Tag Archives: Deep learning

#437769 Q&A: Facebook’s CTO Is at War With ...

Posted on November 29, 2020 by Android

Photo: Patricia de Melo Moreira/AFP/Getty Images

Facebook chief technology officer Mike Schroepfer leads the company’s AI and integrity efforts.

Facebook’s challenge is huge. Billions of pieces of content—short and long posts, images, and combinations of the two—are uploaded to the site daily from around the world. And any tiny piece of that—any phrase, image, or video—could contain so-called bad content.

In its early days, Facebook relied on simple computer filters to identify potentially problematic posts by their words, such as those containing profanity. These automatically filtered posts, as well as posts flagged by users as offensive, went to humans for adjudication.

In 2015, Facebook started using artificial intelligence to cull images that contained nudity, illegal goods, and other prohibited content; those images identified as possibly problematic were sent to humans for further review.

By 2016, more offensive photos were reported by Facebook’s AI systems than by Facebook users (and that is still the case).

In 2018, Facebook CEO Mark Zuckerberg made a bold proclamation: He predicted that within five or ten years, Facebook’s AI would not only look for profanity, nudity, and other obvious violations of Facebook’s policies. The tools would also be able to spot bullying, hate speech, and other misuse of the platform, and put an immediate end to them.

Today, automated systems using algorithms developed with AI scan every piece of content between the time when a user completes a post and when it is visible to others on the site—just fractions of a second. In most cases, a violation of Facebook’s standards is clear, and the AI system automatically blocks the post. In other cases, the post goes to human reviewers for a final decision, a workforce that includes 15,000 content reviewers and another 20,000 employees focused on safety and security, operating out of more than 20 facilities around the world.

In the first quarter of this year, Facebook removed or took other action (like appending a warning label) on more than 9.6 million posts involving hate speech, 8.6 million involving child nudity or exploitation, almost 8 million posts involving the sale of drugs, 2.3 million posts involving bullying and harassment, and tens of millions of posts violating other Facebook rules.

Right now, Facebook has more than 1,000 engineers working on further developing and implementing what the company calls “integrity” tools. Using these systems to screen every post that goes up on Facebook, and doing so in milliseconds, is sucking up computing resources. Facebook chief technology officer Mike Schroepfer, who is heading up Facebook’s AI and integrity efforts, spoke with IEEE Spectrum about the team’s progress on building an AI system that detects bad content.

Since that discussion, Facebook’s policies around hate speech have come under increasing scrutiny, with particular attention on divisive posts by political figures. A group of major advertisers in June announced that they would stop advertising on the platform while reviewing the situation, and civil rights groups are putting pressure on others to follow suit until Facebook makes policy changes related to hate speech and groups that promote hate, misinformation, and conspiracies.

Facebook CEO Mark Zuckerberg responded with news that Facebook will widen the category of what it considers hateful content in ads. Now the company prohibits claims that people from a specific race, ethnicity, national origin, religious affiliation, caste, sexual orientation, gender identity, or immigration status are a threat to the physical safety, health, or survival of others. The policy change also aims to better protect immigrants, migrants, refugees, and asylum seekers from ads suggesting these groups are inferior or expressing contempt. Finally, Zuckerberg announced that the company will label some problematic posts by politicians and government officials as content that violates Facebook’s policies.

However, civil rights groups say that’s not enough. And an independent audit released in July also said that Facebook needs to go much further in addressing civil rights concerns and disinformation.

Schroepfer indicated that Facebook’s AI systems are designed to quickly adapt to changes in policy. “I don’t expect considerable technical changes are needed to adjust,” he told Spectrum.

This interview has been edited and condensed for clarity.

IEEE Spectrum: What are the stakes of content moderation? Is this an existential threat to Facebook? And is it critical that you deal well with the issue of election interference this year?

Schroepfer: It’s probably existential; it’s certainly massive. We are devoting a tremendous amount of our attention to it.

The idea that anyone could meddle in an election is deeply disturbing and offensive to all of us here, just as people and citizens of democracies. We don’t want to see that happen anywhere, and certainly not on our watch. So whether it’s important to the company or not, it’s important to us as people. And I feel a similar way on the content-moderation side.

There are not a lot of easy choices here. The only way to prevent people, with certainty, from posting bad things is to not let them post anything. We can take away all voice and just say, “Sorry, the Internet’s too dangerous. No one can use it.” That will certainly get rid of all hate speech online. But I don’t want to end up in that world. And there are variants of that world that various governments are trying to implement, where they get to decide what’s true or not, and you as a person don’t. I don’t want to get there either.

My hope is that we can build a set of tools that make it practical for us to do a good enough job, so that everyone is still excited about the idea that anyone can share what they want, and so that Facebook is a safe and reasonable place for people to operate in.

Spectrum: You joined Facebook in 2008, before AI was part of the company’s toolbox. When did that change? When did you begin to think that AI tools would be useful to Facebook?

Schroepfer: Ten years ago, AI wasn’t commercially practical; the technology just didn’t work very well. In 2012, there was one of those moments that a lot of people point to as the beginning of the current revolution in deep learning and AI. A computer-vision model—a neural network—was trained using what we call supervised training, and it turned out to be better than all the existing models.

Spectrum: How is that training done, and how did computer-vision models come to Facebook?

Image: Facebook

Just Broccoli? Facebook’s image analysis algorithms can tell the difference between marijuana [left] and tempura broccoli [right] better than some humans.

Schroepfer: Say I take a bunch of photos and I have people look at them. If they see a photo of a cat, they put a text label that says cat; if it’s one of a dog, the text label says dog. If you build a big enough data set and feed that to the neural net, it learns how to tell the difference between cats and dogs.

Prior to 2012, it didn’t work very well. And then in 2012, there was this moment where it seemed like, “Oh wow, this technique might work.” And a few years later we were deploying that form of technology to help us detect problematic imagery.

Spectrum: Do your AI systems work equally well on all types of prohibited content?

Schroepfer: Nudity was technically easiest. I don’t need to understand language or culture to understand that this is either a naked human or not. Violence is a much more nuanced problem, so it was harder technically to get it right. And with hate speech, not only do you have to understand the language, it may be very contextual, even tied to recent events. A week before the Christchurch shooting [New Zealand, 2019], saying “I wish you were in the mosque” probably doesn’t mean anything. A week after, that might be a terrible thing to say.

Spectrum: How much progress have you made on hate speech?

Schroepfer: AI, in the first quarter of 2020, proactively detected 88.8 percent of the hate-speech content we removed, up from 80.2 percent in the previous quarter. In the first quarter of 2020, we took action on 9.6 million pieces of content for violating our hate-speech policies.

Image: Facebook

Off Label: Sometimes image analysis isn’t enough to determine whether a picture posted violates the company’s policies. In considering these candy-colored vials of marijuana, for example, the algorithms can look at any accompanying text and, if necessary, comments on the post.

Spectrum: It sounds like you’ve expanded beyond tools that analyze images and are also using AI tools that analyze text.

Schroepfer: AI started off as very siloed. People worked on language, people worked on computer vision, people worked on video. We’ve put these things together—in production, not just as research—into multimodal classifiers.

[Schroepfer shows a photo of a pan of Rice Krispies treats, with text referring to it as a “potent batch”] This is a case in which you have an image, and then you have the text on the post. This looks like Rice Krispies. On its own, this image is fine. You put the text together with it in a bigger model; that can then understand what’s going on. That didn’t work five years ago.

Spectrum: Today, every post that goes up on Facebook is immediately checked by automated systems. Can you explain that process?

Image: Facebook

Bigger Picture: Identifying hate speech is often a matter of context. Either the text or the photo in this post isn’t hateful standing alone, but putting them together tells a different story.

Schroepfer: You upload an image and you write some text underneath it, and the systems look at both the image and the text to try to see which, if any, policies it violates. Those decisions are based on our Community Standards. It will also look at other signals on the posts, like the comments people make.

It happens relatively instantly, though there may be times things happen after the fact. Maybe you uploaded a post that had misinformation in it, and at the time you uploaded it, we didn’t know it was misinformation. The next day we fact-check something and scan again; we may find your post and take it down. As we learn new things, we’re going to go back through and look for violations of what we now know to be a problem. Or, as people comment on your post, we might update our understanding of it. If people are saying, “That’s terrible,” or “That’s mean,” or “That looks fake,” those comments may be an interesting signal.

Spectrum: How is Facebook applying its AI tools to the problem of election interference?

Schroepfer: I would split election interference into two categories. There are times when you’re going after the content, and there are times you’re going after the behavior or the authenticity of the person.

On content, if you’re sharing misinformation, saying, “It’s super Wednesday, not super Tuesday, come vote on Wednesday,” that’s a problem whether you’re an American sitting in California or a foreign actor.

Other times, people create a series of Facebook pages pretending they’re Americans, but they’re really a foreign entity. That is a problem on its own, even if all the content they’re sharing completely meets our Community Standards. The problem there is that you have a foreign government running an information operation.

There, you need different tools. What you’re trying to do is put pieces together, to say, “Wait a second. All of these pages—Martians for Justice, Moonlings for Justice, and Venusians for Justice”—are all run by an administrator with an IP address that’s outside the United States. So they’re all connected, even though they’re pretending to not be connected. That’s a very different problem than me sitting in my office in Menlo Park [Calif.] sharing misinformation.

I’m not going to go into lots of technical detail, because this is an area of adversarial nature. The fundamental problem you’re trying to solve is that there’s one entity coordinating the activity of a bunch of things that look like they’re not all one thing. So this is a series of Instagram accounts, or a series of Facebook pages, or a series of WhatsApp accounts, and they’re pretending to be totally different things. We’re looking for signals that these things are related in some way. And we’re looking through the graph [what Facebook calls its map of relationships between users] to understand the properties of this network.

Spectrum: What cutting-edge AI tools and methods have you been working on lately?

Schroepfer: Supervised learning, with humans setting up the instruction process for the AI systems, is amazingly effective. But it has a very obvious flaw: the speed at which you can develop these things is limited by how fast you can curate the data sets. If you’re dealing in a problem domain where things change rapidly, you have to rebuild a new data set and retrain the whole thing.

Self-supervision is inspired by the way people learn, by the way kids explore the world around them. To get computers to do it themselves, we take a bunch of raw data and build a way for the computer to construct its own tests. For language, you scan a bunch of Web pages, and the computer builds a test where it takes a sentence, eliminates one of the words, and figures out how to predict what word belongs there. And because it created the test, it actually knows the answer. I can use as much raw text as I can find and store because it’s processing everything itself and doesn’t require us to sit down and build the information set. In the last two years there has been a revolution in language understanding as a result of AI self-supervised learning.

Spectrum: What else are you excited about?

Schroepfer: What we’ve been working on over the last few years is multilingual understanding. Usually, when I’m trying to figure out, say, whether something is hate speech or not I have to go through the whole process of training the model in every language. I have to do that one time for every language. When you make a post, the first thing we have to figure out is what language your post is in. “Ah, that’s Spanish. So send it to the Spanish hate-speech model.”

We’ve started to build a multilingual model—one box where you can feed in text in 40 different languages and it determines whether it’s hate speech or not. This is way more effective and easier to deploy.

To geek out for a second, just the idea that you can build a model that understands a concept in multiple languages at once is crazy cool. And it not only works for hate speech, it works for a variety of things.

When we started working on this multilingual model years ago, it performed worse than every single individual model. Now, it not only works as well as the English model, but when you get to the languages where you don’t have enough data, it’s so much better. This rapid progress is very exciting.

Spectrum: How do you move new AI tools from your research labs into operational use?

Schroepfer: Engineers trying to make the next breakthrough will often say, “Cool, I’ve got a new thing and it achieved state-of-the-art results on machine translation.” And we say, “Great. How long does it take to run in production?” They say, “Well, it takes 10 seconds for every sentence to run on a CPU.” And we say, “It’ll eat our whole data center if we deploy that.” So we take that state-of-the-art model and we make it 10 or a hundred or a thousand times more efficient, maybe at the cost of a little bit of accuracy. So it’s not as good as the state-of-the-art version, but it’s something we can actually put into our data centers and run in production.

Spectrum: What’s the role of the humans in the loop? Is it true that Facebook currently employs 35,000 moderators?

Schroepfer: Yes. Right now our goal is not to reduce that. Our goal is to do a better job catching bad content. People often think that the end state will be a fully automated system. I don’t see that world coming anytime soon.

As automated systems get more sophisticated, they take more and more of the grunt work away, freeing up the humans to work on the really gnarly stuff where you have to spend an hour researching.

We also use AI to give our human moderators power tools. Say I spot this new meme that is telling everyone to vote on Wednesday rather than Tuesday. I have a tool in front of me that says, “Find variants of that throughout the system. Find every photo with the same text, find every video that mentions this thing and kill it in one shot.” Rather than, I found this one picture, but then a bunch of other people upload that misinformation in different forms.

Another important aspect of AI is that anything I can do to prevent a person from having to look at terrible things is time well spent. Whether it’s a person employed by us as a moderator or a user of our services, looking at these things is a terrible experience. If I can build systems that take the worst of the worst, the really graphic violence, and deal with that in an automated fashion, that’s worth a lot to me. Continue reading →

Posted in Human Robots

#437751 Startup and Academics Find Path to ...

Posted on November 28, 2020 by Android

Engineers have been chasing a form of AI that could drastically lower the energy required to do typical AI things like recognize words and images. This analog form of machine learning does one of the key mathematical operations of neural networks using the physics of a circuit instead of digital logic. But one of the main things limiting this approach is that deep learning’s training algorithm, back propagation, has to be done by GPUs or other separate digital systems.

Now University of Montreal AI expert Yoshua Bengio, his student Benjamin Scellier, and colleagues at startup Rain Neuromorphics have come up with way for analog AIs to train themselves. That method, called equilibrium propagation, could lead to continuously learning, low-power analog systems of a far greater computational ability than most in the industry now consider possible, according to Rain CTO Jack Kendall.

Analog circuits could save power in neural networks in part because they can efficiently perform a key calculation, called multiply and accumulate. That calculation multiplies values from inputs according to various weights, and then it sums all those values up. Two fundamental laws of electrical engineering can basically do that, too. Ohm’s Law multiplies voltage and conductance to give current, and Kirchoff’s Current Law sums the currents entering a point. By storing a neural network’s weights in resistive memory devices, such as memristors, multiply-and-accumulate can happen completely in analog, potentially reducing power consumption by orders of magnitude.

The reason analog AI systems can’t train themselves today has a lot to do with the variability of their components. Just like real neurons, those in analog neural networks don’t all behave exactly alike. To do back propagation with analog components, you must build two separate circuit pathways. One going forward to come up with an answer (called inferencing), the other going backward to do the learning so that the answer becomes more accurate. But because of the variability of analog components, the pathways don't match up.

“You end up accumulating error as you go backwards through the network,” says Bengio. To compensate, a network would need lots of power-hungry analog-to-digital and digital-to-analog circuits, defeating the point of going analog.

Equilibrium propagation allows learning and inferencing to happen on the same network, partly by adjusting the behavior of the network as a whole. “What [equilibrium propagation] allows us to do is to say how we should modify each of these devices so that the overall circuit performs the right thing,” he says. “We turn the physical computation that is happening in the analog devices directly to our advantage.”

Right now, equilibrium propagation is only working in simulation. But Rain plans to have a hardware proof-of-principle in late 2021, according to CEO and cofounder Gordon Wilson. “We are really trying to fundamentally reimagine the hardware computational substrate for artificial intelligence, find the right clues from the brain, and use those to inform the design of this,” he says. The result could be what they call end-to-end analog AI systems that capable of running sophisticated robots or even playing a role in data centers. Both of those are currently considered beyond the capabilities of analog AI, which is now focused only on adding inferencing abilities to sensors and other low-power “edge” devices, while leaving the learning to GPUs. Continue reading →

Posted in Human Robots

#437695 Video Friday: Even Robots Know That You ...

Posted on November 25, 2020 by Android

Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We’ll also be posting a weekly calendar of upcoming robotics events for the next few months; here's what we have so far (send us your events!):

CLAWAR 2020 – August 24-26, 2020 – [Online Conference]
Other Than Human – September 3-10, 2020 – Stockholm, Sweden
ICRES 2020 – September 28-29, 2020 – Taipei, Taiwan
AUVSI EXPONENTIAL 2020 – October 5-8, 2020 – [Online Conference]
IROS 2020 – October 25-29, 2020 – Las Vegas, Nev., USA
CYBATHLON 2020 – November 13-14, 2020 – [Online Event]
ICSR 2020 – November 14-16, 2020 – Golden, Colo., USA
Let us know if you have suggestions for next week, and enjoy today's videos.

From the Robotics and Perception Group at UZH comes Flightmare, a simulation environment for drones that combines a slick rendering engine with a robust physics engine that can run as fast as your system can handle.

Flightmare is composed of two main components: a configurable rendering engine built on Unity and a flexible physics engine for dynamics simulation. Those two components are totally decoupled and can run independently from each other. Flightmare comes with several desirable features: (i) a large multi-modal sensor suite, including an interface to extract the 3D point-cloud of the scene; (ii) an API for reinforcement learning which can simulate hundreds of quadrotors in parallel; and (iii) an integration with a virtual-reality headset for interaction with the simulated environment. Flightmare can be used for various applications, including path-planning, reinforcement learning, visual-inertial odometry, deep learning, human-robot interaction, etc.

[ Flightmare ]

Quadruped robots yelling at people to maintain social distancing is really starting to become a thing, for better or worse.

We introduce a fully autonomous surveillance robot based on a quadruped platform that can promote social distancing in complex urban environments. Specifically, to achieve autonomy, we mount multiple cameras and a 3D LiDAR on the legged robot. The robot then uses an onboard real-time social distancing detection system to track nearby pedestrian groups. Next, the robot uses a crowd-aware navigation algorithm to move freely in highly dynamic scenarios. The robot finally uses a crowd aware routing algorithm to effectively promote social distancing by using human-friendly verbal cues to send suggestions to overcrowded pedestrians.

[ Project ]

Thanks Fan!

The Personal Robotics Group at Oregon State University is looking at UV germicidal irradiation for surface disinfection with a Fetch Manipulator Robot.

Fetch Robot disinfecting dance party woo!

[ Oregon State ]

How could you not take a mask from this robot?

[ Reachy ]

This work presents the design, development and autonomous navigation of the alpha-version of our Resilient Micro Flyer, a new type of collision-tolerant small aerial robot tailored to traversing and searching within highly confined environments including manhole-sized tubes. The robot is particularly lightweight and agile, while it implements a rigid collision-tolerant design which renders it resilient during forcible interaction with the environment. Furthermore, the design of the system is enhanced through passive flaps ensuring smoother and more compliant collision which was identified to be especially useful in very confined settings.

[ ARL ]

Pepper can make maps and autonomously navigate, which is interesting, but not as interesting as its posture when it's wandering around.

Dat backing into the charging dock tho.

[ Pepper ]

RatChair a strategy for displacing big objects by attaching relatively small vibration sources. After learning how several random bursts of vibration affect its pose, an optimization algorithm discovers the optimal sequence of vibration patterns required to (slowly but surely) move the object to a specified position.

This is from 2015, why isn't all of my furniture autonomous yet?!

[ KAIST ]

The new SeaDrone Pro is designed to be the underwater equivalent of a quadrotor. This video is a rendering, but we've been assured that it does actually exist.

[ SeaDrone ]

Thanks Eduardo!

Porous Loops is a lightweight composite facade panel that shows the potential of 3D printing of mineral foams for building scale applications.

[ ETH ]

Thanks Fan!

Here's an interesting idea for a robotic gripper- it's what appears to be a snap bracelet coupled to a pneumatic actuator that allows the snap bracelet to be reset.

[ Georgia Tech ]

Graze is developing a commercial robotic lawnmower. They're also doing a sort of crowdfunded investment thing, which probably explains the painfully overproduced nature of the following video:

A couple things about this: the hard part, which the video skips over almost entirely, is the mapping, localization, and understanding where to mow and where not to mow. The pitch deck seems to suggest that this is mostly done through computer vision, a thing that's perhaps easy to do under controlled ideal conditions, but difficult to apply to a world full lawns that are all different. The commercial aspect is interesting because golf courses are likely as standardized as you can get, but the emphasis here on how much money they can make without really addressing any of the technical stuff makes me raise an eyebrow or two.

[ Graze ]

The record & playback X-series arm demo allows the user to record the arm's movements while motors are torqued off. Then, the user may torque the motor's on and watch the movements they just made playback!

[ Interbotix ]

Shadow Robot has a new teleop system for its hand. I'm guessing that it's even trickier to use than it looks.

[ Shadow Robot ]

Quanser Interactive Labs is a collection of virtual hardware-based laboratory activities that supplement traditional or online courses. Same as working with physical systems in the lab, students work with virtual twins of Quanser's most popular plants, develop their mathematical models, implement and simulate the dynamic behavior of these systems, design controllers, and validate them on a high-fidelity 3D real-time virtual models. The virtual systems not only look like the real ones, they also behave, can be manipulated, measured, and controlled like real devices. And finally, when students go to the lab, they can deploy their virtually-validated designs on actual physical equipment.

[ Quanser ]

This video shows robot-assisted heart surgery. It's amazing to watch if you haven't seen this sort of thing before, but be aware that there is a lot of blood.

This video demonstrates a fascinating case of robotic left atrial myxoma excision, narrated by Joel Dunning, Middlesbrough, UK. The Robotic platform provides superior visualisation and enhanced dexterity, through keyhole incisions. Robotic surgery is an integral part of our Minimally Invasive Cardiothoracic Surgery Program.

[ Tristan D. Yan ]

Thanks Fan!

In this talk, we present our work on learning control policies directly in simulation that are deployed onto real drones without any fine tuning. The presentation covers autonomous drone racing, drone acrobatics, and uncertainty estimation in deep networks.

[ RPG ] Continue reading →

Posted in Human Robots

#437620 The Trillion-Transistor Chip That Just ...

Posted on November 22, 2020 by Android

The history of computer chips is a thrilling tale of extreme miniaturization.

The smaller, the better is a trend that’s given birth to the digital world as we know it. So, why on earth would you want to reverse course and make chips a lot bigger? Well, while there’s no particularly good reason to have a chip the size of an iPad in an iPad, such a chip may prove to be genius for more specific uses, like artificial intelligence or simulations of the physical world.

At least, that’s what Cerebras, the maker of the biggest computer chip in the world, is hoping.

The Cerebras Wafer-Scale Engine is massive any way you slice it. The chip is 8.5 inches to a side and houses 1.2 trillion transistors. The next biggest chip, NVIDIA’s A100 GPU, measures an inch to a side and has a mere 54 billion transistors. The former is new, largely untested and, so far, one-of-a-kind. The latter is well-loved, mass-produced, and has taken over the world of AI and supercomputing in the last decade.

So can Goliath flip the script on David? Cerebras is on a mission to find out.

Big Chips Beyond AI
When Cerebras first came out of stealth last year, the company said it could significantly speed up the training of deep learning models.

Since then, the WSE has made its way into a handful of supercomputing labs, where the company’s customers are putting it through its paces. One of those labs, the National Energy Technology Laboratory, is looking to see what it can do beyond AI.

So, in a recent trial, researchers pitted the chip—which is housed in an all-in-one system about the size of a dorm room mini-fridge called the CS-1—against a supercomputer in a fluid dynamics simulation. Simulating the movement of fluids is a common supercomputer application useful for solving complex problems like weather forecasting and airplane wing design.

The trial was described in a preprint paper written by a team led by Cerebras’s Michael James and NETL’s Dirk Van Essendelft and presented at the supercomputing conference SC20 this week. The team said the CS-1 completed a simulation of combustion in a power plant roughly 200 times faster than it took the Joule 2.0 supercomputer to do a similar task.

The CS-1 was actually faster-than-real-time. As Cerebrus wrote in a blog post, “It can tell you what is going to happen in the future faster than the laws of physics produce the same result.”

The researchers said the CS-1’s performance couldn’t be matched by any number of CPUs and GPUs. And CEO and cofounder Andrew Feldman told VentureBeat that would be true “no matter how large the supercomputer is.” At a point, scaling a supercomputer like Joule no longer produces better results in this kind of problem. That’s why Joule’s simulation speed peaked at 16,384 cores, a fraction of its total 86,400 cores.

A comparison of the two machines drives the point home. Joule is the 81st fastest supercomputer in the world, takes up dozens of server racks, consumes up to 450 kilowatts of power, and required tens of millions of dollars to build. The CS-1, by comparison, fits in a third of a server rack, consumes 20 kilowatts of power, and sells for a few million dollars.

While the task is niche (but useful) and the problem well-suited to the CS-1, it’s still a pretty stunning result. So how’d they pull it off? It’s all in the design.

Cut the Commute
Computer chips begin life on a big piece of silicon called a wafer. Multiple chips are etched onto the same wafer and then the wafer is cut into individual chips. While the WSE is also etched onto a silicon wafer, the wafer is left intact as a single, operating unit. This wafer-scale chip contains almost 400,000 processing cores. Each core is connected to its own dedicated memory and its four neighboring cores.

Putting that many cores on a single chip and giving them their own memory is why the WSE is bigger; it’s also why, in this case, it’s better.

Most large-scale computing tasks depend on massively parallel processing. Researchers distribute the task among hundreds or thousands of chips. The chips need to work in concert, so they’re in constant communication, shuttling information back and forth. A similar process takes place within each chip, as information moves between processor cores, which are doing the calculations, and shared memory to store the results.

It’s a little like an old-timey company that does all its business on paper.

The company uses couriers to send and collect documents from other branches and archives across town. The couriers know the best routes through the city, but the trips take some minimum amount of time determined by the distance between the branches and archives, the courier’s top speed, and how many other couriers are on the road. In short, distance and traffic slow things down.

Now, imagine the company builds a brand new gleaming skyscraper. Every branch is moved into the new building and every worker gets a small filing cabinet in their office to store documents. Now any document they need can be stored and retrieved in the time it takes to step across the office or down the hall to their neighbor’s office. The information commute has all but disappeared. Everything’s in the same house.

Cerebras’s megachip is a bit like that skyscraper. The way it shuttles information—aided further by its specially tailored compiling software—is far more efficient compared to a traditional supercomputer that needs to network a ton of traditional chips.

Simulating the World as It Unfolds
It’s worth noting the chip can only handle problems small enough to fit on the wafer. But such problems may have quite practical applications because of the machine’s ability to do high-fidelity simulation in real-time. The authors note, for example, the machine should in theory be able to accurately simulate the air flow around a helicopter trying to land on a flight deck and semi-automate the process—something not possible with traditional chips.

Another opportunity, they note, would be to use a simulation as input to train a neural network also residing on the chip. In an intriguing and related example, a Caltech machine learning technique recently proved to be 1,000 times faster at solving the same kind of partial differential equations at play here to simulate fluid dynamics.

They also note that improvements in the chip (and others like it, should they arrive) will push back the limits of what can be accomplished. Already, Cerebras has teased the release of its next-generation chip, which will have 2.6 trillion transistors, 850,00 cores, and more than double the memory.

Of course, it still remains to be seen whether wafer-scale computing really takes off. The idea has been around for decades, but Cerebras is the first to pursue it seriously. Clearly, they believe they’ve solved the problem in a way that’s useful and economical.

Other new architectures are also being pursued in the lab. Memristor-based neuromorphic chips, for example, mimic the brain by putting processing and memory into individual transistor-like components. And of course, quantum computers are in a separate lane, but tackle similar problems.

It could be that one of these technologies eventually rises to rule them all. Or, and this seems just as likely, computing may splinter into a bizarre quilt of radical chips, all stitched together to make the most of each depending on the situation.

Image credit: Cerebras Continue reading →

Posted in Human Robots

#437543 This Is How We’ll Engineer Artificial ...

Posted on November 17, 2020 by Android

Take a Jeopardy! guess: this body part was once referred to as the “consummation of all perfection as an instrument.”

Answer: “What is the human hand?”

Our hands are insanely complex feats of evolutionary engineering. Densely-packed sensors provide intricate and ultra-sensitive feelings of touch. Dozens of joints synergize to give us remarkable dexterity. A “sixth sense” awareness of where our hands are in space connects them to the mind, making it possible to open a door, pick up a mug, and pour coffee in total darkness based solely on what they feel.

So why can’t robots do the same?

In a new article in Science, Dr. Subramanian Sundaram at Boston and Harvard University argues that it’s high time to rethink robotic touch. Scientists have long dreamed of artificially engineering robotic hands with the same dexterity and feedback that we have. Now, after decades, we’re at the precipice of a breakthrough thanks to two major advances. One, we better understand how touch works in humans. Two, we have the mega computational powerhouse called machine learning to recapitulate biology in silicon.

Robotic hands with a sense of touch—and the AI brain to match it—could overhaul our idea of robots. Rather than charming, if somewhat clumsy, novelties, robots equipped with human-like hands are far more capable of routine tasks—making food, folding laundry—and specialized missions like surgery or rescue. But machines aren’t the only ones to gain. For humans, robotic prosthetic hands equipped with accurate, sensitive, and high-resolution artificial touch is the next giant breakthrough to seamlessly link a biological brain to a mechanical hand.

Here’s what Sundaram laid out to get us to that future.

How Does Touch Work, Anyway?
Let me start with some bad news: reverse engineering the human hand is really hard. It’s jam-packed with over 17,000 sensors tuned to mechanical forces alone, not to mention sensors for temperature and pain. These force “receptors” rely on physical distortions—bending, stretching, curling—to signal to the brain.

The good news? We now have a far clearer picture of how biological touch works. Imagine a coin pressed into your palm. The sensors embedded in the skin, called mechanoreceptors, capture that pressure, and “translate” it into electrical signals. These signals pulse through the nerves on your hand to the spine, and eventually make their way to the brain, where they gets interpreted as “touch.”

At least, that’s the simple version, but one too vague and not particularly useful for recapitulating touch. To get there, we need to zoom in.

The cells on your hand that collect touch signals, called tactile “first order” neurons (enter Star Wars joke) are like upside-down trees. Intricate branches extend from their bodies, buried deep in the skin, to a vast area of the hand. Each neuron has its own little domain called “receptor fields,” although some overlap. Like governors, these neurons manage a semi-dedicated region, so that any signal they transfer to the higher-ups—spinal cord and brain—is actually integrated from multiple sensors across a large distance.

It gets more intricate. The skin itself is a living entity that can regulate its own mechanical senses through hydration. Sweat, for example, softens the skin, which changes how it interacts with surrounding objects. Ever tried putting a glove onto a sweaty hand? It’s far more of a struggle than a dry one, and feels different.

In a way, the hand’s tactile neurons play a game of Morse Code. Through different frequencies of electrical beeps, they’re able to transfer information about an object’s size, texture, weight, and other properties, while also asking the brain for feedback to better control the object.

Biology to Machine
Reworking all of our hands’ greatest features into machines is absolutely daunting. But robots have a leg up—they’re not restricted to biological hardware. Earlier this year, for example, a team from Columbia engineered a “feeling” robotic finger using overlapping light emitters and sensors in a way loosely similar to receptor fields. Distortions in light were then analyzed with deep learning to translate into contact location and force.

Although a radical departure from our own electrical-based system, the Columbia team’s attempt was clearly based on human biology. They’re not alone. “Substantial progress is being made in the creation of soft, stretchable electronic skins,” said Sundaram, many of which can sense forces or pressure, although they’re currently still limited.

What’s promising, however, is the “exciting progress in using visual data,” said Sundaram. Computer vision has gained enormously from ubiquitous cameras and large datasets, making it possible to train powerful but data-hungry algorithms such as deep convolutional neural networks (CNNs).

By piggybacking on their success, we can essentially add “eyes” to robotic hands, a superpower us humans can’t imagine. Even better, CNNs and other classes of algorithms can be readily adopted for processing tactile data. Together, a robotic hand could use its eyes to scan an object, plan its movements for grasp, and use touch for feedback to adjust its grip. Maybe we’ll finally have a robot that easily rescues the phone sadly dropped into a composting toilet. Or something much grander to benefit humanity.

That said, relying too heavily on vision could also be a downfall. Take a robot that scans a wide area of rubble for signs of life during a disaster response. If touch relies on sight, then it would have to keep a continuous line-of-sight in a complex and dynamic setting—something computer vision doesn’t do well in, at least for now.

A Neuromorphic Way Forward
Too Debbie Downer? I got your back! It’s hard to overstate the challenges, but what’s clear is that emerging machine learning tools can tackle data processing challenges. For vision, it’s distilling complex images into “actionable control policies,” said Sundaram. For touch, it’s easy to imagine the same. Couple the two together, and that’s a robotic super-hand in the making.

Going forward, argues Sundaram, we need to closely adhere to how the hand and brain process touch. Hijacking our biological “touch machinery” has already proved useful. In 2019, one team used a nerve-machine interface for amputees to control a robotic arm—the DEKA LUKE arm—and sense what the limb and attached hand were feeling. Pressure on the LUKE arm and hand activated an implanted neural interface, which zapped remaining nerves in a way that the brain processes as touch. When the AI analyzed pressure data similar to biological tactile neurons, the person was able to better identify different objects with their eyes closed.

“Neuromorphic tactile hardware (and software) advances will strongly influence the future of bionic prostheses—a compelling application of robotic hands,” said Sundaram, adding that the next step is to increase the density of sensors.

Two additional themes made the list of progressing towards a cyborg future. One is longevity, in that sensors on a robot need to be able to reliably produce large quantities of high-quality data—something that’s seemingly mundane, but is a practical limitation.

The other is going all-in-one. Rather than just a pressure sensor, we need something that captures the myriad of touch sensations. From feather-light to a heavy punch, from vibrations to temperatures, a tree-like architecture similar to our hands would help organize, integrate, and otherwise process data collected from those sensors.

Just a decade ago, mind-controlled robotics were considered a blue sky, stretch-goal neurotechnological fantasy. We now have a chance to “close the loop,” from thought to movement to touch and back to thought, and make some badass robots along the way.

Image Credit: PublicDomainPictures from Pixabay Continue reading →

Posted in Human Robots