A future app may help you avoid getting the flu by informing you of your local risk on a given day.
Applied mathematician Sara del Valle works at the U.S.'s foremost nuclear weapons lab: Los Alamos. Once colloquially called Atomic City, it's a hidden place 45 minutes into the mountains northwest of Santa Fe. Here, engineers developed the first atomic bomb.
Like AccuWeather, an app for disease prediction could help people alter their behavior to live better lives.
Today, Los Alamos still a small science town, though no longer a secret, nor in the business of building new bombs. Instead, it's tasked with, among other things, keeping the stockpile of nuclear weapons safe and stable: not exploding when they're not supposed to (yes, please) and exploding if someone presses that red button (please, no).
Del Valle, though, doesn't work on any of that. Los Alamos is also interested in other kinds of booms—like the explosion of a contagious disease that could take down a city. Predicting (and, ideally, preventing) such epidemics is del Valle's passion. She hopes to develop an app that's like AccuWeather for germs: It would tell you your chance of getting the flu, or dengue or Zika, in your city on a given day. And like AccuWeather, it could help people alter their behavior to live better lives, whether that means staying home on a snowy morning or washing their hands on a sickness-heavy commute.
Sara del Valle of Los Alamos is working to predict and prevent epidemics using data and machine learning.
Since the beginning of del Valle's career, she's been driven by one thing: using data and predictions to help people behave practically around pathogens. As a kid, she'd always been good at math, but when she found out she could use it to capture the tentacular spread of disease, and not just manipulate abstractions, she was hooked.
When she made her way to Los Alamos, she started looking at what people were doing during outbreaks. Using social media like Twitter, Google search data, and Wikipedia, the team started to sift for trends. Were people talking about hygiene, like hand-washing? Or about being sick? Were they Googling information about mosquitoes? Searching Wikipedia for symptoms? And how did those things correlate with the spread of disease?
It was a new, faster way to think about how pathogens propagate in the real world. Usually, there's a 10- to 14-day lag in the U.S. between when doctors tap numbers into spreadsheets and when that information becomes public. By then, the world has moved on, and so has the disease—to other villages, other victims.
"We found there was a correlation between actual flu incidents in a community and the number of searches online and the number of tweets online," says del Valle. That was when she first let herself dream about a real-time forecast, not a 10-days-later backcast. Del Valle's group—computer scientists, mathematicians, statisticians, economists, public health professionals, epidemiologists, satellite analysis experts—has continued to work on the problem ever since their first Twitter parsing, in 2011.
They've had their share of outbreaks to track. Looking back at the 2009 swine flu pandemic, they saw people buying face masks and paying attention to the cleanliness of their hands. "People were talking about whether or not they needed to cancel their vacation," she says, and also whether pork products—which have nothing to do with swine flu—were safe to buy.
At the latest meeting with all the prediction groups, del Valle's flu models took first and second place.
They watched internet conversations during the measles outbreak in California. "There's a lot of online discussion about anti-vax sentiment, and people trying to convince people to vaccinate children and vice versa," she says.
Today, they work on predicting the spread of Zika, Chikungunya, and dengue fever, as well as the plain old flu. And according to the CDC, that latter effort is going well.
Since 2015, the CDC has run the Epidemic Prediction Initiative, a competition in which teams like de Valle's submit weekly predictions of how raging the flu will be in particular locations, along with other ailments occasionally. Michael Johannson is co-founder and leader of the program, which began with the Dengue Forecasting Project. Its goal, he says, was to predict when dengue cases would blow up, when previously an area just had a low-level baseline of sick people. "You'll get this massive epidemic where all of a sudden, instead of 3,000 to 4,000 cases, you have 20,000 cases," he says. "They kind of come out of nowhere."
But the "kind of" is key: The outbreaks surely come out of somewhere and, if scientists applied research and data the right way, they could forecast the upswing and perhaps dodge a bomb before it hit big-time. Questions about how big, when, and where are also key to the flu.
A big part of these projects is the CDC giving the right researchers access to the right information, and the structure to both forecast useful public-health outcomes and to compare how well the models are doing. The extra information has been great for the Los Alamos effort. "We don't have to call departments and beg for data," says del Valle.
When data isn't available, "proxies"—things like symptom searches, tweets about empty offices, satellite images showing a green, wet, mosquito-friendly landscape—are helpful: You don't have to rely on anyone's health department.
At the latest meeting with all the prediction groups, del Valle's flu models took first and second place. But del Valle wants more than weekly numbers on a government website; she wants that weather-app-inspired fortune-teller, incorporating the many diseases you could get today, standing right where you are. "That's our dream," she says.
This plot shows the the correlations between the online data stream, from Wikipedia, and various infectious diseases in different countries. The results of del Valle's predictive models are shown in brown, while the actual number of cases or illness rates are shown in blue.
(Courtesy del Valle)
The goal isn't to turn you into a germophobic agoraphobe. It's to make you more aware when you do go out. "If you know it's going to rain today, you're more likely to bring an umbrella," del Valle says. "When you go on vacation, you always look at the weather and make sure you bring the appropriate clothing. If you do the same thing for diseases, you think, 'There's Zika spreading in Sao Paulo, so maybe I should bring even more mosquito repellent and bring more long sleeves and pants.'"
They're not there yet (don't hold your breath, but do stop touching your mouth). She estimates it's at least a decade away, but advances in machine learning could accelerate that hypothetical timeline. "We're doing baby steps," says del Valle, starting with the flu in the U.S., dengue in Brazil, and other efforts in Colombia, Ecuador, and Canada. "Going from there to forecasting all diseases around the globe is a long way," she says.
But even AccuWeather started small: One man began predicting weather for a utility company, then helping ski resorts optimize their snowmaking. His influence snowballed, and now private forecasting apps, including AccuWeather's, populate phones across the planet. The company's progression hasn't been without controversy—privacy incursions, inaccuracy of long-term forecasts, fights with the government—but it has continued, for better and for worse.
Disease apps, perhaps spun out of a small, unlikely team at a nuclear-weapons lab, could grow and breed in a similar way. And both the controversies and public-health benefits that may someday spin out of them lie in the future, impossible to predict with certainty.
Giving robots self-awareness as they move through space - and maybe even providing them with gene-like methods for storing rules of behavior - could be important steps toward creating more intelligent machines.
Now roboticists are going a step further, training their creations to do even better: understand their own image in space and interact with humans like humans do. Today, there are already robot-teachers like KeeKo, robot-pets like Moffin, robot-babysitters like iPal, and robotic companions for the elderly like Pepper.
But even these reasonably intelligent creations still have huge limitations, some scientists think. “There are niche applications for the current generations of robots,” says professor Anthony Zador at Cold Spring Harbor Laboratory—but they are not “generalists” who can do varied tasks all on their own, as they mostly lack the abilities to improvise, make decisions based on a multitude of facts or emotions, and adjust to rapidly changing circumstances. “We don’t have general purpose robots that can interact with the world. We’re ages away from that.”
Robotic spatial self-awareness – the achievement by the team at Columbia – is an important step toward creating more intelligent machines. Hod Lipson, professor of mechanical engineering who runs the Columbia lab, says that future robots will need this ability to assist humans better. Knowing how you look and where in space your parts are, decreases the need for human oversight. It also helps the robot to detect and compensate for damage and keep up with its own wear-and-tear. And it allows robots to realize when something is wrong with them or their parts. “We want our robots to learn and continue to grow their minds and bodies on their own,” Chen says. That’s what Zador wants too—and on a much grander level. “I want a robot who can drive my car, take my dog for a walk and have a conversation with me.”
Columbia scientists have trained a robot to become aware of its own "body," so it can map the right path to touch a ball without running into an obstacle, in this case a square.
Jane Nisselson and Yinuo Qin/ Columbia Engineering
Today’s technological advances are making some of these leaps of progress possible. One of them is the so-called Deep Learning—a method that trains artificial intelligence systems to learn and use information similar to how humans do it. Described as a machine learning method based on neural network architectures with multiple layers of processing units, Deep Learning has been used to successfully teach machines to recognize images, understand speech and even write text.
Trained by Google, one of these language machine learning geniuses, BERT, can finish sentences. Another one called GPT3, designed by San Francisco-based company OpenAI, can write little stories. Yet, both of them still make funny mistakes in their linguistic exercises that even a child wouldn’t. According to a paper published by Stanford’s Center for Research on Foundational Models, BERT seems to not understand the word “not.” When asked to fill in the word after “A robin is a __” it correctly answers “bird.” But try inserting the word “not” into that sentence (“A robin is not a __”) and BERT still completes it the same way. Similarly, in one of its stories, GPT3 wrote that if you mix a spoonful of grape juice into your cranberry juice and drink the concoction, you die. It seems that robots, and artificial intelligence systems in general, are still missing some rudimentary facts of life that humans and animals grasp naturally and effortlessly.
How does one give robots a genome? Zador has an idea. We can’t really equip machines with real biological nucleotide-based genes, but we can mimic the neuronal blueprint those genes create.
It's not exactly the robots’ fault. Compared to humans, and all other organisms that have been around for thousands or millions of years, robots are very new. They are missing out on eons of evolutionary data-building. Animals and humans are born with the ability to do certain things because they are pre-wired in them. Flies know how to fly, fish knows how to swim, cats know how to meow, and babies know how to cry. Yet, flies don’t really learn to fly, fish doesn’t learn to swim, cats don’t learn to meow, and babies don’t learn to cry—they are born able to execute such behaviors because they’re preprogrammed to do so. All that happens thanks to the millions of years of evolutions wired into their respective genomes, which give rise to the brain’s neural networks responsible for these behaviors. Robots are the newbies, missing out on that trove of information, Zador argues.
A neuroscience professor who studies how brain circuitry generates various behaviors, Zador has a different approach to developing the robotic mind. Until their creators figure out a way to imbue the bots with that information, robots will remain quite limited in their abilities. Each model will only be able to do certain things it was programmed to do, but it will never go above and beyond its original code. So Zador argues that we have to start giving robots a genome.
How does one do that? Zador has an idea. We can’t really equip machines with real biological nucleotide-based genes, but we can mimic the neuronal blueprint those genes create. Genomes lay out rules for brain development. Specifically, the genome encodes blueprints for wiring up our nervous system—the details of which neurons are connected, the strength of those connections and other specs that will later hold the information learned throughout life. “Our genomes serve as blueprints for building our nervous system and these blueprints give rise to a human brain, which contains about 100 billion neurons,” Zador says.
If you think what a genome is, he explains, it is essentially a very compact and compressed form of information storage. Conceptually, genomes are similar to CliffsNotes and other study guides. When students read these short summaries, they know about what happened in a book, without actually reading that book. And that’s how we should be designing the next generation of robots if we ever want them to act like humans, Zador says. “We should give them a set of behavioral CliffsNotes, which they can then unwrap into brain-like structures.” Robots that have such brain-like structures will acquire a set of basic rules to generate basic behaviors and use them to learn more complex ones.
Currently Zador is in the process of developing algorithms that function like simple rules that generate such behaviors. “My algorithms would write these CliffsNotes, outlining how to solve a particular problem,” he explains. “And then, the neural networks will use these CliffsNotes to figure out which ones are useful and use them in their behaviors.” That’s how all living beings operate. They use the pre-programmed info from their genetics to adapt to their changing environments and learn what’s necessary to survive and thrive in these settings.
For example, a robot’s neural network could draw from CliffsNotes with “genetic” instructions for how to be aware of its own body or learn to adjust its movements. And other, different sets of CliffsNotes may imbue it with the basics of physical safety or the fundamentals of speech.
At the moment, Zador is working on algorithms that are trying to mimic neuronal blueprints for very simple organisms—such as earthworms, which have only 302 neurons and about 7000 synapses compared to the millions we have. That’s how evolution worked, too—expanding the brains from simple creatures to more complex to the Homo Sapiens. But if it took millions of years to arrive at modern humans, how long would it take scientists to forge a robot with human intelligence? That’s a billion-dollar question. Yet, Zador is optimistic. “My hypotheses is that if you can build simple organisms that can interact with the world, then the higher level functions will not be nearly as challenging as they currently are.”
