Recent News & Events

Donate Your Voice to Charity

Earlier this year, Stephen Hawking, who has relied on the same computerized system to communicate for more than 20 years, received a much-needed upgrade. The system, which allowed Hawking to translate text into speech via a sensor on his cheek, had become prohibitively slow as the physicist’s progressing ALS left him with weakened control of his facial muscles.

When Intel began work on his newer, faster computer, Hawking had one requirement: The software could change, but the sound of his speech had to remain the same.

“The voice has become so iconic that he considers that his own personal voice,” Horst Haussecker, the director of Intel’s Computational Imaging Lab and a leader of the project, recently told NPR. “It’s based on, you know, slightly outdated technology, but it makes it very unique and you couldn’t copy it even if you wanted to.”

“It is the best I have heard,” Hawking wrote of his iconic voice on his personal website, “although it gives me an accent that has been described variously as Scandinavian, American, or Scottish.”

But even a case of mistaken nationality isn’t enough to damage the link between the sound of the machine and the man it speaks for; over time, one has become emblematic of the other. Stephen Hawking does not sound like a computer—Stephen Hawking sounds like Stephen Hawking.

Most people who cannot speak, though, do not have the luxury of being Stephen Hawking.

* * *

An estimated eight out of every 1,000 Americans, or 2.5 million people, are severely speech-impaired due to a variety of conditions: head injuries, congenital disorders like cerebral palsy, or degenerative diseases like Hawking’s ALS. Many of them rely on text-to-speech machines, typing words that are then vocalized electronically. They sound like computers. And because computers are manufactured in batches of more than one, they also sound like each other.

In August 2002, Rupal Patel, a speech-science professor at Northeastern University, was at a speech-technology conference in Odese, Denmark to present the results of her latest research. People with dramatic speech impairment, she had found, were still able to control the melody of their voices (also called the “prosody” of a voice, or its pitch, tempo, and volume) even when they couldn’t form words; as a result, many people forewent their communication devices when talking to those closest to them, relying on inflection to help convey meaning.

Walking through the conference’s exhibition hall after her presentation, Patel passed a young woman and older man engaged in conversation, their voices indistinguishable from one another—both were using the same text-to-speech system.

Patel paused, listened. The same sound, she realized, was all around her. People throughout the hall—“nearly half the room,” she recalls—were using nearly identical voices.

“That’s when I put two and two together,” she says. “I thought, well, if they have this part of their voice that’s preserved, maybe I would be able to build a voice for them.”

The idea stayed with her. For the next few years, Patel developed and fine-tuned her process, and in 2007 she received a grant from the National Science Foundation to pursue the project that would become VocaliD (pronounced “vocality”), a for-profit company that creates personalized voices for text-to-speech systems by blending sounds taken from speech-impaired people with words recorded by healthy donors. (The price of a voice, she says, will ultimately depend on demand.)

The company’s technology is based on the “source-filter theory,” which breaks the production of human speech into two components. One is the source, or the sound made by the vibrations of the vocal cords. The other is the filter, or the vocal tract: the path of these vibrations as they echo through the chambers of the neck and head. Conditions that cause speech impairment mainly affect the filter; the prosody of a voice is controlled by the source, which is usually left intact.

To create a voice, Patel says, “we’re taking the filter, the shape of the vocal tract, from the voice donor, and the source from the individual who’s given us something as limited as a vowel.” After taking a short recording from a recipient—who often can only vocalize as much as an “ahhh” sound—the VocaliD team selects a donor with a similar filter and uses a computer algorithm to layer one over the other. Donations come via the company’s “voice bank,” which opened to the public over Thanksgiving weekend. To donate, a person needs a computer, a microphone, and a few hours of time to record the hundreds of sentences Patel has compiled from old stories and common phrases to encompass all of the sounds of the English language.

From there, she explains, “we chop that blended voice into little snippets of speech that can be rearranged any way, by gluing together little bits of a sentence.”

Patel estimates that somewhere between 500 and 600 people have already donated their voices, and that around 24,000 people have signed up to donate in the future—a number she hopes will allow the team to more effectively pair a recipient with a voice.

“In the past, we were doing some really [basic] matching, like age and gender,” she says. “We’re developing some new techniques to do more sophisticated matching for the kind of voice you have,” taking into account things like “voice quality,” or hoarseness; regional accent; and height and weight, both of which affect the vocal tract.

And further down the road, Patel says, she’d like to look into ways to accommodate VocaliD’s recipients as they age. “If you have a recording of one person going through time, you’ll see that voice is changing,” she says. “Maybe it’s not that you have to get a brand-new donor and recipient. Maybe there’s a way to change it computationally … It would be an exciting thing, if we could build someone a voice when they’re a kid and grow it over time.”

* * *

The voice bank may be new, but the ability of machines to generate human speech predates even electricity.

Over a century before the modern computer would be developed, Hungarian inventor Wolfgang von Kempelen began his work on the first speech-synthesis machine in 1770. The final product, which would take him two decades to complete, used a bellows to simulate lungs, a reed to create vibrations, and rubber “mouth,” with tubes and levers that could be manipulated to create the sounds of vowels and consonants. According to his 1791 book The Mechanism of Human Speech, with a Description of a Speaking Machine, von Kempelen’s creation could imitate human speech well enough for people to recognize phrases in French and Italian.

In 1845, using a bellows design similar to von Kempelen’s, German scientist Joseph Faber unveiled his own talking machine, the “Euphonia,” at Philadelphia’s Musical Fund Hall. The machine, emblazoned with the image of a disembodied female head, had a “ghostly monotone,” historian David Lindsay wrote, but could speak every European language and sing “God Save the Queen.”

Both von Kempelen’s and Faber’s devices caught the attention of Alexander Graham Bell, who used their work as inspiration for his own model of the human vocal tract in 1860, 16 years before filing his patent for the telephone. Bell Labs, the company he later founded, was at the forefront of text-to-speech technology through its transition to the digital age: In 1961, the company was the first to synthesize speech with a computer, using an IBM machine to sing the song “Daisy Bell.” (Author Arthur C. Clarke, who happened to witness the demonstration, later recreated it with Hal, the computer in 2001: A Space Odyssey.)

Stephen Hawking’s voice—based on the “outdated technology” that Intel’s Haussecker referenced—comes from DECTalk, one of the first personal text-to-speech devices. Invented in the early 1980s by Dennis Klatt, an engineer at the Massachusetts Institute of Technology, the device originally had only three voices: Hawking’s, “Perfect Paul,” based on Klatt’s own voice; “Beautiful Betty,” based on his wife, and a child’s voice, which he named “Kit the Kid.” DECTalk has since added six additional voices (and dropped the adjectives—the newcomers are simply named “Harry,” “Ursula,” etc.), but Paul quickly became the standard in artificial voices—the voice was so common, in fact, that it was used by the National Weather Service until earlier this year.

“Nowadays, there are more choices than there were 10 years ago,” Patel says, but they remain limited. “[For example,] your GPS can speak in an Australian accent, American accent, male or female. Those are the kinds of choices people can make about their voice, but they’re not specific—there’s not a Bostonian speaking in a Bostonian accent.”

Paul’s ubiquity—and the small size of the pool of current options—throw into stark relief what the voice-impaired have lost. Like fingerprints, each human voice is unique to its owner; even the voices of identical twins have measurable differences.

“It’s really hard to overstate how important the voice is in the way we present ourselves to the world,” says Jody Kreiman, a speech scientist at the University of California Los Angeles’ Bureau of Glottal Affairs and the author of the book Voices and Listeners. “In the same way you look at someone and start drawing conclusions, you hear a voice and start drawing conclusions … Are they in a good mood or a bad mood? Are they healthy? Educational level, [whether] they’re good-looking or not, [whether] they’re a leader.”

“When you lose your voice,” Kreiman adds, “you lose your social self.”

* * *

“The real question is, how fast can we get people their voices?” Patel says. As of a few months ago, the waiting list was about a thousand names long. Each voice takes around 10-15 hours to build, once everything is recorded, but VocaliD has more to do before it can begin work in earnest—there are technological tweaks to be made, money to raise. To keep the eventual cost of a voice as low as possible, Patel is also looking into other ways to market her technology: “You might want [it] if you want an email to be read out loud in your voice,” she said, “or when you’re playing a video game and want to sound like yourself.”

In the meantime, VocaliD has thus far successfully created voices for three people, all teen girls, as part of its beta-testing phase. One of them, Samantha Grimaldo, is featured in a video on the company’s website, where the VocaliD team and Samantha’s mother watch her receive her new voice.

Seated at her family’s kitchen table, she types out a sentence on her tablet: “My favorite food is pizza.”

She’s grinning, though the inflectionless sound that emerges gives no indication of her excitement. Samantha’s new voice doesn’t sound completely natural. It sounds, still, like a robotic voice—but it doesn’t sound like anyone else, either.

This article available online at:


Social network’s hidden resources

wellesarmy-740x493People’s social net­works can be quite exten­sive, often bigger than they realize. So Brooke Fou­cault Welles, an assis­tant pro­fessor of com­mu­ni­ca­tion studies in the Col­lege of Arts, Media and Design, says it’s not sur­prising that past research indi­cates people can’t always recall everyone in their net­work and every­thing they know about them.

For her part, Fou­cault Welles is pur­suing a new line of research she describes as helping people “acti­vate their net­works.” She says people’s social groups, par­tic­u­larly in the work­place and other pro­fes­sional set­tings, con­tain valu­able con­nec­tions and resources that are under­uti­lized. For instance, a col­league could be a useful resource on a work project or someone in your pro­fes­sional net­work could be the ideal con­nec­tion for a new job. Often, the people who are most rel­e­vant to an individual’s needs are those at the edge of his or her net­work, she explained.

I like to think of these net­works as resources that are hidden in plain sight,” said Fou­cault Welles, whose research focuses on how social net­works shape and con­strain human com­mu­ni­ca­tion. “If you don’t have a good sense of who is in your net­work then you can’t leverage what people have to offer.”

Mea­suring and iden­ti­fying the con­se­quences of an individual’s ability to accu­rately acti­vate his or her net­works is a social psy­cho­log­ical con­struct that Fou­cault Welles calls “net­work thinking.” This approach, she says, can be par­tic­u­larly valu­able for the U.S. mil­i­tary, which relies on effi­cient and effec­tive networks.

This fall, Fou­cault Welles received a U.S. Army Research Lab­o­ra­tory young inves­ti­gator grant, with which she will spend the next three years mea­suring and iden­ti­fying “net­work thinking.”

Learning how well someone knows his or her net­work and detecting errors in the person’s rec­ol­lec­tions has tra­di­tion­ally been time-​​consuming and labor-​​intensive, she says. That’s why Fou­cault Welles, with her new grant, will develop a self-​​reporting scale for mea­suring “net­work thinking.”

Over the next year, she will survey North­eastern under­grad­u­ates with ques­tions about their social net­works and then com­pare those responses to what she per­ceives and observes from data gath­ered from their Face­book accounts.

For this project Fou­cault Welles has teamed up with Christo Wilson, an assis­tant pro­fessor in the Col­lege of Com­puter and Infor­ma­tion Sci­ence, who will develop a method­ology for col­lecting this Face­book data. Wilson’s research focuses on online social net­works, secu­rity and pri­vacy, and algo­rithmic society.

Once about 200 North­eastern stu­dents have been sur­veyed, Fou­cault Welles will use the scale to deter­mine how “net­work thinking” affects indi­vidual and team performance.

Fou­cault Welles said this research could have tremen­dous trans­la­tional poten­tial for mil­i­tary prac­tices. She also views the scale as a tool to mea­sure how quickly people adapt to new sit­u­a­tions, such as col­lege life. “Stu­dents who are quicker to rec­og­nize a col­lege sup­port net­work are more likely to have an easier tran­si­tion,” she said.

Col­lab­o­rating with Wilson will also create an oppor­tu­nity to set the social sci­ences stan­dards for col­lecting data from social net­works such as Face­book, Fou­cault Welles said. “Right now there are few eth­ical guide­lines for col­lecting data from Face­book,” she said. “We want to estab­lish a track record of researchers doing this eth­i­cally to gen­erate social sci­en­tific insights.”

Ebert, S’15 Wins Marshall Scholarship


Julia Ebert, S’15, has won a Mar­shall Schol­ar­ship to pursue a one-​​year master’s of research in bio­engi­neering at Impe­rial Col­lege London starting in the fall.

Founded by the British gov­ern­ment in 1953 to com­mem­o­rate the Mar­shall Plan, the post­grad­uate schol­ar­ship allows up to 40 intel­lec­tu­ally dis­tin­guished Amer­ican stu­dents to study in the United Kingdom each year.

Ebert is Northeastern’s second stu­dent to receive the award, whose 2015 win­ners were announced last week.

“It’s an honor to receive the Mar­shall Schol­ar­ship and it shows that all the work

I have done up to this point has paid off,” said Ebert.

Ebert is a fifth-​​year behav­ioral neu­ro­science major who applied for the schol­ar­ship through the University Scholars, which houses the university’s Office of Fel­low­ships.

Her aca­d­emic journey began in high school, when her pas­sion for learning led her to pursue the Inter­na­tional Baccalaureate’s psy­chology course. “I got really inter­ested in psy­chology,” she said, “and I wanted to take a more sci­en­tific approach to under­standing the brain.”

Over the past four years, Ebert has fine-​​tuned her aca­d­emic focus through research posi­tions and co-​​op jobs in campus labs and far-​​flung coun­tries. In the fall of 2011, the honors stu­dent and National Merit Scholar started working as a research assis­tant in Northeastern’s Action Lab, which is ded­i­cated to the exper­i­mental and com­pu­ta­tional study of human motor con­trol. Under the direc­tion of pro­fessor Dagmar Sternad, she col­lected and ana­lyzed data from human par­tic­i­pants in motor con­trol exper­i­ments and sub­se­quently won a Barry Gold­water Schol­ar­ship for her research achieve­ments. She is cur­rently fin­ishing her under­grad­uate thesis on learning and long-​​term reten­tion of a bimanual skill.

Julia has the mak­ings of an excel­lent sci­en­tist,” said Sternad, a pro­fessor of physics, biology, and elec­trical and com­puter engi­neering. “She is extremely bright, works inde­pen­dently, and is self-​​motivated.”

Ebert’s next expe­ri­en­tial learning opportunity—a research co-​​op in the autonomous motion depart­ment at the Max-​​Planck Insti­tute for Intel­li­gent Sys­tems in Tübingen, Ger­manykin­dled her interest in robotics and machine learning. There, she designed exper­i­ments using the Cyber­Glove, an input device that mea­sures 19 joint angles in the hand and can be used to test how humans learn to con­trol a high-​​dimensional system. “Con­ducting research has given me insight into how my course work fits together,” said Ebert, who added a com­puter sci­ence minor fol­lowing her ini­tial expe­ri­ence in the Action Lab. “I started taking more math classes and learned how to pro­gram, which has enabled me to do more tech­nical modeling.”

As a Mar­shall Scholar, Ebert hopes to work in Dr. Eti­enne Burdet’s human robotics lab, which aims to design assis­tive devices and vir­tual reality-​​based training for reha­bil­i­ta­tion and surgery. After London, she plans to return to the U.S. to earn her doc­torate in bio­med­ical engi­neering with the goal of becoming an aca­d­emic researcher.

Ebert summed up her career ambi­tions in her per­sonal state­ment for the Mar­shall Schol­ar­ship appli­ca­tion, writing that “My desire to learn has already led me to a field that draws on my inter­ests in every­thing from music to math. Now I want to expand on my pas­sion for neu­ro­science not only to solve the mys­tery of learning, but to employ that infor­ma­tion to improve lives.”

To Halt Ebola’s Spread, Researchers Race for Data

At a burial in Freetown, Sierra Leone, on Nov. 19, a member of the burial team struggles to regain his footing. Photo by Nikki Kahn/The Washington Post

The Ebola virus has consistently stayed several steps ahead of doctors, public officials and others trying to fight the epidemic. Throughout the first half of 2014, it spread quickly as international and even local leaders failed to recognize the severity of the situation. In recent weeks, with international response in high gear, the virus has thrown more curve balls.

The spread has significantly slowed in Liberia and beds for Ebola patients are empty even as the U.S. is building multiple treatment centers there. Meanwhile the epidemic has escalated greatly in Sierra Leone, which has a serious dearth of treatment centers. And in Mali, where an incursion was successfully contained in October, a rash of new cases has spread from an infected imam.

Predicting the trajectory of Ebola rather than playing catching-up could do much to help prevent and contain the disease. Some experts have called for prioritizing mobile treatment units that can be quickly relocated to the spots most needed. Figuring out where Ebola is likely to strike next or finding emerging hot spots early on would be key to the placement of these treatment centers.

But such modeling requires data, and lots of it.  And for stressed healthcare workers on the ground and government and non-profit agencies scrambling to combat a raging epidemic, collecting and disseminating data is often not a high priority.

Air traffic connections from West African countries to the rest of the world. Guinea, Liberia, and Sierra Leone are not well connected outside the region; Nigeria, in contrast, is. Image source

Population Flows

The crux to combating Ebola is understanding how people move between different cities, villages and countries. Such data are already captured in a variety of metrics. On the macro level, records of border crossings and airline flights create clear pictures. On a more local level, trucking and bus routes and traffic flows help. But especially in rural areas like the forests of Guinea where the epidemic started, even more detailed information is needed.

Alessandro Vespignani is one researcher trying to gather that information. Vespignani seeks population data at the most granular level possible, trying to determine numbers of people and types of dwellings within five by five mile boxes, for example. He uses local census numbers plus data from the LandScan program out of Oak Ridge National Laboratory and Worldpop, a UK-based project to map populations in Africa, Asia and Latin America with a focus on development and health. He integrates that information with Ebola data provided by health agencies, and with data on the movement of people from airlines, borders, transportation records and other sources. In early September he published a high-profile projection of the potential international spread of Ebola, using these data sources.

Mapping with Mobile Phones

Mobile phone records are another promising way to track the movements of people at a more localized level. Phone data stripped of users’ identities helped researchers understand past epidemics including the cholera outbreak following the 2010 Haiti earthquake.

Researchers involved in the Swedish non-profit organization Flowminder have been trying to use mobile phone records to shed light on Ebola’s spread. However they’ve been so far stymied by a lack of cooperation from phone companies and from government regulators in West Africa, who have not made the data available. To demonstrate the potential of its approach, Flowminder created a model showing people’s movements in Senegal and Ivory Coast using several-year-old data from Orange Telecom. Flowminder board member Andy Tatem, a Reader at the University of Southampton, says that negotiations are ongoing with government regulators and companies which could provide the mobile phone data, but it has been slow going.

“This kind of data can give you information about population level movements, how they change over time, how they change over space,” says Tatem, who is also director of Worldpop and an expert in modeling population movements related to malaria. (There is no talk of using phone data to track individual people infected by Ebola. That would be considered a major breach of privacy, and would also likely be impossible given the high number of cases.)

“It’s an area of the world where there are huge seasonal movements, mobility patterns changing month by month. People cross borders between countries, people are moving to the cities looking for alternative work.”

And unlike conventional disease surveillance, Tatem notes, once companies and regulators give the green light, using mobile phone data is basically “free” and involves no continuous action from companies and regulators.

Data Gaps

However even mobile phone data isn’t a perfect relay of what’s going on on the ground. A dense city could have numerous phone towers, allowing fairly precise modeling of human flows, even down to specific neighborhoods. But in rural areas, one tower might cover a radius of 50 miles or more, making it less useful in estimating movements between small villages.

And in the impoverished and often geographically isolated areas decimated by Ebola, many people don’t have mobile phones. As a whole, mobile phone usage in Africa is high and growing. But Guinea, Sierra Leone and Liberia have among the lowest mobile phone usage rates in Sub-Saharan Africa, with between 51 and 54 percent of households having phones, compared to 78 percent for Nigeria and 96 percent for Mauritania, according to a Gallup poll this year. In the U.S., nine in 10 adults have a mobile phone.

Capturing Clinical Data

Using data for a dynamic understanding of the epidemic also naturally involves information compiled by health care workers on the ground, including counts of cases, deaths and people in isolation. Record-keeping during the epidemic has been notoriously flawed and incomplete, and getting records from far-flung clinics that may not have computers is a daunting undertaking.

“We are talking about very weak health care systems in the region that were right away overwhelmed by the situation,” says Vespignani. “In that case you cannot ask people who are really struggling to save lives, to get data.”

But Vespignani is hopeful that as the epidemic ebbs, as it has at least in Liberia, staff and officials will be able to focus more on compiling and providing data from Ebola treatment centers.

For instance, researchers could compare predicted caseloads with actual caseloads in specific areas where initiatives around safe burials, prevention, isolation or other best practices have been implemented. That could reveal whether such efforts have significantly reduced infections compared to what otherwise would have been expected. Another aspect not currently captured in models, he says, is “social behavior” like  community perception of the disease and rate of compliance with government and agency warnings.

And such data could be key to making sure the epidemic is really stamped out, and doesn’t resurface catching people unaware.

“When I hear people saying the epidemic is subsiding I always shiver,” he says. “We see improvement, there’s a slowing down and things are improving in certain places, but we need to have the last battle to really try to contain it. Any decrease of effort, any arrogance to say we are good, could really backfire and we could find ourselves with a disaster.”

Article from Discover Magazine