AI Just Learned How to Boost the Brain’s Memory

[Note:  This item comes from friend Steve Schear.  DLH]

AI Just Learned How to Boost the Brain’s Memory
Feb 6 2018

When it comes to black boxes, there is none more black than the human brain. Our gray matter is so complex, scientists lament, that it can’t quite understand itself.

But if we can’t grok our own brains, maybe the machines can do it for us. In the latest issue of Nature Communications, researchers led by University of Pennsylvania psychologist Michael Kahana show that machine learning algorithms—notoriously inscrutable systems themselves—can be used to decode and then enhance human memory. How? By triggering the delivery of precisely timed pulses of electricity to the brain.

Researchers, in other words, can use one black box to unlock the potential of another. Which on one hand sounds like a rather elegant solution to an absurdly difficult problem, and on the other sounds like the beginning of a techno-pocalypse horror flick.

When it comes to brain measurements, the best recordings come from inside the cranium. But people—and institutional review boards—aren’t usually amenable to cracking open skulls in the name of science. So Kahana and his colleagues collaborated with 25 epilepsy patients, each of whom had between 100 and 200 electrodes implanted in their brain (to monitor seizure-related electrical activity). Kahana and his team piggybacked on those implants, using the electrodes to record high-resolution brain activity during memory tasks.

First, the researchers got a sense of what it looks like when a brain memorizes stuff. As the patients read and attempted to internalize lists of words, Kahana and his team gathered thousands of voltage measurements per second from each of the implanted electrodes. Later, they tested the patients’ recall—building up data about which brain activity patterns were associated with remembering a word vs. forgetting it.

Then they did it again. And again. After two or three visits with each test subject, they’d collected enough training data to produce patient-specific algorithms that could predict which words each patient would likely remember—based on their electrode activity alone.

Here’s the kicker. These electrodes don’t just read neural activity; they can stimulate it, too. So the researchers tried prodding the brain to improve—or, as they put it, “rescue”—the formation of memories in real time. Every few seconds, the subject would see a new word, and the newly trained algorithm would decide whether the brain was ready to remember it. “A closed loop system lets us record the state of the subject’s brain, analyze it, and decide whether to trigger a stimulation, all in a few hundred milliseconds,” Kahana says.

And it worked. The researchers’ system improved patients’ ability to recall words by an average of 15 percent.

This isn’t the first time Kahana’s lab has explored the impacts of brain stimulation on memory. Last year, the group showed that electrode pulses seemed to improve or worsen recall, depending on when the researchers delivered them. In that study, test subjects scored higher when the researchers stimulated memory-specific regions of the brain during periods of low functionality (stimulation during high-functioning times had the opposite effect). It was a major finding, but therapeutically useless; the researchers could only identify the link between memory and brain states after the memory tests were performed. What you really want, from a brain-enhancement standpoint, is to deliver pulses in the middle of memorization.

Now, Kahana and his colleagues appear to have closed the loop with the help of their machine learning algorithm. “Only instead of using it to identify images of cats, we’re using it to build a decoder—something that can look at electrical activity and say whether the brain is in a state that’s conducive to learning,” Kahana says. If the brain looks like it’s encoding memories effectively, the researchers leave it alone. If it isn’t, their system quickly delivers electrical pulses to jostle it into a higher-functioning state—like a pacemaker for the brain.



Democrats can’t expand opportunity without reducing inequality

[Note:  This item comes from friend Robert Berger.  DLH]

Democrats can’t expand opportunity without reducing inequality
By Elizabeth Bruenig
Mar 18 2018

THE UNEXPECTED outcome of the 2016 presidential election has put Democrats and other left-of-center people of goodwill in a uniquely opportune position. They have the chance to reimagine what the character of mainstream-left politics will be in the United States, and to articulate those principles anew to a populace that appears to be listening. But what to imagine, and what to say? 

As The Post’s James Hohmann recently reported, the center-left Third Way think tank has proposed that, in the words of Senior Vice President for Policy Jim Kessler, “Democrats need a positive, modern, forward-looking economic agenda that gives people hope that they’re going to have the opportunity to earn a good life in the future.” But Kessler also “noted that while the Bernie Sanders wing of the Democratic Party is focused on income inequality, Third Way has decided to focus on the concentration of opportunities.” In other words, the mainstream left has a decision to make: focus on inequality or focus on opportunity? 

But, speaking as a member of the “Bernie Sanders wing,” I don’t actually view the two focal points — opportunity and inequality — as fully separable. The impact of inequality on the overall economy and political climate means that, unless inequality itself is reduced, individual initiative won’t have much impact and conditions for ordinary people in America will continue, in key ways, to worsen.

Sanders himself agrees. “You have an economic situation where a tiny number of people have enormous amounts of wealth,” Sanders told me at his Senate office Wednesday morning, “and politically, you have the Koch brothers and a handful of billionaires buying elections.” Sanders pointed out that the top 1 percent of earners would rake in roughly 83 percent of the benefits of President Trump’s tax cuts, a series of policies for which Koch-led groups spent more than $20 million. As Yale University political scientist Jacob Hacker pointed out in The Post, the decision to supply a generous tax cut to the rich — greatly encouraged by the vast wealth of billionaire donors — is a de facto decision to reduce expenditures that help ordinary Americans, “like public investments in infrastructure, education, research and development, and the regulation of labor and financial markets.” Put simply, inequality allows the wealthiest Americans to exert undue control over politics, thereby maintaining the conditions that made them rich in the first place, and hamstringing government efforts that could increase opportunities for the rest of us.

And inequality weakens the economy itself. Economic researchers have argued that American inequality worsened and prolonged the negative effects of the Great Recession and that continued inequality has left less well-off families more vulnerable to economic shocks going forward, meaning that future crises might unfold even more disastrously than the last. 

Most important, rising inequality can swamp the gains that broadening opportunity is supposed to deliver. A recent study by Harvard University researcher Robert Manduca found that, while the median black family had risen from the 25th percentile of the overall income distribution in 1968 to the 35th percentile in 2016, and closed the black-white gap in income rank by about 28 percent, the concurrent rise in overall inequality was so dramatic that it completely negated the black gains. The ratio between black and white income, in other words, remained unchanged: As the distribution of income skews ever more toward top of the spectrum, succeeding in climbing up from the bottom doesn’t amount to much.


The Perfect Selfishness of Mapping Apps

The Perfect Selfishness of Mapping Apps
Apps like Waze, Google Maps, and Apple Maps may make traffic conditions worse in some areas, new research suggests.
Mar 15 2018

What is the price of anarchy?

Technically, in transportation engineering, the price of anarchy describes the difference between what happens when every driver selfishly picks the fastest route and what the socially optimal traffic outcome would be.

In the pre-mobile-app days, drivers’ selfishness was limited by their knowledge of the road network. In those conditions, both simulation and real-world experience showed that most people stuck to the freeways and arterial roads. Sure, there were always people who knew the crazy, back-road route, but the bulk of people just stuck to the routes that transportation planners had designated as the preferred way to get from A to B.

Now, however, a new information layer is destroying the nudging infrastructure that traffic planners built into cities. Commuters armed with mobile mapping apps, route-following Lyft and Uber drivers, and software-optimized truckers can all act with a more perfect selfishness.

In some happy universe, this would lead to socially optimal outcomes, too. But a new body of research at the University of California’s Institute of Transportation Studies suggests that the reality is far more complicated. In some scenarios, traffic-beating apps might work for an individual, but make congestion worse overall. And autonomous vehicles, touted as an answer to traffic-y streets, could deepen the problem.

“This problem has been vastly overlooked,” Alexandre Bayen, the director of UC Berkeley’s Institute of Transportation Studies, told me. “It is just the beginning of something that is gonna be much worse.”

Bayen and a team of researchers presented their work earlier this year at the Transportation Research Board’s annual meeting and at the Cal Future conference at Berkeley in May 2017. They’ve also published work examining the negative externalities of high levels of automatic routing.

In the Cal Future talk, Bayen walked through a simulation created in the commercial-transportation simulator Aimsun. The video below shows how the flow of a freeway changes in response to an accident under two conditions: when no drivers use routing apps and when only 20 percent of drivers use routing apps. When there are more app-using drivers, congestion builds up at off-ramps, creating more traffic on the freeway.

“The situation then gets much worse because hundreds of people just like you want to go on the side streets, which were never designed to handle the traffic,” Bayen says. “So, now, in addition to congesting the freeway, you’ve also congested the side streets and the intersections.”

While it’s clear that traffic on local roads gets worse with the use of these apps, Bayen said that nobody has managed to do a multi-scale analysis that can determine if the apps, even if they create local problems, are better or worse for whole traffic basins.

Nonetheless, the adoption and use of these apps continues to grow, too, Bayen said. Over the last 10 years, traffic-routing apps have become a standard accessory for the driving public. According to a 2015 Pew survey, 90 percent of Americans with smartphones use maps for driving directions at least some of the time. As smartphone penetration reaches up above 70 percent, a vast number of people now have access to real-time traffic data on their phones. The driving public is better informed about routes and road conditions than ever before.

In many local neighborhoods, residents have complained about increased traffic volumes. Bayen’s team has been gathering stories from around California of resident complaints and found dozens in even a short time window, as shown in the map below.


Investors in Bitcoin and other cryptocurrencies face hefty tax bills

Investors in Bitcoin and other cryptocurrencies face hefty tax bills
According to the Internal Revenue Service, anything purchased using a digital currency is liable to be taxed as a capital gain
By Edward Helmore
Mar 18 2018

The rollercoaster ride for some cryptocurrency investors could be about to take another tax-time lurch, according to experts, as the taxman looks for his share of transactions made using bitcoin and its like. 

Wild fluctuations in the value of digital currencies – bitcoin surged from less than one dollar in 2010 to $997 at the start of the 2017 to nearly $20,000 before settling back to around $8,500 on Friday – have exposed investors to tax bills the value of their coins may no longer meet. 

On Reddit earlier this week, one contributor, under the heading “I just discovered that I owe the IRS $50k that I don’t have, because I traded in cryptos. Am I fucked?”, wrote they had ended up with a $50,000 tax liability on trades after they sold $120,000 worth of bitcoin to buy different coins. The current value of those coins is about $30,000. “I feel like I might have accidentally ruined my life because I didn’t know about the taxes,” the poster wrote. One complication for crypto investors is that digital currencies that were, in part, devised to operate outside of government and banking industry oversight, are still of interest to the US tax authorities, who look at cryptocurrency as property and not currency.

According to the Internal Revenue Service, anything purchased using a digital currency is liable to be taxed as a capital gain. So anyone who has cashed out or paid for anything using cryptocurrency may have capital gains to report to the IRS.

Another source of confusion is that crypto-brokers are not required to issue 1099 disclosure forms – the forms used by the IRS to report income other than wages, bonuses and tips – on digital currencies, but individuals are still responsible for reporting gains.

In November, a US district court judge in California ordered Coinbase, a popular platform for trading bitcoin, to turn over identifying information on accounts worth at least $20,000 during 2013 to 2015.

The IRS case came about after the agency discovered that only about 800 taxpayers claimed bitcoin gains in each year from 2013 to 2015. But the Coinbase agreement only affects about 10,000 accounts, not the 480,000 accounts the IRS first requested.

Not reporting gains, it should be said, could amount to tax evasion.

And the capital gains ruling is not the only crypto-complication. If an investor sells a cryptocurrency after holding it longer than a year, then the profits are typically long-term capital gains. Nor are losses deductible against future tax years.

William Perez, a tax accountant at the online tax filing and advisory service Visor, has noticed that accountants are often unwilling to familiarize themselves with crypto-accounting rulings. “Among crypto-investors, I see resistance to reporting it,” he says. “Then there’s another group who’ve got a 1099 from Coinbase but they don’t know what it means.”

Investors are getting caught out in basic ways. For instance, crypto-to-crypto transactions are taxable – if, for example, you use your bitcoin to buy rival ethereum. “That often catches people off guard, but once you break it out you’ve sold one coin and invested in another. That’s one bear trap,” said Perez.

The second bear trap, Perez explains, is when crypto is used for purchasing. But crypto is not like PayPal or a gift card, and not merely a conduit of exchange. “Under accounting rules, you have property that you exchanged for something else. 

“People think because they’ve paid a sales tax so that’s the end of the story. But it’s not. We’re talking about a property denominated in dollars. If you exchange that then there’s a tax liability.”

The IRS rules on crypto, Perez says, are straightforward. He anticipates the agency will leave the preliminary guidelines (issued in 2014) in place for a few years to see how they work out.


Take This App and Call Me in the Morning

Take This App and Call Me in the Morning
A new category of prescription medical treatments, what executives call digital therapeutics, comes in the form of mobile apps.
By Natasha Singer
Mar 18 2018

Health tech companies are making a big push to digitize medicine, introducing novel tools like digital pills that track when patients take their drugs and smart spoons that can automatically adjust to hand tremors.

Now they want some patients to get prescription treatments from the app store as well.

Later this year, doctors treating patients addicted to substances like cocaine and amphetamines will be able to prescribe Reset, an app that gives patients lessons to help them modify their behavior. The Food and Drug Administration cleared it in September as the first mobile medical app to help treat substance-use disorders.

“It’s all the things you would traditionally associate with a pill or any other medication,” said Dr. Corey McCann, the chief executive of Pear Therapeutics, the start-up behind Reset. “But it just so happens to be a piece of software.”

Pear Therapeutics is at the forefront of a new category of medical treatment, offering what company executives call “prescription digital therapeutics.” These products, they say, are medical apps that have been studied in randomized clinical trials, cleared by the Food and Drug Administration, require a doctor’s prescription and allow doctors to track patients’ progress.

This month, in a sign of momentum behind the idea and how it might expand to other health conditions, Novartis, the Swiss pharmaceutical giant, announced that it had teamed up with Pear Therapeutics. Together, they plan to develop prescription apps for schizophrenia and multiple sclerosis.

Companies like Pear are trying to stand apart in a global market of more than 318,000 health apps by arguing that their products provide assurance of effectiveness and safety. The advantage of their prescription treatment software, they say, is that apps making general wellness claims — like meditation apps marketed to soothe you to sleep — do not require F.D.A. review. But medical devices like apps that claim to treat or manage disease must submit clinical evidence to the F.D.A. and get clearance.

A few medical experts, however, argue that the apps, while promoting themselves as a new medical therapy, are essentially just repackaging and rebranding existing treatments. They note that behavioral therapy done on a computer, instead of on a smartphone, was already a longstanding health treatment and backed by research.

“This is a branding effort that is meaningless,” said Dr. Allen J. Frances, a psychiatrist who is a professor emeritus at the Duke University School of Medicine. “F.D.A. approval comes easily, and a prescription doesn’t guarantee greater efficacy.”

Dr. McCann, a former neuroscientist and venture capitalist, co-founded Pear Therapeutics in 2013. The start-up has since raised $70 million and licensed a variety of digital therapeutics from researchers and other companies. The Reset mobile app, for instance, is based on a web-based addiction therapy program. It was originally developed in the late 1990s by behavior modification researchers who digitized long-established methods of in-person addiction therapy.

Reset is not the first prescription mobile medical app. The F.D.A. previously cleared software like BlueStar Rx, a prescription diabetes management app.

But Reset is different because its primary focus is not disease management. It delivers an established behavior-modification treatment for addiction — which traditionally involves face-to-face outpatient therapy — entirely in digital form. In other words, the app itself is the medicine.

Anyone will be able to download the Reset app, expected to be available in the second half of the year. But to get it to work on a smartphone, patients must enter a prescription access code. A prescription will be good for 90 days. The F.D.A. cleared the app to be used in conjunction with outpatient therapy.

The Reset app contains 61 therapy units, with titles like “Coping With Thoughts About Using,” including lessons, skills-building exercises and quizzes on the material. The recommended dose: four therapy lessons per week.

The idea behind the app is to teach patient skills — such as avoiding situations that trigger their substance use or learning how to refuse an offer of drugs — to modify their behavior, thereby promoting abstinence.


FedEx Follows Amazon Into the Robotic Future

FedEx Follows Amazon Into the Robotic Future
Mar 18 2018

KERNERSVILLE, N.C. — As soon as the first robot arrived at a FedEx shipping hub in the heart of North Carolina tobacco country early last year, talk of pink slips was in the air.

Workers had been driving the “tuggers” that navigated large and irregular items across the vast concrete floor of the 630,000-square-foot freight depot since it opened in 2011.

Their initial robotic colleague drew a three-dimensional digital map of the place as it tugged freight around. A few months later, three other robots — nicknamed Lucky, Dusty and Ned in a nod to the movie “¡Three Amigos!” — arrived, using the digital map to get around on their own.

By March, they were joined by two others, Jefe and El Guapo. Horns honking and warning lights flashing, the autonomous vehicles snaked through the hub, next to about 20 tuggers that still needed humans behind the wheel.

The robot team, part of the automation trend rippling through the worlds of shipping and online retail, was the first significant deployment of mobile robots inside a FedEx hub. Amazon and our e-commerce shopping habits are big reasons it’s happening.

In 2012, Amazon acquired a robotics company called Kiva. Since then, it has moved many of that company’s robots into its network of more than 210 fulfillment and package-sorting centers. Now, many Amazon partners and competitors are moving in the same direction, including big shipping and logistics operations like FedEx and DHL.

But what has happened at the FedEx hub may be a surprise to people who fear that they are about to be replaced by a smart machine: a robot might take your role, but not necessarily your job.

Yes, the robots replaced a few jobs right away. And in time, they will replace about 25 jobs in a facility that employs about 1,300 people. But the hub creates about 100 new jobs every year — and a robot work force still seems like the distant future.

“Everyone will have a job,” said Galen Steele, the senior manager who oversees the depot. “It just might be in a different place.”

As people have become more comfortable buying online, big and bulky goods like car tires, canoes and boxes as big as a coffin have accounted for an increasing percentage of the packages flowing through FedEx’s distribution centers, said Ted Dengel, who oversees operations technology for the FedEx Ground network, which includes 35 shipping hubs across the United States and Canada, including the facility in North Carolina.

These ungainly items can’t fit on a conveyor belt. That’s where the robots, which cost several thousand dollars and are made by a Massachusetts company called Vecna, come in.

Kernersville, the home of the FedEx hub, sits in the middle of the North Carolina Piedmont, the plateau between the coastal lowlands in the east and the mountains in the west, in an area that used to be known as the Tobacco Road. In the 20th century, it embraced textile and apparel manufacturing. But as automation has increased inside factories and textile work has moved overseas, the local economy has sputtered.

“In the middle of a boom elsewhere, things really aren’t that great here,” said Andrew Brod, senior research fellow at the University of North Carolina, Greensboro, who tracks the local economy. The local unemployment rate is 4.6 percent, modestly above the 4.1 percent national rate.

Mr. Brod estimates that the arrival of distribution centers like the one FedEx opened in 2011 has provided a boost, creating as many as 120,000 jobs, a majority of which go to workers with no more than a high school education.

And Mr. Steele, the manager of the FedEx facility in Kernersville, did not envision pink slips when he saw the first Vecna robot.

“I understand people thinking this will take their jobs,” Mr. Steele said on a recent Monday morning inside the hub. “But over time, they realize that is not the case at all.”


Army, Struggling to Get Technology in Soldiers’ Hands, Tries the Unconventional

Army, Struggling to Get Technology in Soldiers’ Hands, Tries the Unconventional
Mar 18 2018

WASHINGTON — The platoon of Army Special Operations soldiers was on a routine night patrol in eastern Afghanistan when one of them suddenly opened fire on what looked to the others to be a bush.

The bush, it turned out, had been obscuring a militant fighter. He was detectable only to the one platoon member wearing prototype night vision goggles that could detect heat signatures — a happenstance that Army officials say probably saved many lives.

That incident took place in 2015. Three years later, soldiers in the field still do not have the new night vision goggles, and that is just one example of a process that can take a decade to get new weapons from the lab to the hands of troops. Worried about that lag, the Army is creating a new and decidedly unconventional department to address it: the Futures Command.

“Washington and Marshall are looking over me like ghosts,” said Ryan McCarthy, a former Ranger who is now the Army under secretary, in a reference to George Washington and George Marshall, two of the most famous proponents of an American Army that keeps ahead of all adversaries. “Things can take too long; historically services will experiment things to death and never buy anything, or don’t experiment and then buy a billion-dollar PowerPoint. We have to move away from that.”

So on March 26, top Army leaders will travel to Huntsville, Ala., to announce details of their plan for the Futures Command, which will focus solely on developing new weapons and getting them “downrange” faster. The doors of the command are expected to open by the end of July, and it is supposed to be fully operational a year after that.

The Futures Command, Army leaders say, is part of a movement to get the Army, focused for nearly two decades on fighting Islamist militants in Afghanistan and Iraq, in shape to fight a potential great-power land war.

At the Pentagon, the talk inside the military’s biggest service is all about “modernization” and “readiness,” the favorite children of the Army chief of staff, Gen. Mark A. Milley. Besides speeding up the lengthy procurement process, Army leaders want to get rid of layers of bureaucracy that can eat away at the military’s competitive edge.

And they want to enlist the talents of Americans who may not necessarily see themselves as bound for the military, by locating the Futures Command not at a traditional military base like Fort Hood, Tex., or Fort Sill, Okla., but in a city with easy access to big universities and cutting-edge technological research.

Seattle, San Francisco and Boston are all among the cities Army planners say they are looking at for the Futures Command. Mark T. Esper, the secretary of the Army, said that whoever is chosen to lead the command will be “not a traditional person who spent all their time in combat units, but someone who understands the acquisition process, and who understands the corporate Army.” Ten cities should be selected within the next two months as finalists for hosting the command.

The Army is making many of the changes in an effort to figure out how to fight multiple types of wars at the same time.

More than 16 years of counterinsurgency warfare has taught military leaders one thing: The burden of fighting jihadists in multiple places is unlikely to go away anytime soon. Even today, almost eight years after President Barack Obama announced the end of combat in Iraq and three years after he announced the end of combat in Afghanistan, American troops are deployed in both countries, not to mention Syria, Niger, Somalia, Libya and Mali.

But at the same time, the Army has been charged with getting ready for a possible war on the Korean Peninsula. President Trump may have agreed to one-on-one talks with North Korea’s leader, Kim Jong-un, but the Pentagon is still trying to make sure that American forces are ready for what would probably be a completely different kind of war.

General Milley, the Army chief of staff, has repeatedly expressed concern that the Army has lost what he calls its “muscle memory” of how to fight big land wars. Beyond that, Army leaders — many of them students of history — say that for almost 200 years, the Army has often gotten the first battle of major wars wrong.