Tag Archives: failure

#437103 How to Make Sense of Uncertainty in a ...

As the internet churns with information about Covid-19, about the virus that causes the disease, and about what we’re supposed to do to fight it, it can be difficult to see the forest for the trees. What can we realistically expect for the rest of 2020? And how do we even know what’s realistic?

Today, humanity’s primary, ideal goal is to eliminate the virus, SARS-CoV-2, and Covid-19. Our second-choice goal is to control virus transmission. Either way, we have three big aims: to save lives, to return to public life, and to keep the economy functioning.

To hit our second-choice goal—and maybe even our primary goal—countries are pursuing five major public health strategies. Note that many of these advances cross-fertilize: for example, advances in virus testing and antibody testing will drive data-based prevention efforts.

Five major public health strategies are underway to bring Covid-19 under control and to contain the spread of SARS-CoV-2.
These strategies arise from things we can control based on the things that we know at any given moment. But what about the things we can’t control and don’t yet know?

The biology of the virus and how it interacts with our bodies is what it is, so we should seek to understand it as thoroughly as possible. How long any immunity gained from prior infection lasts—and indeed whether people develop meaningful immunity at all after infection—are open questions urgently in need of greater clarity. Similarly, right now it’s important to focus on understanding rather than making assumptions about environmental factors like seasonality.

But the biggest question on everyone’s lips is, “When?” When will we see therapeutic progress against Covid-19? And when will life get “back to normal”? There are lots of models out there on the internet; which of those models are right? The simple answer is “none of them.” That’s right—it’s almost certain that every model you’ve seen is wrong in at least one detail, if not all of them. But modeling is meant to be a tool for deeper thinking, a way to run mental (and computational) experiments before—and while—taking action. As George E. P. Box famously wrote in 1976, “All models are wrong, but some are useful.”

Here, we’re seeking useful insights, as opposed to exact predictions, which is why we’re pulling back from quantitative details to get at the mindsets that will support agency and hope. To that end, I’ve been putting together timelines that I believe will yield useful expectations for the next year or two—and asking how optimistic I need to be in order to believe a particular timeline.

For a moderately optimistic scenario to be relevant, breakthroughs in science and technology come at paces expected based on previous efforts and assumptions that turn out to be basically correct; accessibility of those breakthroughs increases at a reasonable pace; regulation achieves its desired effects, without major surprises; and compliance with regulations is reasonably high.

In contrast, if I’m being highly optimistic, breakthroughs in science and technology and their accessibility come more quickly than they ever have before; regulation is evidence-based and successful in the first try or two; and compliance with those regulations is high and uniform. If I’m feeling not-so-optimistic, then I anticipate serious setbacks to breakthroughs and accessibility (with the overturning of many important assumptions), repeated failure of regulations to achieve their desired outcomes, and low compliance with those regulations.

The following scenarios outline the things that need to happen in the fight against Covid-19, when I expect to see them, and how confident I feel in those expectations. They focus on North America and Europe because there are data missing about China’s 2019 outbreak and other regions are still early in their outbreaks. Perhaps the most important thing to keep in mind throughout: We know more today than we did yesterday, but we still have much to learn. New knowledge derived from greater study and debate will almost certainly inspire ongoing course corrections.

As you dive into the scenarios below, practice these three mindset shifts. First, defeating Covid-19 will be a marathon, not a sprint. We shouldn’t expect life to look like 2019 for the next year or two—if ever. As Ed Yong wrote recently in The Atlantic, “There won’t be an obvious moment when everything is under control and regular life can safely resume.” Second, remember that you have important things to do for at least a year. And third, we are all in this together. There is no “us” and “them.” We must all be alert, responsive, generous, and strong throughout 2020 and 2021—and willing to throw away our assumptions when scientific evidence invalidates them.

The Middle Way: Moderate Optimism
Let’s start with the case in which I have the most confidence: moderate optimism.

This timeline considers milestones through late 2021, the earliest that I believe vaccines will become available. The “normal” timeline for developing a vaccine for diseases like seasonal flu is 18 months, which leads to my projection that we could potentially have vaccines as soon as 18 months from the first quarter of 2020. While Melinda Gates agrees with that projection, others (including AI) believe that 3 to 5 years is far more realistic, based on past vaccine development and the need to test safety and efficacy in humans. However, repurposing existing vaccines against other diseases—or piggybacking off clever synthetic platforms—could lead to vaccines being available sooner. I tried to balance these considerations for this moderately optimistic scenario. Either way, deploying vaccines at the end of 2021 is probably much later than you may have been led to believe by the hype engine. Again, if you take away only one message from this article, remember that the fight against Covid-19 is a marathon, not a sprint.

Here, I’ve visualized a moderately optimistic scenario as a baseline. Think of these timelines as living guides, as opposed to exact predictions. There are still many unknowns. More or less optimistic views (see below) and new information could shift these timelines forward or back and change the details of the strategies.
Based on current data, I expect that the first wave of Covid-19 cases (where we are now) will continue to subside in many areas, leading governments to ease restrictions in an effort to get people back to work. We’re already seeing movement in that direction, with a variety of benchmarks and changes at state and country levels around the world. But depending on the details of the changes, easing restrictions will probably cause a second wave of sickness (see Germany and Singapore), which should lead governments to reimpose at least some restrictions.

In tandem, therapeutic efforts will be transitioning from emergency treatments to treatments that have been approved based on safety and efficacy data in clinical trials. In a moderately optimistic scenario, assuming clinical trials currently underway yield at least a few positive results, this shift to mostly approved therapies could happen as early as the third or fourth quarter of this year and continue from there. One approval that should come rather quickly is for plasma therapies, in which the blood from people who have recovered from Covid-19 is used as a source of antibodies for people who are currently sick.

Companies around the world are working on both viral and antibody testing, focusing on speed, accuracy, reliability, and wide accessibility. While these tests are currently being run in hospitals and research laboratories, at-home testing is a critical component of the mass testing we’ll need to keep viral spread in check. These are needed to minimize the impact of asymptomatic cases, test the assumption that infection yields resistance to subsequent infection (and whether it lasts), and construct potential immunity passports if this assumption holds. Testing is also needed for contact tracing efforts to prevent further spread and get people back to public life. Finally, it’s crucial to our fundamental understanding of the biology of SARS-CoV-2 and Covid-19.

We need tests that are very reliable, both in the clinic and at home. So, don’t go buying any at-home test kits just yet, even if you find them online. Wait for reliable test kits and deeper understanding of how a test result translates to everyday realities. If we’re moderately optimistic, in-clinic testing will rapidly expand this quarter and/or next, with the possibility of broadly available, high-quality at-home sampling (and perhaps even analysis) thereafter.

Note that testing is not likely to be a “one-and-done” endeavor, as a person’s infection and immunity status change over time. Expect to be testing yourself—and your family—often as we move later into 2020.

Testing data are also going to inform distancing requirements at the country and local levels. In this scenario, restrictions—at some level of stringency—could persist at least through the end of 2020, as most countries are way behind the curve on testing (Iceland is an informative exception). Governments will likely continue to ask citizens to work from home if at all possible; to wear masks or face coverings in public; to employ heightened hygiene and social distancing in workplaces; and to restrict travel and social gatherings. So while it’s likely we’ll be eating in local restaurants again in 2020 in this scenario, at least for a little while, it’s not likely we’ll be heading to big concerts any time soon.

The Extremes: High and Low Optimism
How would high and low levels of optimism change our moderately optimistic timeline? The milestones are the same, but the time required to achieve them is shorter or longer, respectively. Quantifying these shifts is less important than acknowledging and incorporating a range of possibilities into our view. It pays to pay attention to our bias. Here are a few examples of reasonable possibilities that could shift the moderately optimistic timeline.

When vaccines become available
Vaccine repurposing could shorten the time for vaccines to become available; today, many vaccine candidates are in various stages of testing. On the other hand, difficulties in manufacture and distribution, or faster-than-expected mutation of SARS-CoV-2, could slow vaccine development. Given what we know now, I am not strongly concerned about either of these possibilities—drug companies are rapidly expanding their capabilities, and viral mutation isn’t an urgent concern at this time based on sequencing data—but they could happen.

At first, governments will likely supply vaccines to essential workers such as healthcare workers, but it is essential that vaccines become widely available around the world as quickly and as safely as possible. Overall, I suggest a dose of skepticism when reading highly optimistic claims about a vaccine (or multiple vaccines) being available in 2020. Remember, a vaccine is a knockout punch, not a first line of defense for an outbreak.

When testing hits its stride
While I am confident that testing is a critical component of our response to Covid-19, reliability is incredibly important to testing for SARS-CoV-2 and for immunity to the disease, particularly at home. For an individual, a false negative (being told you don’t have antibodies when you really do) could be just as bad as a false positive (being told you do have antibodies when you really don’t). Those errors are compounded when governments are trying to make evidence-based policies for social and physical distancing.

If you’re highly optimistic, high-quality testing will ramp up quickly as companies and scientists innovate rapidly by cleverly combining multiple test modalities, digital signals, and cutting-edge tech like CRISPR. Pop-up testing labs could also take some pressure off hospitals and clinics.

If things don’t go well, reliability issues could hinder testing, manufacturing bottlenecks could limit availability, and both could hamstring efforts to control spread and ease restrictions. And if it turns out that immunity to Covid-19 isn’t working the way we assumed, then we must revisit our assumptions about our path(s) back to public life, as well as our vaccine-development strategies.

How quickly safe and effective treatments appear
Drug development is known to be long, costly, and fraught with failure. It’s not uncommon to see hope in a drug spike early only to be dashed later on down the road. With that in mind, the number of treatments currently under investigation is astonishing, as is the speed through which they’re proceeding through testing. Breakthroughs in a therapeutic area—for example in treating the seriously ill or in reducing viral spread after an infection takes hold—could motivate changes in the focus of distancing regulations.

While speed will save lives, we cannot overlook the importance of knowing a treatment’s efficacy (does it work against Covid-19?) and safety (does it make you sick in a different, or worse, way?). Repurposing drugs that have already been tested for other diseases is speeding innovation here, as is artificial intelligence.

Remarkable collaborations among governments and companies, large and small, are driving innovation in therapeutics and devices such as ventilators for treating the sick.

Whether government policies are effective and responsive
Those of us who have experienced lockdown are eager for it to be over. Businesses, economists, and governments are also eager to relieve the terrible pressure that is being exerted on the global economy. However, lifting restrictions will almost certainly lead to a resurgence in sickness.

Here, the future is hard to model because there are many, many factors at play, and at play differently in different places—including the extent to which individuals actually comply with regulations.

Reliable testing—both in the clinic and at home—is crucial to designing and implementing restrictions, monitoring their effectiveness, and updating them; delays in reliable testing could seriously hamper this design cycle. Lack of trust in governments and/or companies could also suppress uptake. That said, systems are already in place for contact tracing in East Asia. Other governments could learn important lessons, but must also earn—and keep—their citizens’ trust.

Expect to see restrictions descend and then lift in response to changes in the number of Covid-19 cases and in the effectiveness of our prevention strategies. Also expect country-specific and perhaps even area-specific responses that differ from each other. The benefit of this approach? Governments around the world are running perhaps hundreds of real-time experiments and design cycles in balancing health and the economy, and we can learn from the results.

A Way Out
As Jeremy Farrar, head of the Wellcome Trust, told Science magazine, “Science is the exit strategy.” Some of our greatest technological assistance is coming from artificial intelligence, digital tools for collaboration, and advances in biotechnology.

Our exit strategy also needs to include empathy and future visioning—because in the midst of this crisis, we are breaking ground for a new, post-Covid future.

What do we want that future to look like? How will the hard choices we make now about data ethics impact the future of surveillance? Will we continue to embrace inclusiveness and mass collaboration? Perhaps most importantly, will we lay the foundation for successfully confronting future challenges? Whether we’re thinking about the next pandemic (and there will be others) or the cascade of catastrophes that climate change is bringing ever closer—it’s important to remember that we all have the power to become agents of that change.

Special thanks to Ola Kowalewski and Jason Dorrier for significant conversations.

Image Credit: Drew Beamer / Unsplash Continue reading

Posted in Human Robots

#436491 The Year’s Most Fascinating Tech ...

Last Saturday we took a look at some of the most-read Singularity Hub articles from 2019. This week, we’re featuring some of our favorite articles from the last year. As opposed to short pieces about what’s happening, these are long reads about why it matters and what’s coming next. Some of them make the news while others frame the news, go deep on big ideas, go behind the scenes, or explore the human side of technological progress.

We hope you find them as fascinating, inspiring, and illuminating as we did.

DeepMind and Google: The Battle to Control Artificial Intelligence
Hal Hodson | 1843
“[DeepMind cofounder and CEO Demis] Hassabis thought DeepMind would be a hybrid: it would have the drive of a startup, the brains of the greatest universities, and the deep pockets of one of the world’s most valuable companies. Every element was in place to hasten the arrival of [artificial general intelligence] and solve the causes of human misery.”

The Most Powerful Person in Silicon Valley
Katrina Brooker | Fast Company
“Billionaire Masayoshi Son—not Elon Musk, Jeff Bezos, or Mark Zuckerberg—has the most audacious vision for an AI-powered utopia where machines control how we live. And he’s spending hundreds of billions of dollars to realize it. Are you ready to live in Masa World?”

AR Will Spark the Next Big Tech Platform—Call It Mirrorworld
Kevin Kelly | Wired
“Eventually this melded world will be the size of our planet. It will be humanity’s greatest achievement, creating new levels of wealth, new social problems, and uncountable opportunities for billions of people. There are no experts yet to make this world; you are not late.”

Behind the Scenes of a Radical New Cancer Cure
Ilana Yurkiewicz | Undark
“I remember the first time I watched a patient get his Day 0 infusion. It felt anti-climactic. The entire process took about 15 minutes. The CAR-T cells are invisible to the naked eye, housed in a small plastic bag containing clear liquid. ‘That’s it?’ my patient asked when the nurse said it was over. The infusion part is easy. The hard part is everything that comes next.”

The Promise and Price of Cellular Therapies
Siddhartha Mukherjee | The New Yorker
“We like to imagine medical revolutions as, well, revolutionary—propelled forward through leaps of genius and technological innovation. But they are also evolutionary, nudged forward through the optimization of design and manufacture.”

Impossible Foods’ Rising Empire of Almost Meat
Chris Ip | Engadget
“Impossible says it wants to ultimately create a parallel universe of ersatz animal products from steak to eggs. …Yet as Impossible ventures deeper into the culinary uncanny valley, it also needs society to discard a fundamental cultural idea that dates back millennia and accept a new truth: Meat doesn’t have to come from animals.”

Inside the Amazon Warehouse Where Humans and Machines Become One
Matt Simon | Wired
“Seen from above, the scale of the system is dizzying. My robot, a little orange slab known as a ‘drive’ (or more formally and mythically, Pegasus), is just one of hundreds of its kind swarming a 125,000-square-foot ‘field’ pockmarked with chutes. It’s a symphony of electric whirring, with robots pausing for one another at intersections and delivering their packages to the slides.”

Boston Dynamics’ Robots Are Preparing to Leave the Lab—Is the World Ready?
James Vincent | The Verge
“After decades of kicking machines in parking lots, the company is set to launch its first ever commercial bot later this year: the quadrupedal Spot. It’s a crucial test for a company that’s spent decades pursuing long-sighted R&D. And more importantly, the success—or failure—of Spot will tell us a lot about our own robot future. Are we ready for machines to walk among us?”

I Cut the ‘Big Five’ Tech Giants From My Life. It Was Hell
Kashmir Hill | Gizmodo
“Critics of the big tech companies are often told, ‘If you don’t like the company, don’t use its products.’ I did this experiment to find out if that is possible, and I found out that it’s not—with the exception of Apple. …These companies are unavoidable because they control internet infrastructure, online commerce, and information flows.”

Why I (Still) Love Tech: In Defense of a Difficult Industry
Paul Ford | Wired
“The mysteries of software caught my eye when I was a boy, and I still see it with the same wonder, even though I’m now an adult. Proudshamed, yes, but I still love it, the mess of it, the code and toolkits, down to the pixels and the processors, and up to the buses and bridges. I love the whole made world. But I can’t deny that the miracle is over, and that there is an unbelievable amount of work left for us to do.”

The Peculiar Blindness of Experts
David Epstein | The Atlantic
“In business, esteemed (and lavishly compensated) forecasters routinely are wildly wrong in their predictions of everything from the next stock-market correction to the next housing boom. Reliable insight into the future is possible, however. It just requires a style of thinking that’s uncommon among experts who are certain that their deep knowledge has granted them a special grasp of what is to come.”

The Most Controversial Tree in the World
Rowan Jacobson | Pacific Standard
“…we are all GMOs, the beneficiaries of freakishly unlikely genetic mash-ups, and the real Island of Dr. Moreau is that blue-green botanical garden positioned third from the sun. Rather than changing the nature of nature, as I once thought, this might just be the very nature of nature.”

How an Augmented Reality Game Escalated Into Real-World Spy Warfare
Elizabeth Ballou | Vice
“In Ingress, players accept that every park and train station could be the site of an epic showdown, but that’s only the first step. The magic happens when other people accept that, too. When players feel like that magic is real, there are few limits to what they’ll do or where they’ll go for the sake of the game. ”

The Shady Cryptocurrency Boom on the Post-Soviet Frontier
Hannah Lucinda Smith | Wired
“…although the tourists won’t guess it as they stand at Kuchurgan’s gates, admiring how the evening light reflects off the silver plaque of Lenin, this plant is pumping out juice to a modern-day gold rush: a cryptocurrency boom that is underway all across the former Soviet Union, from the battlefields of eastern Ukraine to time-warp enclaves like Transnistria and freshly annexed Crimea.”

Scientists Are Totally Rethinking Animal Cognition
Ross Andersen | The Atlantic
“This idea that animals are conscious was long unpopular in the West, but it has lately found favor among scientists who study animal cognition. …For many scientists, the resonant mystery is no longer which animals are conscious, but which are not.”

I Wrote This on a 30-Year-Old Computer
Ian Bogost | The Atlantic
“[Back then] computing was an accompaniment to life, rather than the sieve through which all ideas and activities must filter. That makes using this 30-year-old device a surprising joy, one worth longing for on behalf of what it was at the time, rather than for the future it inaugurated.”

Image Credit: Wes Hicks / Unsplash Continue reading

Posted in Human Robots

#436188 The Blogger Behind “AI ...

Sure, artificial intelligence is transforming the world’s societies and economies—but can an AI come up with plausible ideas for a Halloween costume?

Janelle Shane has been asking such probing questions since she started her AI Weirdness blog in 2016. She specializes in training neural networks (which underpin most of today’s machine learning techniques) on quirky data sets such as compilations of knitting instructions, ice cream flavors, and names of paint colors. Then she asks the neural net to generate its own contributions to these categories—and hilarity ensues. AI is not likely to disrupt the paint industry with names like “Ronching Blue,” “Dorkwood,” and “Turdly.”

Shane’s antics have a serious purpose. She aims to illustrate the serious limitations of today’s AI, and to counteract the prevailing narrative that describes AI as well on its way to superintelligence and complete human domination. “The danger of AI is not that it’s too smart,” Shane writes in her new book, “but that it’s not smart enough.”

The book, which came out on Tuesday, is called You Look Like a Thing and I Love You. It takes its odd title from a list of AI-generated pick-up lines, all of which would at least get a person’s attention if shouted, preferably by a robot, in a crowded bar. Shane’s book is shot through with her trademark absurdist humor, but it also contains real explanations of machine learning concepts and techniques. It’s a painless way to take AI 101.

She spoke with IEEE Spectrum about the perils of placing too much trust in AI systems, the strange AI phenomenon of “giraffing,” and her next potential Halloween costume.

Janelle Shane on . . .

The un-delicious origin of her blog
“The narrower the problem, the smarter the AI will seem”
Why overestimating AI is dangerous
Giraffing!
Machine and human creativity

The un-delicious origin of her blog IEEE Spectrum: You studied electrical engineering as an undergrad, then got a master’s degree in physics. How did that lead to you becoming the comedian of AI?
Janelle Shane: I’ve been interested in machine learning since freshman year of college. During orientation at Michigan State, a professor who worked on evolutionary algorithms gave a talk about his work. It was full of the most interesting anecdotes–some of which I’ve used in my book. He told an anecdote about people setting up a machine learning algorithm to do lens design, and the algorithm did end up designing an optical system that works… except one of the lenses was 50 feet thick, because they didn’t specify that it couldn’t do that.
I started working in his lab on optics, doing ultra-short laser pulse work. I ended up doing a lot more optics than machine learning, but I always found it interesting. One day I came across a list of recipes that someone had generated using a neural net, and I thought it was hilarious and remembered why I thought machine learning was so cool. That was in 2016, ages ago in machine learning land.
Spectrum: So you decided to “establish weirdness as your goal” for your blog. What was the first weird experiment that you blogged about?
Shane: It was generating cookbook recipes. The neural net came up with ingredients like: “Take ¼ pounds of bones or fresh bread.” That recipe started out: “Brown the salmon in oil, add creamed meat to the mixture.” It was making mistakes that showed the thing had no memory at all.
Spectrum: You say in the book that you can learn a lot about AI by giving it a task and watching it flail. What do you learn?
Shane: One thing you learn is how much it relies on surface appearances rather than deep understanding. With the recipes, for example: It got the structure of title, category, ingredients, instructions, yield at the end. But when you look more closely, it has instructions like “Fold the water and roll it into cubes.” So clearly this thing does not understand water, let alone the other things. It’s recognizing certain phrases that tend to occur, but it doesn’t have a concept that these recipes are describing something real. You start to realize how very narrow the algorithms in this world are. They only know exactly what we tell them in our data set.
BACK TO TOP↑ “The narrower the problem, the smarter the AI will seem” Spectrum: That makes me think of DeepMind’s AlphaGo, which was universally hailed as a triumph for AI. It can play the game of Go better than any human, but it doesn’t know what Go is. It doesn’t know that it’s playing a game.
Shane: It doesn’t know what a human is, or if it’s playing against a human or another program. That’s also a nice illustration of how well these algorithms do when they have a really narrow and well-defined problem.
The narrower the problem, the smarter the AI will seem. If it’s not just doing something repeatedly but instead has to understand something, coherence goes down. For example, take an algorithm that can generate images of objects. If the algorithm is restricted to birds, it could do a recognizable bird. If this same algorithm is asked to generate images of any animal, if its task is that broad, the bird it generates becomes an unrecognizable brown feathered smear against a green background.
Spectrum: That sounds… disturbing.
Shane: It’s disturbing in a weird amusing way. What’s really disturbing is the humans it generates. It hasn’t seen them enough times to have a good representation, so you end up with an amorphous, usually pale-faced thing with way too many orifices. If you asked it to generate an image of a person eating pizza, you’ll have blocks of pizza texture floating around. But if you give that image to an image-recognition algorithm that was trained on that same data set, it will say, “Oh yes, that’s a person eating pizza.”
BACK TO TOP↑ Why overestimating AI is dangerous Spectrum: Do you see it as your role to puncture the AI hype?
Shane: I do see it that way. Not a lot of people are bringing out this side of AI. When I first started posting my results, I’d get people saying, “I don’t understand, this is AI, shouldn’t it be better than this? Why doesn't it understand?” Many of the impressive examples of AI have a really narrow task, or they’ve been set up to hide how little understanding it has. There’s a motivation, especially among people selling products based on AI, to represent the AI as more competent and understanding than it actually is.
Spectrum: If people overestimate the abilities of AI, what risk does that pose?
Shane: I worry when I see people trusting AI with decisions it can’t handle, like hiring decisions or decisions about moderating content. These are really tough tasks for AI to do well on. There are going to be a lot of glitches. I see people saying, “The computer decided this so it must be unbiased, it must be objective.”

“If the algorithm’s task is to replicate human hiring decisions, it’s going to glom onto gender bias and race bias.”
—Janelle Shane, AI Weirdness blogger
That’s another thing I find myself highlighting in the work I’m doing. If the data includes bias, the algorithm will copy that bias. You can’t tell it not to be biased, because it doesn’t understand what bias is. I think that message is an important one for people to understand.
If there’s bias to be found, the algorithm is going to go after it. It’s like, “Thank goodness, finally a signal that’s reliable.” But for a tough problem like: Look at these resumes and decide who’s best for the job. If its task is to replicate human hiring decisions, it’s going to glom onto gender bias and race bias. There’s an example in the book of a hiring algorithm that Amazon was developing that discriminated against women, because the historical data it was trained on had that gender bias.
Spectrum: What are the other downsides of using AI systems that don’t really understand their tasks?
Shane: There is a risk in putting too much trust in AI and not examining its decisions. Another issue is that it can solve the wrong problems, without anyone realizing it. There have been a couple of cases in medicine. For example, there was an algorithm that was trained to recognize things like skin cancer. But instead of recognizing the actual skin condition, it latched onto signals like the markings a surgeon makes on the skin, or a ruler placed there for scale. It was treating those things as a sign of skin cancer. It’s another indication that these algorithms don’t understand what they’re looking at and what the goal really is.
BACK TO TOP↑ Giraffing Spectrum: In your blog, you often have neural nets generate names for things—such as ice cream flavors, paint colors, cats, mushrooms, and types of apples. How do you decide on topics?
Shane: Quite often it’s because someone has written in with an idea or a data set. They’ll say something like, “I’m the MIT librarian and I have a whole list of MIT thesis titles.” That one was delightful. Or they’ll say, “We are a high school robotics team, and we know where there’s a list of robotics team names.” It’s fun to peek into a different world. I have to be careful that I’m not making fun of the naming conventions in the field. But there’s a lot of humor simply in the neural net’s complete failure to understand. Puns in particular—it really struggles with puns.
Spectrum: Your blog is quite absurd, but it strikes me that machine learning is often absurd in itself. Can you explain the concept of giraffing?
Shane: This concept was originally introduced by [internet security expert] Melissa Elliott. She proposed this phrase as a way to describe the algorithms’ tendency to see giraffes way more often than would be likely in the real world. She posted a whole bunch of examples, like a photo of an empty field in which an image-recognition algorithm has confidently reported that there are giraffes. Why does it think giraffes are present so often when they’re actually really rare? Because they’re trained on data sets from online. People tend to say, “Hey look, a giraffe!” And then take a photo and share it. They don’t do that so often when they see an empty field with rocks.
There’s also a chatbot that has a delightful quirk. If you show it some photo and ask it how many giraffes are in the picture, it will always answer with some non zero number. This quirk comes from the way the training data was generated: These were questions asked and answered by humans online. People tended not to ask the question “How many giraffes are there?” when the answer was zero. So you can show it a picture of someone holding a Wii remote. If you ask it how many giraffes are in the picture, it will say two.
BACK TO TOP↑ Machine and human creativity Spectrum: AI can be absurd, and maybe also creative. But you make the point that AI art projects are really human-AI collaborations: Collecting the data set, training the algorithm, and curating the output are all artistic acts on the part of the human. Do you see your work as a human-AI art project?
Shane: Yes, I think there is artistic intent in my work; you could call it literary or visual. It’s not so interesting to just take a pre-trained algorithm that’s been trained on utilitarian data, and tell it to generate a bunch of stuff. Even if the algorithm isn’t one that I’ve trained myself, I think about, what is it doing that’s interesting, what kind of story can I tell around it, and what do I want to show people.

The Halloween costume algorithm “was able to draw on its knowledge of which words are related to suggest things like sexy barnacle.”
—Janelle Shane, AI Weirdness blogger
Spectrum: For the past three years you’ve been getting neural nets to generate ideas for Halloween costumes. As language models have gotten dramatically better over the past three years, are the costume suggestions getting less absurd?
Shane: Yes. Before I would get a lot more nonsense words. This time I got phrases that were related to real things in the data set. I don’t believe the training data had the words Flying Dutchman or barnacle. But it was able to draw on its knowledge of which words are related to suggest things like sexy barnacle and sexy Flying Dutchman.
Spectrum: This year, I saw on Twitter that someone made the gothy giraffe costume happen. Would you ever dress up for Halloween in a costume that the neural net suggested?
Shane: I think that would be fun. But there would be some challenges. I would love to go as the sexy Flying Dutchman. But my ambition may constrict me to do something more like a list of leg parts.
BACK TO TOP↑ Continue reading

Posted in Human Robots

#436180 Bipedal Robot Cassie Cal Learns to ...

There’s no particular reason why knowing how to juggle would be a useful skill for a robot. Despite this, robots are frequently taught how to juggle things. Blind robots can juggle, humanoid robots can juggle, and even drones can juggle. Why? Because juggling is hard, man! You have to think about a bunch of different things at once, and also do a bunch of different things at once, which this particular human at least finds to be overly stressful. While juggling may not stress robots out, it does require carefully coordinated sensing and computing and actuation, which means that it’s as good a task as any (and a more entertaining task than most) for testing the capabilities of your system.

UC Berkeley’s Cassie Cal robot, which consists of two legs and what could be called a torso if you were feeling charitable, has just learned to juggle by bouncing a ball on what would be her head if she had one of those. The idea is that if Cassie can juggle while balancing at the same time, she’ll be better able to do other things that require dynamic multitasking, too. And if that doesn’t work out, she’ll still be able to join the circus.

Cassie’s juggling is assisted by an external motion capture system that tracks the location of the ball, but otherwise everything is autonomous. Cassie is able to juggle the ball by leaning forwards and backwards, left and right, and moving up and down. She does this while maintaining her own balance, which is the whole point of this research—successfully executing two dynamic behaviors that may sometimes be at odds with one another. The end goal here is not to make a better juggling robot, but rather to explore dynamic multitasking, a skill that robots will need in order to be successful in human environments.

This work is from the Hybrid Robotics Lab at UC Berkeley, led by Koushil Sreenath, and is being done by Katherine Poggensee, Albert Li, Daniel Sotsaikich, Bike Zhang, and Prasanth Kotaru.

For a bit more detail, we spoke with Albert Li via email.

Image: UC Berkeley

UC Berkeley’s Cassie Cal getting ready to juggle.

IEEE Spectrum: What would be involved in getting Cassie to juggle without relying on motion capture?

Albert Li: Our motivation for starting off with motion capture was to first address the control challenge of juggling on a biped without worrying about implementing the perception. We actually do have a ball detector working on a camera, which would mean we wouldn’t have to rely on the motion capture system. However, we need to mount the camera in a way that it would provide the best upwards field of view, and we also have develop a reliable estimator. The estimator is particularly important because when the ball gets close enough to the camera, we actually can’t track the ball and have to assume our dynamic models describe its motion accurately enough until it bounces back up.

What keeps Cassie from juggling indefinitely?

There are a few factors that affect how long Cassie can sustain a juggle. While in simulation the paddle exhibits homogeneous properties like its stiffness and damping, in reality every surface has anisotropic contact properties. So, there are parts of the paddle which may be better for juggling than others (and importantly, react differently than modeled). These differences in contact are also exacerbated due to how the paddle is cantilevered when mounted on Cassie. When the ball hits these areas, it leads to a larger than expected error in a juggle. Due to the small size of the paddle, the ball may then just hit the paddle’s edge and end the juggling run. Over a very long run, this is a likely occurrence. Additionally, some large juggling errors could cause Cassie’s feet to slip slightly, which ends up changing the stable standing position over time. Since this version of the controller assumes Cassie is stationary, this change in position eventually leads to poor juggles and failure.

Would Cassie be able to juggle while walking (or hovershoe-ing)?

Walking (and hovershoe-ing) while juggling is a far more challenging problem and is certainly a goal for future research. Some of these challenges include getting the paddle to precise poses to juggle the ball while also moving to avoid any destabilizing effects of stepping incorrectly. The number of juggles per step of walking could also vary and make the mathematics of the problem more challenging. The controller goal is also more involved. While the current goal of the juggling controller is to juggle the ball to a static apex position, with a walking juggling controller, we may instead want to hit the ball forwards and also walk forwards to bounce it, juggle the ball along a particular path, etc. Solving such challenges would be the main thrusts of the follow-up research.

Can you give an example of a practical task that would be made possible by using a controller like this?

Studying juggling means studying contact behavior and leveraging our models of it to achieve a known objective. Juggling could also be used to study predictable post-contact flight behavior. Consider the scenario where a robot is attempting to make a catch, but fails, letting the ball to bounce off of its hand, and then recovering the catch. This behavior could also be intentional: It is often easier to first execute a bounce to direct the target and then perform a subsequent action. For example, volleyball players could in principle directly hit a spiked ball back, but almost always bump the ball back up and then return it.

Even beyond this motivating example, the kinds of models we employ to get juggling working are more generally applicable to any task that involves contact, which could include tasks besides bouncing like sliding and rolling. For example, clearing space on a desk by pushing objects to the side may be preferable than individually manipulating each and every object on it.

You mention collaborative juggling or juggling multiple balls—is that something you’ve tried yet? Can you talk a bit more about what you’re working on next?

We haven’t yet started working on collaborative or multi-ball juggling, but that’s also a goal for future work. Juggling multiple balls statically is probably the most reasonable next goal, but presents additional challenges. For instance, you have to encode a notion of juggling urgency (if the second ball isn’t hit hard enough, you have less time to get the first ball up before you get back to the second one).

On the other hand, collaborative human-robot juggling requires a more advanced decision-making framework. To get robust multi-agent juggling, the robot will need to employ some sort of probabilistic model of the expected human behavior (are they likely to move somewhere? Are they trying to catch the ball high or low? Is it safe to hit the ball back?). In general, developing such human models is difficult since humans are fairly unpredictable and often don’t exhibit rational behavior. This will be a focus of future work.

[ Hybrid Robotics Lab ] Continue reading

Posted in Human Robots

#436123 A Path Towards Reasonable Autonomous ...

Editor’s Note: The debate on autonomous weapons systems has been escalating over the past several years as the underlying technologies evolve to the point where their deployment in a military context seems inevitable. IEEE Spectrum has published a variety of perspectives on this issue. In summary, while there is a compelling argument to be made that autonomous weapons are inherently unethical and should be banned, there is also a compelling argument to be made that autonomous weapons could potentially make conflicts less harmful, especially to non-combatants. Despite an increasing amount of international attention (including from the United Nations), progress towards consensus, much less regulatory action, has been slow. The following workshop paper on autonomous weapons systems policy is remarkable because it was authored by a group of experts with very different (and in some cases divergent) views on the issue. Even so, they were able to reach consensus on a roadmap that all agreed was worth considering. It’s collaborations like this that could be the best way to establish a reasonable path forward on such a contentious issue, and with the permission of the authors, we’re excited to be able to share this paper (originally posted on Georgia Tech’s Mobile Robot Lab website) with you in its entirety.

Autonomous Weapon Systems: A Roadmapping Exercise
Over the past several years, there has been growing awareness and discussion surrounding the possibility of future lethal autonomous weapon systems that could fundamentally alter humanity’s relationship with violence in war. Lethal autonomous weapons present a host of legal, ethical, moral, and strategic challenges. At the same time, artificial intelligence (AI) technology could be used in ways that improve compliance with the laws of war and reduce non-combatant harm. Since 2014, states have come together annually at the United Nations to discuss lethal autonomous weapons systems1. Additionally, a growing number of individuals and non-governmental organizations have become active in discussions surrounding autonomous weapons, contributing to a rapidly expanding intellectual field working to better understand these issues. While a wide range of regulatory options have been proposed for dealing with the challenge of lethal autonomous weapons, ranging from a preemptive, legally binding international treaty to reinforcing compliance with existing laws of war, there is as yet no international consensus on a way forward.

The lack of an international policy consensus, whether codified in a formal document or otherwise, poses real risks. States could fall victim to a security dilemma in which they deploy untested or unsafe weapons that pose risks to civilians or international stability. Widespread proliferation could enable illicit uses by terrorists, criminals, or rogue states. Alternatively, a lack of guidance on which uses of autonomy are acceptable could stifle valuable research that could reduce the risk of non-combatant harm.

International debate thus far has predominantly centered around whether or not states should adopt a preemptive, legally-binding treaty that would ban lethal autonomous weapons before they can be built. Some of the authors of this document have called for such a treaty and would heartily support it, if states were to adopt it. Other authors of this document have argued an overly expansive treaty would foreclose the possibility of using AI to mitigate civilian harm. Options for international action are not binary, however, and there are a range of policy options that states should consider between adopting a comprehensive treaty or doing nothing.

The purpose of this paper is to explore the possibility of a middle road. If a roadmap could garner sufficient stakeholder support to have significant beneficial impact, then what elements could it contain? The exercise whose results are presented below was not to identify recommendations that the authors each prefer individually (the authors hold a broad spectrum of views), but instead to identify those components of a roadmap that the authors are all willing to entertain2. We, the authors, invite policymakers to consider these components as they weigh possible actions to address concerns surrounding autonomous weapons3.

Summary of Issues Surrounding Autonomous Weapons

There are a variety of issues that autonomous weapons raise, which might lend themselves to different approaches. A non-exhaustive list of issues includes:

The potential for beneficial uses of AI and autonomy that could improve precision and reliability in the use of force and reduce non-combatant harm.
Uncertainty about the path of future technology and the likelihood of autonomous weapons being used in compliance with the laws of war, or international humanitarian law (IHL), in different settings and on various timelines.
A desire for some degree of human involvement in the use of force. This has been expressed repeatedly in UN discussions on lethal autonomous weapon systems in different ways.
Particular risks surrounding lethal autonomous weapons specifically targeting personnel as opposed to vehicles or materiel.
Risks regarding international stability.
Risk of proliferation to terrorists, criminals, or rogue states.
Risk that autonomous systems that have been verified to be acceptable can be made unacceptable through software changes.
The potential for autonomous weapons to be used as scalable weapons enabling a small number of individuals to inflict very large-scale casualties at low cost, either intentionally or accidentally.

Summary of Components

A time-limited moratorium on the development, deployment, transfer, and use of anti-personnel lethal autonomous weapon systems4. Such a moratorium could include exceptions for certain classes of weapons.
Define guiding principles for human involvement in the use of force.
Develop protocols and/or technological means to mitigate the risk of unintentional escalation due to autonomous systems.
Develop strategies for preventing proliferation to illicit uses, such as by criminals, terrorists, or rogue states.
Conduct research to improve technologies and human-machine systems to reduce non-combatant harm and ensure IHL compliance in the use of future weapons.

Component 1:

States should consider adopting a five-year, renewable moratorium on the development, deployment, transfer, and use of anti-personnel lethal autonomous weapon systems. Anti-personnel lethal autonomous weapon systems are defined as weapons systems that, once activated, can select and engage dismounted human targets without further intervention by a human operator, possibly excluding systems such as:

Fixed-point defensive systems with human supervisory control to defend human-occupied bases or installations
Limited, proportional, automated counter-fire systems that return fire in order to provide immediate, local defense of humans
Time-limited pursuit deterrent munitions or systems
Autonomous weapon systems with size above a specified explosive weight limit that select as targets hand-held weapons, such as rifles, machine guns, anti-tank weapons, or man-portable air defense systems, provided there is adequate protection for non-combatants and ensuring IHL compliance5

The moratorium would not apply to:

Anti-vehicle or anti-materiel weapons
Non-lethal anti-personnel weapons
Research on ways of improving autonomous weapon technology to reduce non-combatant harm in future anti-personnel lethal autonomous weapon systems
Weapons that find, track, and engage specific individuals whom a human has decided should be engaged within a limited predetermined period of time and geographic region

Motivation:

This moratorium would pause development and deployment of anti-personnel lethal autonomous weapons systems to allow states to better understand the systemic risks of their use and to perform research that improves their safety, understandability, and effectiveness. Particular objectives could be to:

ensure that, prior to deployment, anti-personnel lethal autonomous weapons can be used in ways that are equal to or outperform humans in their compliance with IHL (other conditions may also apply prior to deployment being acceptable);
lay the groundwork for a potentially legally binding diplomatic instrument; and
decrease the geopolitical pressure on countries to deploy anti-personnel lethal autonomous weapons before they are reliable and well-understood.

Compliance Verification:

As part of a moratorium, states could consider various approaches to compliance verification. Potential approaches include:

Developing an industry cooperation regime analogous to that mandated under the Chemical Weapons Convention, whereby manufacturers must know their customers and report suspicious purchases of significant quantities of items such as fixed-wing drones, quadcopters, and other weaponizable robots.
Encouraging states to declare inventories of autonomous weapons for the purposes of transparency and confidence-building.
Facilitating scientific exchanges and military-to-military contacts to increase trust, transparency, and mutual understanding on topics such as compliance verification and safe operation of autonomous systems.
Designing control systems to require operator identity authentication and unalterable records of operation; enabling post-hoc compliance checks in case of plausible evidence of non-compliant autonomous weapon attacks.
Relating the quantity of weapons to corresponding capacities for human-in-the-loop operation of those weapons.
Designing weapons with air-gapped firing authorization circuits that are connected to the remote human operator but not to the on-board automated control system.
More generally, avoiding weapon designs that enable conversion from compliant to non-compliant categories or missions solely by software updates.
Designing weapons with formal proofs of relevant properties—e.g., the property that the weapon is unable to initiate an attack without human authorization. Proofs can, in principle, be provided using cryptographic techniques that allow the proofs to be checked by a third party without revealing any details of the underlying software.
Facilitate access to (non-classified) AI resources (software, data, methods for ensuring safe operation) to all states that remain in compliance and participate in transparency activities.

Component 2:

Define and universalize guiding principles for human involvement in the use of force.

Humans, not machines, are legal and moral agents in military operations.
It is a human responsibility to ensure that any attack, including one involving autonomous weapons, complies with the laws of war.
Humans responsible for initiating an attack must have sufficient understanding of the weapons, the targets, the environment and the context for use to determine whether that particular attack is lawful.
The attack must be bounded in space, time, target class, and means of attack in order for the determination about the lawfulness of that attack to be meaningful.
Militaries must invest in training, education, doctrine, policies, system design, and human-machine interfaces to ensure that humans remain responsible for attacks.

Component 3:

Develop protocols and/or technological means to mitigate the risk of unintentional escalation due to autonomous systems.

Specific potential measures include:

Developing safe rules for autonomous system behavior when in proximity to adversarial forces to avoid unintentional escalation or signaling. Examples include:

No-first-fire policy, so that autonomous weapons do not initiate hostilities without explicit human authorization.
A human must always be responsible for providing the mission for an autonomous system.
Taking steps to clearly distinguish exercises, patrols, reconnaissance, or other peacetime military operations from attacks in order to limit the possibility of reactions from adversary autonomous systems, such as autonomous air or coastal defenses.

Developing resilient communications links to ensure recallability of autonomous systems. Additionally, militaries should refrain from jamming others’ ability to recall their autonomous systems in order to afford the possibility of human correction in the event of unauthorized behavior.

Component 4:

Develop strategies for preventing proliferation to illicit uses, such as by criminals, terrorists, or rogue states:

Targeted multilateral controls to prevent large-scale sale and transfer of weaponizable robots and related military-specific components for illicit use.
Employ measures to render weaponizable robots less harmful (e.g., geofencing; hard-wired kill switch; onboard control systems largely implemented in unalterable, non-reprogrammable hardware such as application-specific integrated circuits).

Component 5:

Conduct research to improve technologies and human-machine systems to reduce non-combatant harm and ensure IHL-compliance in the use of future weapons, including:

Strategies to promote human moral engagement in decisions about the use of force
Risk assessment for autonomous weapon systems, including the potential for large-scale effects, geopolitical destabilization, accidental escalation, increased instability due to uncertainty about the relative military balance of power, and lowering thresholds to initiating conflict and for violence within conflict
Methodologies for ensuring the reliability and security of autonomous weapon systems
New techniques for verification, validation, explainability, characterization of failure conditions, and behavioral specifications.

About the Authors (in alphabetical order)

Ronald Arkin directs the Mobile Robot Laboratory at Georgia Tech.

Leslie Kaelbling is co-director of the Learning and Intelligent Systems Group at MIT.

Stuart Russell is a professor of computer science and engineering at UC Berkeley.

Dorsa Sadigh is an assistant professor of computer science and of electrical engineering at Stanford.

Paul Scharre directs the Technology and National Security Program at the Center for a New American Security (CNAS).

Bart Selman is a professor of computer science at Cornell.

Toby Walsh is a professor of artificial intelligence at the University of New South Wales (UNSW) Sydney.

The authors would like to thank Max Tegmark for organizing the three-day meeting from which this document was produced.

1 Autonomous Weapons System (AWS): A weapon system that, once activated, can select and engage targets without further intervention by a human operator. BACK TO TEXT↑

2 There is no implication that some authors would not personally support stronger recommendations. BACK TO TEXT↑

3 For ease of use, this working paper will frequently shorten “autonomous weapon system” to “autonomous weapon.” The terms should be treated as synonymous, with the understanding that “weapon” refers to the entire system: sensor, decision-making element, and munition. BACK TO TEXT↑

4 Anti-personnel lethal autonomous weapon system: A weapon system that, once activated, can select and engage dismounted human targets with lethal force and without further intervention by a human operator. BACK TO TEXT↑

5 The authors are not unanimous about this item because of concerns about ease of repurposing for mass-casualty missions targeting unarmed humans. The purpose of the lower limit on explosive payload weight would be to minimize the risk of such repurposing. There is precedent for using explosive weight limit as a mechanism of delineating between anti-personnel and anti-materiel weapons, such as the 1868 St. Petersburg Declaration Renouncing the Use, in Time of War, of Explosive Projectiles Under 400 Grammes Weight. BACK TO TEXT↑ Continue reading

Posted in Human Robots