Tag Archives: form

#439073 There’s a ‘New’ Nirvana Song Out, ...

One of the primary capabilities separating human intelligence from artificial intelligence is our ability to be creative—to use nothing but the world around us, our experiences, and our brains to create art. At present, AI needs to be extensively trained on human-made works of art in order to produce new work, so we’ve still got a leg up. That said, neural networks like OpenAI’s GPT-3 and Russian designer Nikolay Ironov have been able to create content indistinguishable from human-made work.

Now there’s another example of AI artistry that’s hard to tell apart from the real thing, and it’s sure to excite 90s alternative rock fans the world over: a brand-new, never-heard-before Nirvana song. Or, more accurately, a song written by a neural network that was trained on Nirvana’s music.

The song is called “Drowned in the Sun,” and it does have a pretty Nirvana-esque ring to it. The neural network that wrote it is Magenta, which was launched by Google in 2016 with the goal of training machines to create art—or as the tool’s website puts it, exploring the role of machine learning as a tool in the creative process. Magenta was built using TensorFlow, Google’s massive open-source software library focused on deep learning applications.

The song was written as part of an album called Lost Tapes of the 27 Club, a project carried out by a Toronto-based organization called Over the Bridge focused on mental health in the music industry.

Here’s how a computer was able to write a song in the unique style of a deceased musician. Music, 20 to 30 tracks, was fed into Magenta’s neural network in the form of MIDI files. MIDI stands for Musical Instrument Digital Interface, and the format contains the details of a song written in code that represents musical parameters like pitch and tempo. Components of each song, like vocal melody or rhythm guitar, were fed in one at a time.

The neural network found patterns in these different components, and got enough of a handle on them that when given a few notes to start from, it could use those patterns to predict what would come next; in this case, chords and melodies that sound like they could’ve been written by Kurt Cobain.

To be clear, Magenta didn’t spit out a ready-to-go song complete with lyrics. The AI wrote the music, but a different neural network wrote the lyrics (using essentially the same process as Magenta), and the team then sifted through “pages and pages” of output to find lyrics that fit the melodies Magenta created.

Eric Hogan, a singer for a Nirvana tribute band who the Over the Bridge team hired to sing “Drowned in the Sun,” felt that the lyrics were spot-on. “The song is saying, ‘I’m a weirdo, but I like it,’” he said. “That is total Kurt Cobain right there. The sentiment is exactly what he would have said.”

Cobain isn’t the only musician the Lost Tapes project tried to emulate; songs in the styles of Jimi Hendrix, Jim Morrison, and Amy Winehouse were also included. What all these artists have in common is that they died by suicide at the age of 27.

The project is meant to raise awareness around mental health, particularly among music industry professionals. It’s not hard to think of great artists of all persuasions—musicians, painters, writers, actors—whose lives are cut short due to severe depression and other mental health issues for which it can be hard to get help. These issues are sometimes romanticized, as suffering does tend to create art that’s meaningful, relatable, and timeless. But according to the Lost Tapes website, suicide attempts among music industry workers are more than double that of the general population.

How many more hit songs would these artists have written if they were still alive? We’ll never know, but hopefully Lost Tapes of the 27 Club and projects like it will raise awareness of mental health issues, both in the music industry and in general, and help people in need find the right resources. Because no matter how good computers eventually get at creating music, writing, or other art, as Lost Tapes’ website pointedly says, “Even AI will never replace the real thing.”

Image Credit: Edward Xu on Unsplash Continue reading

Posted in Human Robots

#439062 Xenobots 2.0: These Living Robots ...

The line between animals and machines was already getting blurry after a team of scientists and roboticists unveiled the first living robots last year. Now the same team has released version 2.0 of their so-called xenobots, and they’re faster, stronger, and more capable than ever.

In January 2020, researchers from Tufts University and the University of Vermont laid out a method for building tiny biological machines out of the eggs of the African claw frog Xenopus laevis. Dubbed xenobots after their animal forebear, they could move independently, push objects, and even team up to create swarms.

Remarkably, building them involved no genetic engineering. Instead, the team used an evolutionary algorithm running on a supercomputer to test out thousands of potential designs made up of different configurations of cells.

Once they’d found some promising candidates that could solve the tasks they were interested in, they used microsurgical tools to build real-world versions out of living cells. The most promising design was built by splicing heart muscle cells (which could contract to propel the xenobots), and skin cells (which provided a rigid support).

Impressive as that might sound, having to build each individual xenobot by hand is obviously tedious. But now the team has devised a new approach that works from the bottom up by getting the xenobots to self-assemble their bodies from single cells. Not only is the approach more scalable, the new xenobots are faster, live longer, and even have a rudimentary memory.

In a paper in Science Robotics, the researchers describe how they took stem cells from frog embryos and allowed them to grow into clumps of several thousand cells called spheroids. After a few days, the stem cells had turned into skin cells covered in small hair-like projections called cilia, which wriggle back and forth.

Normally, these structures are used to spread mucus around on the frog’s skin. But when divorced from their normal context they took on a function more similar to that seen in microorganisms, which use cilia to move about by acting like tiny paddles.

“We are witnessing the remarkable plasticity of cellular collectives, which build a rudimentary new ‘body’ that is quite distinct from their default—in this case, a frog—despite having a completely normal genome,” corresponding author Michael Levin from Tufts University said in a press release.

“We see that cells can re-purpose their genetically encoded hardware, like cilia, for new functions such as locomotion. It is amazing that cells can spontaneously take on new roles and create new body plans and behaviors without long periods of evolutionary selection for those features,” he said.

Not only were the new xenobots faster and longer-lived, they were also much better at tasks like working together as a swarm to gather piles of iron oxide particles. And while the form and function of the xenobots was achieved without any genetic engineering, in an extra experiment the team injected them with RNA that caused them to produce a fluorescent protein that changes color when exposed to a particular color of light.

This allowed the xenobots to record whether they had come into contact with a specific light source while traveling about. The researchers say this is a proof of principle that the xenobots can be imbued with a molecular memory, and future work could allow them to record multiple stimuli and potentially even react to them.

What exactly these xenobots could eventually be used for is still speculative, but they have features that make them a promising alternative to non-organic alternatives. For a start, robots made of stem cells are completely biodegradable and also have their own power source in the form of “yolk platelets” found in all amphibian embryos. They are also able to self-heal in as little as five minutes if cut, and can take advantage of cells’ ability to process all kinds of chemicals.

That suggests they could have applications in everything from therapeutics to environmental engineering. But the researchers also hope to use them to better understand the processes that allow individual cells to combine and work together to create a larger organism, and how these processes might be harnessed and guided for regenerative medicine.

As these animal-machine hybrids advance, they are sure to raise ethical concerns and question marks over the potential risks. But it looks like the future of robotics could be a lot more wet and squishy than we imagined.

Image Credit: Doug Blackiston/Tufts University Continue reading

Posted in Human Robots

#438925 Nanophotonics Could Be the ‘Dark ...

The race to build the first practical quantum computers looks like a two-horse contest between machines built from superconducting qubits and those that use trapped ions. But new research suggests a third contender—machines based on optical technology—could sneak up on the inside.

The most advanced quantum computers today are the ones built by Google and IBM, which rely on superconducting circuits to generate the qubits that form the basis of quantum calculations. They are now able to string together tens of qubits, and while controversial, Google claims its machines have achieved quantum supremacy—the ability to carry out a computation beyond normal computers.

Recently this approach has been challenged by a wave of companies looking to use trapped ion qubits, which are more stable and less error-prone than superconducting ones. While these devices are less developed, engineering giant Honeywell has already released a machine with 10 qubits, which it says is more powerful than a machine made of a greater number of superconducting qubits.

But despite this progress, both of these approaches have some major drawbacks. They require specialized fabrication methods, incredibly precise control mechanisms, and they need to be cooled to close to absolute zero to protect the qubits from any outside interference.

That’s why researchers at Canadian quantum computing hardware and software startup Xanadu are backing an alternative quantum computing approach based on optics, which was long discounted as impractical. In a paper published last week in Nature, they unveiled the first fully programmable and scalable optical chip that can run quantum algorithms. Not only does the system run at room temperature, but the company says it could scale to millions of qubits.

The idea isn’t exactly new. As Chris Lee notes in Ars Technica, people have been experimenting with optical approaches to quantum computing for decades, because encoding information in photons’ quantum states and manipulating those states is relatively easy. The biggest problem was that optical circuits were very large and not readily programmable, which meant you had to build a new computer for every new problem you wanted to solve.

That started to change thanks to the growing maturity of photonic integrated circuits. While early experiments with optical computing involved complex table-top arrangements of lasers, lenses, and detectors, today it’s possible to buy silicon chips not dissimilar to electronic ones that feature hundreds of tiny optical components.

In recent years, the reliability and performance of these devices has improved dramatically, and they’re now regularly used by the telecommunications industry. Some companies believe they could be the future of artificial intelligence too.

This allowed the Xanadu researchers to design a silicon chip that implements a complex optical network made up of beam splitters, waveguides, and devices called interferometers that cause light sources to interact with each other.

The chip can generate and manipulate up to eight qubits, but unlike conventional qubits, which can simultaneously be in two states, these qubits can be in any configuration of three states, which means they can carry more information.

Once the light has travelled through the network, it is then fed out to cutting-edge photon-counting detectors that provide the result. This is one of the potential limitations of the system, because currently these detectors need to be cryogenically cooled, although the rest of the chip does not.

But most importantly, the chip is easily re-programmable, which allows it to tackle a variety of problems. The computation can be controlled by adjusting the settings of these interferometers, but the researchers have also developed a software platform that hides the physical complexity from users and allows them to program it using fairly conventional code.

The company announced that its chips were available on the cloud in September of 2020, but the Nature paper is the first peer-reviewed test of their system. The researchers verified that the computations being done were genuinely quantum mechanical in nature, but they also implemented two more practical algorithms: one for simulating molecules and the other for judging how similar two graphs are, which has applications in a variety of pattern recognition problems.

In an accompanying opinion piece, Ulrik Andersen from the Technical University of Denmark says the quality of the qubits needs to be improved considerably and photon losses reduced if the technology is ever to scale to practical problems. But, he says, this breakthrough suggests optical approaches “could turn out to be the dark horse of quantum computing.”

Image Credit: Shahadat Rahman on Unsplash Continue reading

Posted in Human Robots

#438807 Visible Touch: How Cameras Can Help ...

The dawn of the robot revolution is already here, and it is not the dystopian nightmare we imagined. Instead, it comes in the form of social robots: Autonomous robots in homes and schools, offices and public spaces, able to interact with humans and other robots in a socially acceptable, human-perceptible way to resolve tasks related to core human needs.

To design social robots that “understand” humans, robotics scientists are delving into the psychology of human communication. Researchers from Cornell University posit that embedding the sense of touch in social robots could teach them to detect physical interactions and gestures. They describe a way of doing so by relying not on touch but on vision.

A USB camera inside the robot captures shadows of hand gestures on the robot’s surface and classifies them with machine-learning software. They call this method ShadowSense, which they define as a modality between vision and touch, bringing “the high resolution and low cost of vision-sensing to the close-up sensory experience of touch.”

Touch-sensing in social or interactive robots is usually achieved with force sensors or capacitive sensors, says study co-author Guy Hoffman of the Sibley School of Mechanical and Aerospace Engineering at Cornell University. The drawback to his group’s approach has been that, even to achieve coarse spatial resolution, many sensors are needed in a small area.

However, working with non-rigid, inflatable robots, Hoffman and his co-researchers installed a consumer-grade USB camera to which they attached a fisheye lens for a wider field of vision.

“Given that the robot is already hollow, and has a soft and translucent skin, we could do touch interaction by looking at the shadows created by people touching the robot,” says Hoffman. They used deep neural networks to interpret the shadows. “And we were able to do it with very high accuracy,” he says. The robot was able to interpret six different gestures, including one- or two-handed touch, pointing, hugging and punching, with an accuracy of 87.5 to 96 percent, depending on the lighting.

This is not the first time that computer vision has been used for tactile sensing, though the scale and application of ShadowSense is unique. “Photography has been used for touch mainly in robotic grasping,” says Hoffman. By contrast, Hoffman and collaborators wanted to develop a sense that could be “felt” across the whole of the device.

The potential applications for ShadowSense include mobile robot guidance using touch, and interactive screens on soft robots. A third concerns privacy, especially in home-based social robots. “We have another paper currently under review that looks specifically at the ability to detect gestures that are further away [from the robot’s skin],” says Hoffman. This way, users would be able to cover their robot’s camera with a translucent material and still allow it to interpret actions and gestures from shadows. Thus, even though it’s prevented from capturing a high-resolution image of the user or their surrounding environment, using the right kind of training datasets, the robot can continue to monitor some kinds of non-tactile activities.

In its current iteration, Hoffman says, ShadowSense doesn’t do well in low-light conditions. Environmental noise, or shadows from surrounding objects, also interfere with image classification. Relying on one camera also means a single point of failure. “I think if this were to become a commercial product, we would probably [have to] work a little bit better on image detection,” says Hoffman.

As it was, the researchers used transfer learning—reusing a pre-trained deep-learning model in a new problem—for image analysis. “One of the problems with multi-layered neural networks is that you need a lot of training data to make accurate predictions,” says Hoffman. “Obviously, we don’t have millions of examples of people touching a hollow, inflatable robot. But we can use pre-trained networks trained on general images, which we have billions of, and we only retrain the last layers of the network using our own dataset.” Continue reading

Posted in Human Robots

#438801 This AI Thrashes the Hardest Atari Games ...

Learning from rewards seems like the simplest thing. I make coffee, I sip coffee, I’m happy. My brain registers “brewing coffee” as an action that leads to a reward.

That’s the guiding insight behind deep reinforcement learning, a family of algorithms that famously smashed most of Atari’s gaming catalog and triumphed over humans in strategy games like Go. Here, an AI “agent” explores the game, trying out different actions and registering ones that let it win.

Except it’s not that simple. “Brewing coffee” isn’t one action; it’s a series of actions spanning several minutes, where you’re only rewarded at the very end. By just tasting the final product, how do you learn to fine-tune grind coarseness, water to coffee ratio, brewing temperature, and a gazillion other factors that result in the reward—tasty, perk-me-up coffee?

That’s the problem with “sparse rewards,” which are ironically very abundant in our messy, complex world. We don’t immediately get feedback from our actions—no video-game-style dings or points for just grinding coffee beans—yet somehow we’re able to learn and perform an entire sequence of arm and hand movements while half-asleep.

This week, researchers from UberAI and OpenAI teamed up to bestow this talent on AI.

The trick is to encourage AI agents to “return” to a previous step, one that’s promising for a winning solution. The agent then keeps a record of that state, reloads it, and branches out again to intentionally explore other solutions that may have been left behind on the first go-around. Video gamers are likely familiar with this idea: live, die, reload a saved point, try something else, repeat for a perfect run-through.

The new family of algorithms, appropriately dubbed “Go-Explore,” smashed notoriously difficult Atari games like Montezuma’s Revenge that were previously unsolvable by its AI predecessors, while trouncing human performance along the way.

It’s not just games and digital fun. In a computer simulation of a robotic arm, the team found that installing Go-Explore as its “brain” allowed it to solve a challenging series of actions when given very sparse rewards. Because the overarching idea is so simple, the authors say, it can be adapted and expanded to other real-world problems, such as drug design or language learning.

Growing Pains
How do you reward an algorithm?

Rewards are very hard to craft, the authors say. Take the problem of asking a robot to go to a fridge. A sparse reward will only give the robot “happy points” if it reaches its destination, which is similar to asking a baby, with no concept of space and danger, to crawl through a potential minefield of toys and other obstacles towards a fridge.

“In practice, reinforcement learning works very well, if you have very rich feedback, if you can tell, ‘hey, this move is good, that move is bad, this move is good, that move is bad,’” said study author Joost Huinzinga. However, in situations that offer very little feedback, “rewards can intentionally lead to a dead end. Randomly exploring the space just doesn’t cut it.”

The other extreme is providing denser rewards. In the same robot-to-fridge example, you could frequently reward the bot as it goes along its journey, essentially helping “map out” the exact recipe to success. But that’s troubling as well. Over-holding an AI’s hand could result in an extremely rigid robot that ignores new additions to its path—a pet, for example—leading to dangerous situations. It’s a deceptive AI solution that seems effective in a simple environment, but crashes in the real world.

What we need are AI agents that can tackle both problems, the team said.

Intelligent Exploration
The key is to return to the past.

For AI, motivation usually comes from “exploring new or unusual situations,” said Huizinga. It’s efficient, but comes with significant downsides. For one, the AI agent could prematurely stop going back to promising areas because it thinks it had already found a good solution. For another, it could simply forget a previous decision point because of the mechanics of how it probes the next step in a problem.

For a complex task, the end result is an AI that randomly stumbles around towards a solution while ignoring potentially better ones.

“Detaching from a place that was previously visited after collecting a reward doesn’t work in difficult games, because you might leave out important clues,” Huinzinga explained.

Go-Explore solves these problems with a simple principle: first return, then explore. In essence, the algorithm saves different approaches it previously tried and loads promising save points—once more likely to lead to victory—to explore further.

Digging a bit deeper, the AI stores screen caps from a game. It then analyzes saved points and groups images that look alike as a potential promising “save point” to return to. Rinse and repeat. The AI tries to maximize its final score in the game, and updates its save points when it achieves a new record score. Because Atari doesn’t usually allow people to revisit any random point, the team used an emulator, which is a kind of software that mimics the Atari system but with custom abilities such as saving and reloading at any time.

The trick worked like magic. When pitted against 55 Atari games in the OpenAI gym, now commonly used to benchmark reinforcement learning algorithms, Go-Explore knocked out state-of-the-art AI competitors over 85 percent of the time.

It also crushed games previously unbeatable by AI. Montezuma’s Revenge, for example, requires you to move Pedro, the blocky protagonist, through a labyrinth of underground temples while evading obstacles such as traps and enemies and gathering jewels. One bad jump could derail the path to the next level. It’s a perfect example of sparse rewards: you need a series of good actions to get to the reward—advancing onward.

Go-Explore didn’t just beat all levels of the game, a first for AI. It also scored higher than any previous record for reinforcement learning algorithms at lower levels while toppling the human world record.

Outside a gaming environment, Go-Explore was also able to boost the performance of a simulated robot arm. While it’s easy for humans to follow high-level guidance like “put the cup on this shelf in a cupboard,” robots often need explicit training—from grasping the cup to recognizing a cupboard, moving towards it while avoiding obstacles, and learning motions to not smash the cup when putting it down.

Here, similar to the real world, the digital robot arm was only rewarded when it placed the cup onto the correct shelf, out of four possible shelves. When pitted against another algorithm, Go-Explore quickly figured out the movements needed to place the cup, while its competitor struggled with even reliably picking the cup up.

Combining Forces
By itself, the “first return, then explore” idea behind Go-Explore is already powerful. The team thinks it can do even better.

One idea is to change the mechanics of save points. Rather than reloading saved states through the emulator, it’s possible to train a neural network to do the same, without needing to relaunch a saved state. It’s a potential way to make the AI even smarter, the team said, because it can “learn” to overcome one obstacle once, instead of solving the same problem again and again. The downside? It’s much more computationally intensive.

Another idea is to combine Go-Explore with an alternative form of learning, called “imitation learning.” Here, an AI observes human behavior and mimics it through a series of actions. Combined with Go-Explore, said study author Adrien Ecoffet, this could make more robust robots capable of handling all the complexity and messiness in the real world.

To the team, the implications go far beyond Go-Explore. The concept of “first return, then explore” seems to be especially powerful, suggesting “it may be a fundamental feature of learning in general.” The team said, “Harnessing these insights…may be essential…to create generally intelligent agents.”

Image Credit: Adrien Ecoffet, Joost Huizinga, Joel Lehman, Kenneth O. Stanley, and Jeff Clune Continue reading

Posted in Human Robots