentally crushing people. In recent years there’s been growing interest in “co-bots,” collaborative robots designed to work side-by-side with their human colleagues and even learn from them.

Earlier this year saw the demise of ReThink robotics, the pioneer of the approach. But the simple single arm devices made by Danish firm Universal Robotics are becoming ubiquitous in workshops and warehouses around the world, accounting for about half of global co-bot sales. Last year they released their latest e-Series, with enhanced safety features and force/torque sensing.

10. Sony’s aibo
After a nearly 20-year hiatus, Sony’s robotic dog aibo is back, and it’s had some serious upgrades. As well as a revamp to its appearance, the new robotic pet takes advantage of advances in AI, with improved environmental and command awareness and the ability to develop a unique character based on interactions with its owner.

The editors note that this new context awareness mark the device out as a significant evolution in social robots, which many hope could aid in childhood learning or provide companionship for the elderly.

Image Credit: DelFly Nimble / CC BY – SA 4.0 Continue reading

Posted in Human Robots

#434559 Can AI Tell the Difference Between a ...

Scarcely a day goes by without another headline about neural networks: some new task that deep learning algorithms can excel at, approaching or even surpassing human competence. As the application of this approach to computer vision has continued to improve, with algorithms capable of specialized recognition tasks like those found in medicine, the software is getting closer to widespread commercial use—for example, in self-driving cars. Our ability to recognize patterns is a huge part of human intelligence: if this can be done faster by machines, the consequences will be profound.

Yet, as ever with algorithms, there are deep concerns about their reliability, especially when we don’t know precisely how they work. State-of-the-art neural networks will confidently—and incorrectly—classify images that look like television static or abstract art as real-world objects like school-buses or armadillos. Specific algorithms could be targeted by “adversarial examples,” where adding an imperceptible amount of noise to an image can cause an algorithm to completely mistake one object for another. Machine learning experts enjoy constructing these images to trick advanced software, but if a self-driving car could be fooled by a few stickers, it might not be so fun for the passengers.

These difficulties are hard to smooth out in large part because we don’t have a great intuition for how these neural networks “see” and “recognize” objects. The main insight analyzing a trained network itself can give us is a series of statistical weights, associating certain groups of points with certain objects: this can be very difficult to interpret.

Now, new research from UCLA, published in the journal PLOS Computational Biology, is testing neural networks to understand the limits of their vision and the differences between computer vision and human vision. Nicholas Baker, Hongjing Lu, and Philip J. Kellman of UCLA, alongside Gennady Erlikhman of the University of Nevada, tested a deep convolutional neural network called VGG-19. This is state-of-the-art technology that is already outperforming humans on standardized tests like the ImageNet Large Scale Visual Recognition Challenge.

They found that, while humans tend to classify objects based on their overall (global) shape, deep neural networks are far more sensitive to the textures of objects, including local color gradients and the distribution of points on the object. This result helps explain why neural networks in image recognition make mistakes that no human ever would—and could allow for better designs in the future.

In the first experiment, a neural network was trained to sort images into 1 of 1,000 different categories. It was then presented with silhouettes of these images: all of the local information was lost, while only the outline of the object remained. Ordinarily, the trained neural net was capable of recognizing these objects, assigning more than 90% probability to the correct classification. Studying silhouettes, this dropped to 10%. While human observers could nearly always produce correct shape labels, the neural networks appeared almost insensitive to the overall shape of the images. On average, the correct object was ranked as the 209th most likely solution by the neural network, even though the overall shapes were an exact match.

A particularly striking example arose when they tried to get the neural networks to classify glass figurines of objects they could already recognize. While you or I might find it easy to identify a glass model of an otter or a polar bear, the neural network classified them as “oxygen mask” and “can opener” respectively. By presenting glass figurines, where the texture information that neural networks relied on for classifying objects is lost, the neural network was unable to recognize the objects by shape alone. The neural network was similarly hopeless at classifying objects based on drawings of their outline.

If you got one of these right, you’re better than state-of-the-art image recognition software. Image Credit: Nicholas Baker, Hongjing Lu, Gennady Erlikhman, Philip J. Kelman. “Deep convolutional networks do not classify based on global object shape.” Plos Computational Biology. 12/7/18. / CC BY 4.0
When the neural network was explicitly trained to recognize object silhouettes—given no information in the training data aside from the object outlines—the researchers found that slight distortions or “ripples” to the contour of the image were again enough to fool the AI, while humans paid them no mind.

The fact that neural networks seem to be insensitive to the overall shape of an object—relying instead on statistical similarities between local distributions of points—suggests a further experiment. What if you scrambled the images so that the overall shape was lost but local features were preserved? It turns out that the neural networks are far better and faster at recognizing scrambled versions of objects than outlines, even when humans struggle. Students could classify only 37% of the scrambled objects, while the neural network succeeded 83% of the time.

Humans vastly outperform machines at classifying object (a) as a bear, while the machine learning algorithm has few problems classifying the bear in figure (b). Image Credit: Nicholas Baker, Hongjing Lu, Gennady Erlikhman, Philip J. Kelman. “Deep convolutional networks do not classify based on global object shape.” Plos Computational Biology. 12/7/18. / CC BY 4.0
“This study shows these systems get the right answer in the images they were trained on without considering shape,” Kellman said. “For humans, overall shape is primary for object recognition, and identifying images by overall shape doesn’t seem to be in these deep learning systems at all.”

Naively, one might expect that—as the many layers of a neural network are modeled on connections between neurons in the brain and resemble the visual cortex specifically—the way computer vision operates must necessarily be similar to human vision. But this kind of research shows that, while the fundamental architecture might resemble that of the human brain, the resulting “mind” operates very differently.

Researchers can, increasingly, observe how the “neurons” in neural networks light up when exposed to stimuli and compare it to how biological systems respond to the same stimuli. Perhaps someday it might be possible to use these comparisons to understand how neural networks are “thinking” and how those responses differ from humans.

But, as yet, it takes a more experimental psychology to probe how neural networks and artificial intelligence algorithms perceive the world. The tests employed against the neural network are closer to how scientists might try to understand the senses of an animal or the developing brain of a young child rather than a piece of software.

By combining this experimental psychology with new neural network designs or error-correction techniques, it may be possible to make them even more reliable. Yet this research illustrates just how much we still don’t understand about the algorithms we’re creating and using: how they tick, how they make decisions, and how they’re different from us. As they play an ever-greater role in society, understanding the psychology of neural networks will be crucial if we want to use them wisely and effectively—and not end up missing the woods for the trees.

Image Credit: Irvan Pratama / Shutterstock.com Continue reading

Posted in Human Robots