Tag Archives: kind
What’s the world’s hardest machine learning problem? Autonomous vehicles? Robots that can walk? Cancer detection?
Nope, says Julian Sanchez. It’s agriculture.
Sanchez might be a little biased. He is the director of precision agriculture for John Deere, and is in charge of adding intelligence to traditional farm vehicles. But he does have a little perspective, having spent time working on software for both medical devices and air traffic control systems.
I met with Sanchez and Alexey Rostapshov, head of digital innovation at John Deere Labs, at the organization’s San Francisco offices last month. Labs launched in 2017 to take advantage of the area’s tech expertise, both to apply machine learning to in-house agricultural problems and to work with partners to build technologies that play nicely with Deere’s big green machines. Deere’s neighbors in San Francisco’s tech-heavy South of Market are LinkedIn, Salesforce, and Planet Labs, which puts it in a good position for recruiting.
“We’ve literally had folks knock on the door and say, ‘What are you doing here?’” says Rostapshov, and some return to drop off resumes.
Here’s why Sanchez believes agriculture is such a big challenge for artificial intelligence.
“It’s not just about driving tractors around,” he says, although autonomous driving technologies are part of the mix. (John Deere is doing a lot of work with precision GPS to improve autonomous driving, for example, and allow tractors to plan their own routes around fields.)
But more complex than the driving problem, says Sanchez, are the classification problems.
Corn: A Classic Classification Problem
Photo: Tekla Perry
One key effort, Sanchez says, are AI systems “that allow me to tell whether grain being harvested is good quality or low quality and to make automatic adjustment systems for the harvester.” The company is already selling an early version of this image analysis technology. But the many differences between grain types, and grains grown under different conditions, make this task a tough one for machine learning.
“Take corn,” Sanchez says. “Let’s say we are building a deep learning algorithm to detect this corn. And we take lots of pictures of kernels to give it. Say we pick those kernels in central Illinois. But, one mile over, the farmer planted a slightly different hybrid which has slightly different coloration of yellow. Meanwhile, this other farm harvested three days later in a field five miles away; it’s the same hybrid, but it also looks different.
“It’s an overwhelming classification challenge, and that’s just for corn. But you are not only doing it for corn, you have to add 20 more varieties of grain to the mix; and some, like canola, are almost microscopic.”
Even the ground conditions vary dramatically—far more than road conditions, Sanchez points out.
“Let’s say we are building a deep learning algorithm to detect how much residue is left on the soil after a harvest, including stubble and some chaff. Let’s drive 2,000 acres of fields in the Midwest looking at residue. That’s great, but I guarantee that if you go drive those the next year, it will look significantly different.
“Deep learning is great at interpolating conditions between what it knows; it is not good at extrapolating to situations it hasn’t seen. And in agriculture, you always feel that there is a set of conditions that you haven’t yet classified.”
A Flood of Big Data
The scale of the data is also daunting, Rostapshov points out. “We are one of the largest users of cloud computing services in the world,” he says. “We are gathering 5 to 15 million measurements per second from 130,000 connected machines globally. We have over 150 million acres in our databases, using petabytes and petabytes [of storage]. We process more data than Twitter does.”
Much of this information is so-called dirty data, that is, it doesn’t share the same format or structure, because it’s coming not only from a wide variety of John Deere machines, but also includes data from some 100 other companies that have access to the platform, including weather information, aerial imagery, and soil analyses.
As a result, says Sanchez, Deere has had to make “tremendous investments in back-end data cleanup.”
Deep learning is great at interpolating conditions between what it knows; it is not good at extrapolating to situations it hasn’t seen.”
—Julian Sanchez, John Deere
“We have gotten progressively more skilled at that problem,” he says. “We started simply by cleaning up our own data. You’d think it would be nice and neat, since it’s coming from our own machines, but there is a wide variety of different models and different years. Then we started geospatially tagging the agronomic data—the information about where you are applying herbicides and fertilizer and the like—coming in from our vehicles. When we started bringing in other data, from drones, say, we were already good at cleaning it up.”
John Deere’s Hiring Pitch
Hard problems can be a good thing to have for a company looking to hire machine learning engineers.
“Our opening line to potential recruits,” Sanchez says, “is ‘This stuff matters.’ Then, if we get a chance to talk to them more, we follow up with ‘Not only does this stuff matter, but the problems are really hard and interesting.’ When we explain the variability in farming and how we have to apply all the latest tools to these problems, we get their attention.”
Software engineers “know that feeding a growing population is a massive problem and are excited about the prospect of making a difference,” Rostapshov says.
Only 20 engineers work in the San Francisco labs right now, and that’s on a busy day—some of the researchers spend part of their time at Blue River Technology, a startup based in Sunnyvale that was acquired by Deere in 2017. About half of the researchers are focusing on AI. The Lab is in the process of doubling its office space (no word on staffing plans for that expansion yet).
“We are one of the largest users of cloud computing services in the world.”
—Alexey Rostapshov, John Deere Labs
Company-wide, Deere has thousands of software engineers, with many using AI and machine learning tools in their work, and about the same number of mechanical and electrical engineers, Sanchez reports. “If you look at our hiring 10 years ago,” he says, “it was heavily weighted to mechanical engineers. But if you look at those numbers now, it is by a large majority [engineers working] in the software space. We still need mechanical engineers—we do build green machines—but if you go by our footprint of tech talent, it is pretty safe to call John Deere a software company. And if you follow the key conversations that are happening in the company right now, 95 percent of them are software-related.”
For now, these software engineers are focused on developing technologies that allow farmers to “do more with less,” Sanchez says. Meaning, to get more and better crops from less fuel, less seed, less fertilizer, less pesticide, and fewer workers, and putting together building blocks that, he says, could eventually lead to fully autonomous farm vehicles. The data Deere collects today, for the most part, stays in silos (the virtual kind), with AI algorithms that analyze specific sets of data to provide guidance to individual farmers. At some point, however, with tools to anonymize data and buy-in from farmers, aggregating data could provide some powerful insights.
“We are not asking farmers for that yet,” Sanchez says. “We are not doing aggregation to look for patterns. We are focused on offering technology that allows an individual farmer to use less, on positioning ourselves to be in a neutral spot. We are not about selling you more seed or more fertilizer. So we are building up a good trust level. In the long term, we can have conversations about doing more with deep learning.” Continue reading
It’s been nearly six years since NASA unveiled Valkyrie, a state-of-the-art full-size humanoid robot. After the DARPA Robotics Challenge, NASA has continued to work with Valkyrie at Johnson Space Center, and has also provided Valkyrie robots to several different universities. Although it’s not a new platform anymore (six years is a long time in robotics), Valkyrie is still very capable, with plenty of potential for robotics research.
With that in mind, we were caught by surprise when over the last several months, Jacobs, a Dallas-based engineering company that appears to provide a wide variety of technical services to anyone who wants them, has posted several open jobs in need of roboticists in the Houston, Texas, area who are interested in working with NASA on “the next generation of humanoid robot.”
Here are the relevant bullet points from the one of the job descriptions (which you can view at this link):
Work directly with NASA Johnson Space Center in designing the next generation of humanoid robot.
Join the Valkyrie humanoid robot team in NASA’s Robotic Systems Technology Branch.
Build on the success of the existing Valkyrie and Robonaut 2 humanoid robots and advance NASA’s ability to project a remote human presence and dexterous manipulation capability into challenging, dangerous, and distant environments both in space and here on earth.
The question is, why is NASA developing its own humanoid robot (again) when it could instead save a whole bunch of time and money by using a platform that already exists, whether it’s Atlas, Digit, Valkyrie itself, or one of the small handful of other humanoids that are more or less available? The only answer that I can come up with is that no existing platforms meet NASA’s requirements, whatever those may be. And if that’s the case, what kind of requirements are we talking about? The obvious one would be the ability to work in the kinds of environments that NASA specializes in—space, the Moon, and Mars.
Artist’s concept of NASA’s Valkyrie humanoid robot working on the surface of Mars.
NASA’s existing humanoid robots, including Robonaut 2 and Valkyrie, were designed to operate on Earth. Robonaut 2 ended up going to space anyway (it’s recently returned to Earth for repairs), but its hardware was certainly never intended to function outside of the International Space Station. Working in a vacuum involves designing for a much more rigorous set of environmental challenges, and things get even worse on the Moon or on Mars, where highly abrasive dust gets everywhere.
We know that it’s possible to design robots for long term operation in these kinds of environments because we’ve done it before. But if you’re not actually going to send your robot off-world, there’s very little reason to bother making sure that it can operate through (say) 300° Celsius temperature swings like you’d find on the Moon. In the past, NASA has quite sensibly focused on designing robots that can be used as platforms for the development of software and techniques that could one day be applied to off-world operations, without over-engineering those specific robots to operate in places that they would almost certainly never go. As NASA increasingly focuses on a return to the Moon, though, maybe it’s time to start thinking about a humanoid robot that could actually do useful stuff on the lunar surface.
Artist’s concept of the Gateway moon-orbiting space station (seen on the right) with an Orion crew vehicle approaching.
The other possibility that I can think of, and perhaps the more likely one, is that this next humanoid robot will be a direct successor to Robonaut 2, intended for NASA’s Gateway space station orbiting the Moon. Some of the robotics folks at NASA that we’ve talked to recently have emphasized how important robotics will be for Gateway:
Trey Smith, NASA Ames: Everybody at NASA is really excited about work on the Gateway space station that would be in near lunar space. We don’t have definite plans for what would happen on the Gateway yet, but there’s a general recognition that intra-vehicular robots are important for space stations. And so, it would not be surprising to see a mobile manipulator like Robonaut, and a free flyer like Astrobee, on the Gateway.
If you have an un-crewed cargo vehicle that shows up stuffed to the rafters with cargo bags and it docks with the Gateway when there’s no crew there, it would be very useful to have intra-vehicular robots that can pull all those cargo bags out, unpack them, stow all the items, and then even allow the cargo vehicle to detach before the crew show up so that the crew don’t have to waste their time with that.
Julia Badger, NASA JSC: One of the systems on board Gateway is going to be intravehicular robots. They’re not going to necessarily look like Robonaut, but they’ll have some of the same functionality as Robonaut—being mobile, being able to carry payloads from one part of the module to another, doing some dexterous manipulation tasks, inspecting behind panels, those sorts of things.
Artist’s concept of NASA’s Valkyrie humanoid robot working inside a spacecraft.
Since Gateway won’t be crewed by humans all of the time, it’ll be important to have a permanent robotic presence to keep things running while nobody is home while saving on resources by virtue of the fact that robots aren’t always eating food, drinking water, consuming oxygen, demanding that the temperature stays just so, and producing a variety of disgusting kinds of waste. Obviously, the robot won’t be as capable as humans, but if they can manage to do even basic continuing maintenance tasks (most likely through at least partial teleoperation), that would be very useful.
Photo: Evan Ackerman/IEEE Spectrum
NASA’s Robonaut team plans to perform a variety of mobility and motion-planning experiments using the robot’s new legs, which can grab handrails on the International Space Station.
As for whether robots designed for Gateway would really fall into the “humanoid” category, it’s worth considering that Gateway is designed for humans, implying that an effective robotic system on Gateway would need to be able to interact with the station in similar ways to how a human astronaut would. So, you’d expect to see arms with end-effectors that can grip things as well as push buttons, and some kind of mobility system—the legged version of Robonaut 2 seems like a likely template, but redesigned from the ground up to work in space, incorporating all the advances in robotics hardware and computing that have taken place over the last decade.
We’ve been pestering NASA about this for a little bit now, and they’re not ready to comment on this project, or even to confirm it. And again, everything in this article (besides the job post, which you should totally check out and consider applying for) is just speculation on our part, and we could be wrong about absolutely all of it. As soon as we hear more, we’ll definitely let you know. Continue reading
Engineers have built microrobots to perform all sorts of tasks in the body, and can now add to that list another key skill: delivering stem cells. In a paper published today in Science Robotics, researchers describe propelling a magnetically-controlled, stem-cell-carrying bot through a live mouse.
Under a rotating magnetic field, the microrobots moved with rolling and corkscrew-style locomotion. The researchers, led by Hongsoo Choi and his team at the Daegu Gyeongbuk Institute of Science & Technology (DGIST), in South Korea, also demonstrated their bot’s moves in slices of mouse brain, in blood vessels isolated from rat brains, and in a multi-organ-on-a chip.
The invention provides an alternative way to deliver stem cells, which are increasingly important in medicine. Such cells can be coaxed into becoming nearly any kind of cell, making them great candidates for treating neurodegenerative disorders such as Alzheimer’s.
But delivering stem cells typically requires an injection with a needle, which lowers the survival rate of the stem cells, and limits their reach in the body. Microrobots, however, have the potential to deliver stem cells to precise, hard-to-reach areas, with less damage to surrounding tissue, and better survival rates, says Jin-young Kim, a principle investigator at DGIST-ETH Microrobotics Research Center, and an author on the paper.
The virtues of microrobots have inspired several research groups to propose and test different designs in simple conditions, such as microfluidic channels and other static environments. A group out of Hong Kong last year described a burr-shaped bot that carried cells through live, transparent zebrafish.
The new research presents a magnetically-actuated microrobot that successfully carried stem cells through a live mouse. In additional experiments, the cells, which had differentiated into brain cells such as astrocytes, oligodendrocytes, and neurons, transferred to microtissues on the multi-organ-on-a-chip. Taken together, the proof-of-concept experiments demonstrate the potential for microrobots to be used in human stem cell therapy, says Kim.
The team fabricated the robots with 3D laser lithography, and designed them in two shapes: spherical and helical. Using a rotating magnetic field, the scientists navigated the spherical-shaped bots with a rolling motion, and the helical bots with a corkscrew motion. These styles of locomotion proved more efficient than that from a simple pulling force, and were more suitable for use in biological fluids, the scientists reported.
The big challenge in navigating microbots in a live animal (or human body) is being able to see them in real time. Imaging with fMRI doesn’t work, because the magnetic fields interfere with the system. “To precisely control microbots in vivo, it is important to actually see them as they move,” the authors wrote in their paper.
That wasn’t possible during experiments in a live mouse, so the researchers had to check the location of the microrobots before and after the experiments using an optical tomography system called IVIS. They also had to resort to using a pulling force with a permanent magnet to navigate the microrobots inside the mouse, due to the limitations of the IVIS system.
Kim says he and his colleagues are developing imaging systems that will enable them to view in real time the locomotion of their microrobots in live animals. Continue reading
At Amazon’s re:MARS conference in Las Vegas today, who else but Amazon is introducing two new robots designed to make its fulfillment centers even more fulfilling. Xanthus (named after a mythological horse that could very briefly talk but let’s not read too much into that) is a completely redesigned drive unit, one of the robotic mobile bases that carries piles of stuff around for humans to pick from. It has a thinner profile, a third of the parts, costs half as much, and can wear different modules on top to perform a much wider variety of tasks than its predecessor.
Pegasus (named after a mythological horse that could fly but let’s not read too much into that either) is also a mobile robot, but much smaller than Xanthus, designed to help the company quickly and accurately sort individual packages. For Amazon, it’s a completely new large-scale robotic system involving tightly coordinated fleets of robots tossing boxes down chutes, and it’s just as fun to watch as it sounds.
Amazon has 800 Pegasus units already deployed at a sorting facility in the United States, adding to their newly updated total of 200,000 robotic drive units worldwide.
If the Pegasus system looks familiar, it’s because other warehouse automation companies have had something that’s at least superficially very similar up and running for years.
Pegasus is one of Amazon’s new warehouse robots, equipped with a conveyor belt on top and used in the company’s sorting facilities.
But the most interesting announcement that Amazon made, kind of low key and right at the end of their re:MARS talk, is that they’re working on ways of making some of their mobile robots actually collaborative, leveraging some of the technology that they acquired from Boulder, Colo.-based warehouse robotics startup Canvas Technology earlier this year:
“With our recent acquisition of Canvas, we expect to be able to combine this drive platform with AI and autonomous mobility capabilities, and for the first time, allow our robots to move outside of our robotic drive fields, and interact collaboratively with our associates to do a number of mobility tasks,” said Brad Porter, VP of robotics at Amazon.
At the moment, Amazon’s robots are physically separated from humans except for one highly structured station where the human only interacts with the robot in one or two very specific ways. We were told a few months ago that Amazon would like to have mobile robots that are able to move things through the areas of fulfillment centers that have people in them, but that they’re (quite rightly) worried about the safety aspects of having robots and humans work around each other. Other companies are already doing this on a smaller scale, and it means developing a reliable safety system that can handle randomly moving humans, environmental changes, and all kinds of other stuff. It’s much more difficult than having a nice, clean, roped-off area to work in where a wayward human would be an exception rather than just another part of the job.
Photo: Canvas Technology
A robot created by Canvas Technology, a Boulder, Colo.-based warehouse robotics startup acquired by Amazon earlier this year.
It now seems like Canvas has provided the secret sauce that Amazon needed to start implementing this level of autonomy. As for what it’s going to look like, our best guess is that Amazon is going to have to do a little bit more than slap some extra sensors onto Xanthus or Pegasus, if for no other reason than the robots will almost certainly need more ground clearance to let them operate away from the reliably flat floors that they’re accustomed to. We’re expecting to see them performing many of the tasks that companies like Fetch Robotics and OTTO Motors are doing already—moving everything from small boxes to large pallets to keep humans from having to waste time walking.
Of course, this all feeds back into what drives Amazon more than anything else: efficiency. And for better or worse, humans are not uniquely good at moving things from place to place, so it’s no surprise that Amazon wants to automate that, too. The good news is that, at least for now, Amazon still needs humans to babysit all those robots.
[ Amazon ] Continue reading