Tag Archives: pretty
#437585 Dart-Shooting Drone Attacks Trees for ...
We all know how robots are great at going to places where you can’t (or shouldn’t) send a human. We also know how robots are great at doing repetitive tasks. These characteristics have the potential to make robots ideal for setting up wireless sensor networks in hazardous environments—that is, they could deploy a whole bunch of self-contained sensor nodes that create a network that can monitor a very large area for a very long time.
When it comes to using drones to set up sensor networks, you’ve generally got two options: A drone that just drops sensors on the ground (easy but inaccurate and limited locations), or using a drone with some sort of manipulator on it to stick sensors in specific places (complicated and risky). A third option, under development by researchers at Imperial College London’s Aerial Robotics Lab, provides the accuracy of direct contact with the safety and ease of use of passive dropping by instead using the drone as a launching platform for laser-aimed, sensor-equipped darts.
These darts (which the researchers refer to as aerodynamically stabilized, spine-equipped sensor pods) can embed themselves in relatively soft targets from up to 4 meters away with an accuracy of about 10 centimeters after being fired from a spring-loaded launcher. They’re not quite as accurate as a drone with a manipulator, but it’s pretty good, and the drone can maintain a safe distance from the surface that it’s trying to add a sensor to. Obviously, the spine is only going to work on things like wood, but the researchers point out that there are plenty of attachment mechanisms that could be used, including magnets, adhesives, chemical bonding, or microspines.
Indoor tests using magnets showed the system to be quite reliable, but at close range (within a meter of the target) the darts sometimes bounced off rather than sticking. From between 1 and 4 meters away, the darts stuck between 90 and 100 percent of the time. Initial outdoor tests were also successful, although the system was under manual control. The researchers say that “regular and safe operations should be carried out autonomously,” which, yeah, you’d just have to deal with all of the extra sensing and hardware required to autonomously fly beneath the canopy of a forest. That’s happening next, as the researchers plan to add “vision state estimation and positioning, as well as a depth sensor” to avoid some trees and fire sensors into others.
And if all of that goes well, they’ll consider trying to get each drone to carry multiple darts. Look out, trees: You’re about to be pierced for science.
“Unmanned Aerial Sensor Placement for Cluttered Environments,” by André Farinha, Raphael Zufferey, Peter Zheng, Sophie F. Armanini, and Mirko Kovac from Imperial College London, was published in IEEE Robotics and Automation Letters.
< Back to IEEE Journal Watch Continue reading
Posted in Human Robots
#437504 A New and Improved Burger Robot’s on ...
No doubt about it, the pandemic has changed the way we eat. Never before have so many people who hated cooking been forced to learn how to prepare a basic meal for themselves. With sit-down restaurants limiting their capacity or shutting down altogether, consumption of fast food and fast-casual food has skyrocketed. Don’t feel like slaving over a hot stove? Just hit the drive through and grab a sandwich and some fries (the health implications of increased fast food consumption are another matter…).
Given our sudden immense need for paper-wrapped burgers and cardboard cartons of fries, fast food workers are now counted as essential. But what about their safety, both from a virus standpoint and from the usual risks of working in a busy kitchen (like getting burned by the stove or the hot oil from the fryer, cut by a slicer, etc.)? And how many orders of burgers and fries can humans possibly churn out in an hour?
Enter the robot. Three and a half years ago, a burger-flipping robot aptly named Flippy, made by Miso Robotics, made its debut at a fast food restaurant in California called CaliBurger. Now Flippy is on the market for anyone who wishes to purchase their own, with a price tag of $30,000 and a range of new capabilities—this burger bot has progressed far beyond just flipping burgers.
Flippy’s first iteration was already pretty impressive. It used machine learning software to locate and identify objects in front of it (rather than needing to have objects lined up in specific spots), and was able to learn from experience to improve its accuracy. Sensors on its grill-facing side took in thermal and 3D data to gauge the cooking process for multiple patties at a time, and cameras allowed the robot to ‘see’ its surroundings.
A system that digitally sent tickets to the kitchen from the restaurant’s front counter kept Flippy on top of how many burgers it should be cooking at any given time. Its key tasks were pulling raw patties from a stack and placing them on the grill, tracking each burger’s cook time and temperature, and transferring cooked burgers to a plate.
The new and improved Flippy can do all this and more. It can cook 19 different foods, including chicken wings, onion rings, french fries, and even the Impossible Burger (which, as you may know, isn’t actually made of meat, and that means it’s a little trickier to grill it to perfection).
Flippy’s handiwork. Image Credit: Miso Robotics
And instead of its body sitting on a cart on wheels (which took up a lot of space and meant the robot’s arm could get in the way of human employees), it’s now attached to a rail along the stove’s hood, and can move along the rail to access both the grill and the fryer (provided they’re next to each other, which in many fast food restaurants they are). In fact, Flippy has a new acronym attached to its name: ROAR, which stands for Robot on a Rail.
Flippy ROAR in action, artist rendering. Image Credit: Miso Robotics
Sensors equipped with laser make it safer for human employees to work near Flippy. The bot can automatically switch between different tools, such as a spatula for flipping patties and tongs for gripping the handle of a fryer basket. Its AI software will enable it to learn new skills over time.
Flippy’s interface. Image Credit: Miso Robotics
The first big restaurant chain to go all-in on Flippy was White Castle, which in July announced plans to pilot Flippy ROAR before year’s end. And just last month, Miso made the bot commercially available. The current cost is $30,000 (plus a monthly fee of $1,500 for use of the software), but the company hopes to bring the price down to $20,000 within the next year.
According to Business Insider, demand for the fast food robot is through the roof, probably given a significant boost by the pandemic—thanks, Covid-19. The pace of automation has picked up across multiple sectors, and will likely continue to accelerate as companies look to insure themselves against additional losses.
So for the immediate future, it seems that no matter what happens, we don’t have to worry about the supply of burgers, fries, onion rings, chicken wings, and the like running out.
Now if only Flippy had a cousin—perhaps named Leafy—who could chop vegetables and greens and put together fresh-made salads…
Maybe that can be Miso Robotics’ next project.
Image Credit: Miso Robotics Continue reading
Posted in Human Robots
#437466 How Future AI Could Recognize a Kangaroo ...
AI is continuously taking on new challenges, from detecting deepfakes (which, incidentally, are also made using AI) to winning at poker to giving synthetic biology experiments a boost. These impressive feats result partly from the huge datasets the systems are trained on. That training is costly and time-consuming, and it yields AIs that can really only do one thing well.
For example, to train an AI to differentiate between a picture of a dog and one of a cat, it’s fed thousands—if not millions—of labeled images of dogs and cats. A child, on the other hand, can see a dog or cat just once or twice and remember which is which. How can we make AIs learn more like children do?
A team at the University of Waterloo in Ontario has an answer: change the way AIs are trained.
Here’s the thing about the datasets normally used to train AI—besides being huge, they’re highly specific. A picture of a dog can only be a picture of a dog, right? But what about a really small dog with a long-ish tail? That sort of dog, while still being a dog, looks more like a cat than, say, a fully-grown Golden Retriever.
It’s this concept that the Waterloo team’s methodology is based on. They described their work in a paper published on the pre-print (or non-peer-reviewed) server arXiv last month. Teaching an AI system to identify a new class of objects using just one example is what they call “one-shot learning.” But they take it a step further, focusing on “less than one shot learning,” or LO-shot learning for short.
LO-shot learning consists of a system learning to classify various categories based on a number of examples that’s smaller than the number of categories. That’s not the most straightforward concept to wrap your head around, so let’s go back to the dogs and cats example. Say you want to teach an AI to identify dogs, cats, and kangaroos. How could that possibly be done without several clear examples of each animal?
The key, the Waterloo team says, is in what they call soft labels. Unlike hard labels, which label a data point as belonging to one specific class, soft labels tease out the relationship or degree of similarity between that data point and multiple classes. In the case of an AI trained on only dogs and cats, a third class of objects, say, kangaroos, might be described as 60 percent like a dog and 40 percent like a cat (I know—kangaroos probably aren’t the best animal to have thrown in as a third category).
“Soft labels can be used to represent training sets using fewer prototypes than there are classes, achieving large increases in sample efficiency over regular (hard-label) prototypes,” the paper says. Translation? Tell an AI a kangaroo is some fraction cat and some fraction dog—both of which it’s seen and knows well—and it’ll be able to identify a kangaroo without ever having seen one.
If the soft labels are nuanced enough, you could theoretically teach an AI to identify a large number of categories based on a much smaller number of training examples.
The paper’s authors use a simple machine learning algorithm called k-nearest neighbors (kNN) to explore this idea more in depth. The algorithm operates under the assumption that similar things are most likely to exist near each other; if you go to a dog park, there will be lots of dogs but no cats or kangaroos. Go to the Australian grasslands and there’ll be kangaroos but no cats or dogs. And so on.
To train a kNN algorithm to differentiate between categories, you choose specific features to represent each category (i.e. for animals you could use weight or size as a feature). With one feature on the x-axis and the other on the y-axis, the algorithm creates a graph where data points that are similar to each other are clustered near each other. A line down the center divides the categories, and it’s pretty straightforward for the algorithm to discern which side of the line new data points should fall on.
The Waterloo team kept it simple and used plots of color on a 2D graph. Using the colors and their locations on the graphs, the team created synthetic data sets and accompanying soft labels. One of the more simplistic graphs is pictured below, along with soft labels in the form of pie charts.
Image Credit: Ilia Sucholutsky & Matthias Schonlau
When the team had the algorithm plot the boundary lines of the different colors based on these soft labels, it was able to split the plot up into more colors than the number of data points it was given in the soft labels.
While the results are encouraging, the team acknowledges that they’re just the first step, and there’s much more exploration of this concept yet to be done. The kNN algorithm is one of the least complex models out there; what might happen when LO-shot learning is applied to a far more complex algorithm? Also, to apply it, you still need to distill a larger dataset down into soft labels.
One idea the team is already working on is having other algorithms generate the soft labels for the algorithm that’s going to be trained using LO-shot; manually designing soft labels won’t always be as easy as splitting up some pie charts into different colors.
LO-shot’s potential for reducing the amount of training data needed to yield working AI systems is promising. Besides reducing the cost and the time required to train new models, the method could also make AI more accessible to industries, companies, or individuals who don’t have access to large datasets—an important step for democratization of AI.
Image Credit: pen_ash from Pixabay Continue reading
Posted in Human Robots
#437373 Microsoft’s New Deepfake Detector Puts ...
The upcoming US presidential election seems set to be something of a mess—to put it lightly. Covid-19 will likely deter millions from voting in person, and mail-in voting isn’t shaping up to be much more promising. This all comes at a time when political tensions are running higher than they have in decades, issues that shouldn’t be political (like mask-wearing) have become highly politicized, and Americans are dramatically divided along party lines.
So the last thing we need right now is yet another wrench in the spokes of democracy, in the form of disinformation; we all saw how that played out in 2016, and it wasn’t pretty. For the record, disinformation purposely misleads people, while misinformation is simply inaccurate, but without malicious intent. While there’s not a ton tech can do to make people feel safe at crowded polling stations or up the Postal Service’s budget, tech can help with disinformation, and Microsoft is trying to do so.
On Tuesday the company released two new tools designed to combat disinformation, described in a blog post by VP of Customer Security and Trust Tom Burt and Chief Scientific Officer Eric Horvitz.
The first is Microsoft Video Authenticator, which is made to detect deepfakes. In case you’re not familiar with this wicked byproduct of AI progress, “deepfakes” refers to audio or visual files made using artificial intelligence that can manipulate peoples’ voices or likenesses to make it look like they said things they didn’t. Editing a video to string together words and form a sentence someone didn’t say doesn’t count as a deepfake; though there’s manipulation involved, you don’t need a neural network and you’re not generating any original content or footage.
The Authenticator analyzes videos or images and tells users the percentage chance that they’ve been artificially manipulated. For videos, the tool can even analyze individual frames in real time.
Deepfake videos are made by feeding hundreds of hours of video of someone into a neural network, “teaching” the network the minutiae of the person’s voice, pronunciation, mannerisms, gestures, etc. It’s like when you do an imitation of your annoying coworker from accounting, complete with mimicking the way he makes every sentence sound like a question and his eyes widen when he talks about complex spreadsheets. You’ve spent hours—no, months—in his presence and have his personality quirks down pat. An AI algorithm that produces deepfakes needs to learn those same quirks, and more, about whoever the creator’s target is.
Given enough real information and examples, the algorithm can then generate its own fake footage, with deepfake creators using computer graphics and manually tweaking the output to make it as realistic as possible.
The scariest part? To make a deepfake, you don’t need a fancy computer or even a ton of knowledge about software. There are open-source programs people can access for free online, and as far as finding video footage of famous people—well, we’ve got YouTube to thank for how easy that is.
Microsoft’s Video Authenticator can detect the blending boundary of a deepfake and subtle fading or greyscale elements that the human eye may not be able to see.
In the blog post, Burt and Horvitz point out that as time goes by, deepfakes are only going to get better and become harder to detect; after all, they’re generated by neural networks that are continuously learning from and improving themselves.
Microsoft’s counter-tactic is to come in from the opposite angle, that is, being able to confirm beyond doubt that a video, image, or piece of news is real (I mean, can McDonald’s fries cure baldness? Did a seal slap a kayaker in the face with an octopus? Never has it been so imperative that the world know the truth).
A tool built into Microsoft Azure, the company’s cloud computing service, lets content producers add digital hashes and certificates to their content, and a reader (which can be used as a browser extension) checks the certificates and matches the hashes to indicate the content is authentic.
Finally, Microsoft also launched an interactive “Spot the Deepfake” quiz it developed in collaboration with the University of Washington’s Center for an Informed Public, deepfake detection company Sensity, and USA Today. The quiz is intended to help people “learn about synthetic media, develop critical media literacy skills, and gain awareness of the impact of synthetic media on democracy.”
The impact Microsoft’s new tools will have remains to be seen—but hey, we’re glad they’re trying. And they’re not alone; Facebook, Twitter, and YouTube have all taken steps to ban and remove deepfakes from their sites. The AI Foundation’s Reality Defender uses synthetic media detection algorithms to identify fake content. There’s even a coalition of big tech companies teaming up to try to fight election interference.
One thing is for sure: between a global pandemic, widespread protests and riots, mass unemployment, a hobbled economy, and the disinformation that’s remained rife through it all, we’re going to need all the help we can get to make it through not just the election, but the rest of the conga-line-of-catastrophes year that is 2020.
Image Credit: Darius Bashar on Unsplash Continue reading
Posted in Human Robots
#437269 DeepMind’s Newest AI Programs Itself ...
When Deep Blue defeated world chess champion Garry Kasparov in 1997, it may have seemed artificial intelligence had finally arrived. A computer had just taken down one of the top chess players of all time. But it wasn’t to be.
Though Deep Blue was meticulously programmed top-to-bottom to play chess, the approach was too labor-intensive, too dependent on clear rules and bounded possibilities to succeed at more complex games, let alone in the real world. The next revolution would take a decade and a half, when vastly more computing power and data revived machine learning, an old idea in artificial intelligence just waiting for the world to catch up.
Today, machine learning dominates, mostly by way of a family of algorithms called deep learning, while symbolic AI, the dominant approach in Deep Blue’s day, has faded into the background.
Key to deep learning’s success is the fact the algorithms basically write themselves. Given some high-level programming and a dataset, they learn from experience. No engineer anticipates every possibility in code. The algorithms just figure it.
Now, Alphabet’s DeepMind is taking this automation further by developing deep learning algorithms that can handle programming tasks which have been, to date, the sole domain of the world’s top computer scientists (and take them years to write).
In a paper recently published on the pre-print server arXiv, a database for research papers that haven’t been peer reviewed yet, the DeepMind team described a new deep reinforcement learning algorithm that was able to discover its own value function—a critical programming rule in deep reinforcement learning—from scratch.
Surprisingly, the algorithm was also effective beyond the simple environments it trained in, going on to play Atari games—a different, more complicated task—at a level that was, at times, competitive with human-designed algorithms and achieving superhuman levels of play in 14 games.
DeepMind says the approach could accelerate the development of reinforcement learning algorithms and even lead to a shift in focus, where instead of spending years writing the algorithms themselves, researchers work to perfect the environments in which they train.
Pavlov’s Digital Dog
First, a little background.
Three main deep learning approaches are supervised, unsupervised, and reinforcement learning.
The first two consume huge amounts of data (like images or articles), look for patterns in the data, and use those patterns to inform actions (like identifying an image of a cat). To us, this is a pretty alien way to learn about the world. Not only would it be mind-numbingly dull to review millions of cat images, it’d take us years or more to do what these programs do in hours or days. And of course, we can learn what a cat looks like from just a few examples. So why bother?
While supervised and unsupervised deep learning emphasize the machine in machine learning, reinforcement learning is a bit more biological. It actually is the way we learn. Confronted with several possible actions, we predict which will be most rewarding based on experience—weighing the pleasure of eating a chocolate chip cookie against avoiding a cavity and trip to the dentist.
In deep reinforcement learning, algorithms go through a similar process as they take action. In the Atari game Breakout, for instance, a player guides a paddle to bounce a ball at a ceiling of bricks, trying to break as many as possible. When playing Breakout, should an algorithm move the paddle left or right? To decide, it runs a projection—this is the value function—of which direction will maximize the total points, or rewards, it can earn.
Move by move, game by game, an algorithm combines experience and value function to learn which actions bring greater rewards and improves its play, until eventually, it becomes an uncanny Breakout player.
Learning to Learn (Very Meta)
So, a key to deep reinforcement learning is developing a good value function. And that’s difficult. According to the DeepMind team, it takes years of manual research to write the rules guiding algorithmic actions—which is why automating the process is so alluring. Their new Learned Policy Gradient (LPG) algorithm makes solid progress in that direction.
LPG trained in a number of toy environments. Most of these were “gridworlds”—literally two-dimensional grids with objects in some squares. The AI moves square to square and earns points or punishments as it encounters objects. The grids vary in size, and the distribution of objects is either set or random. The training environments offer opportunities to learn fundamental lessons for reinforcement learning algorithms.
Only in LPG’s case, it had no value function to guide that learning.
Instead, LPG has what DeepMind calls a “meta-learner.” You might think of this as an algorithm within an algorithm that, by interacting with its environment, discovers both “what to predict,” thereby forming its version of a value function, and “how to learn from it,” applying its newly discovered value function to each decision it makes in the future.
Prior work in the area has had some success, but according to DeepMind, LPG is the first algorithm to discover reinforcement learning rules from scratch and to generalize beyond training. The latter was particularly surprising because Atari games are so different from the simple worlds LPG trained in—that is, it had never seen anything like an Atari game.
Time to Hand Over the Reins? Not Just Yet
LPG is still behind advanced human-designed algorithms, the researchers said. But it outperformed a human-designed benchmark in training and even some Atari games, which suggests it isn’t strictly worse, just that it specializes in some environments.
This is where there’s room for improvement and more research.
The more environments LPG saw, the more it could successfully generalize. Intriguingly, the researchers speculate that with enough well-designed training environments, the approach might yield a general-purpose reinforcement learning algorithm.
At the least, though, they say further automation of algorithm discovery—that is, algorithms learning to learn—will accelerate the field. In the near term, it can help researchers more quickly develop hand-designed algorithms. Further out, as self-discovered algorithms like LPG improve, engineers may shift from manually developing the algorithms themselves to building the environments where they learn.
Deep learning long ago left Deep Blue in the dust at games. Perhaps algorithms learning to learn will be a winning strategy in the real world too.
Image credit: Mike Szczepanski / Unsplash Continue reading
Posted in Human Robots