The Siebel Scholars community consists of more than 870 of the world's brightest minds in business, computer science, and bioengineering. The Siebel Scholars community actively fosters leadership, collaboration, and increased potential for Siebel Scholars to achieve even more through their work with an incomparable group of equally talented peers.

Mrinmaya Sachan and Anuj Kumar were selected as a Siebel Scholars based upon their academic excellence and demonstrated leadership. The deans of their colleges selected five students from those who rank at the very top of their class to be recognized as Siebel Scholars. They will each receive a $35,000 award for their final year of graduate studies. The merit-based Siebel Scholars award is granted for outstanding contributions and achievements, without regard to financial need.

The Siebel Scholars program was founded in 2000 to recognize the most talented students at the world's leading graduate schools of business, computer science, and bioengineering and to form an active, lifelong community among an ever-growing group of leaders.

Each year, 85 graduate students at the top of their class are selected from some of the most prestigious universities to be honored as Siebel Scholars. These universities include: Carnegie Mellon University; Harvard University; Johns Hopkins University; Massachusetts Institute of Technology; Northwestern University; Princeton University; Stanford University; Tsinghua University; University of California, Berkeley; University of California, San Diego; University of Chicago; University of Illinois at Urbana-Champaign; and University of Pennsylvania.

It is with great pleasure that we welcome Mrinmaya Sachan and Anuj Kumar to the Siebel Scholars Class of 2014.

Carnegie Mellon Method Uses Network of Cameras To Track People in Complex Indoor Setting

Real-Life "Marauder's Map" Has Applications in Health Care, Security

PITTSBURGH—Researchers at Carnegie Mellon University have developed a method for tracking the locations of multiple individuals in complex, indoor settings using a network of video cameras, creating something similar to the fictional Marauder's Map used by Harry Potter to track comings and goings at the Hogwarts School.

The method was able to automatically follow the movements of 13 people within a nursing home, even though individuals sometimes slipped out of view of the cameras. None of Potter's magic was needed to track them for prolonged periods; rather, the researchers made use of multiple cues from the video feed: apparel color, person detection, trajectory and, perhaps most significantly, facial recognition.

Multi-camera, multi-object tracking has been an active field of research for a decade, but automated techniques have only focused on well-controlled lab environments. The Carnegie Mellon team, by contrast, proved their technique with actual residents and employees in a nursing facility — with camera views compromised by long hallways, doorways, people mingling in the hallways, variations in lighting and too few cameras to provide comprehensive, overlapping views.

The performance of the Carnegie Mellon algorithm significantly improved on two of the leading algorithms in multi-camera, multi-object tracking. It located individuals within one meter of their actual position 88 percent of the time, compared with 35 percent and 56 percent for the other algorithms.

The researchers— Alexander Hauptmann, principal systems scientist in the Computer Science Department (CSD); Shoou-I Yu, a Ph.D. student in the Language Technologies Institute; and Yi Yang, a CSD post-doctoral researcher — will present their findings June 27 at the Computer Vision and Pattern Recognition Conference in Portland, Ore.

Though Harry Potter could activate the Marauder's Map only by first solemnly swearing "I am up to no good," the Carnegie Mellon researchers developed their tracking technique as part of an effort to monitor the health of nursing home residents. "The goal is not to be Big Brother, but to alert the caregivers of subtle changes in activity levels or behaviors that indicate a change of health status," Hauptmann said. All of the people in this study consented to being tracked.

These automated tracking techniques also would be useful in airports, public facilities and other areas where security is a concern. Despite the importance of cameras in identifying perpetrators following this spring's Boston Marathon bombing and the 2005 London bombings, much of the video analysis necessary for tracking people continues to be done manually, Hauptmann noted.

The CMU work on monitoring nursing home residents began in 2005 as part of a National Institutes of Health-sponsored project called CareMedia, which is now associated with the Quality of Life Technology Center, a National Science Foundation engineering research center at CMU and the University of Pittsburgh.

"We thought it would be easy," Hauptmann said of multi-camera tracking, "but it turned out to be incredibly challenging."

Something as simple as tracking based on color of clothing proved difficult, for instance, because the same color apparel can appear different to cameras in different locations, depending on variations in lighting. Likewise, a camera's view of an individual can often be blocked by other people passing in hallways, by furniture and when an individual enters a room or other area not covered by cameras, so individuals must be regularly re-identified by the system.

Face detection helps immensely in re-identifying individuals on different cameras. But Yang noted that faces can be recognized in less than 10 percent of the video frames. So the researchers developed mathematical models that enabled them to combine information, such as appearance, facial recognition and motion trajectories.

Using all of the information is key to the tracking process, but Yu said facial recognition proved to be the greatest help. When the researchers removed facial recognition information from the mix, their on-track performance in the nursing home data dropped from 88 percent to 58 percent, not much better than one of the existing tracking algorithms.

The nursing home video analyzed by the researchers was recorded in 2005 using 15 cameras; the recordings are just more than six minutes long.

Further work will be necessary to extend the technique during longer periods of time and enable real-time monitoring. The researchers also are looking at additional ways to use video to monitor resident activity while preserving privacy, such as by only recording the outlines of people together with distance information from depth cameras similar to the Microsoft Kinect.

The National Science Foundation sponsored this research.

Original Article: http://news.cs.cmu.edu/article.php?a=3757

See also:
http://news.yahoo.com/harry-potter-meets-high-tech-surveillance-tracking-system-154700763.html
http://www.ecnmag.com/news/2013/06/carnegie-mellon-method-uses-network-cameras-track-people-complex-indoor-settings-0
http://www.wired.co.uk/news/archive/2013-05/31/real-life-marauders-map
http://www.newscientist.com/article/mg21829196.000-smart-map-tracks-people-through-camera-networks.html

Bhavana Dalvi Mishra, a Ph.D. student advised by William Cohen and Jamie Callan, was recently awarded the 2013 Google U.S./Canada Fellowship in Information Extraction.

This fellowship was one of just 15 named PhD fellowships that Google awarded in U.S. and Canada. Google's announcement of Bhavana's award said "The student nominations we received were exemplary in their quality, but Bhavana especially stood out and was endorsed by the research scientists and distinguished engineers within Google who participated in the review."

Bhavana's research develops new methods of extracting information from text to create and populate taxonomies and knowledge bases. She has developed unsupervised algorithms that locate tables in large web datasets, and use the content of the tables, as well as other text, to identify new concepts (e.g., "football teams", "musical notes") and instances of those concepts (e.g., "Pittsburgh Steelers", "B flat"). Recently, Bhavana developed a new method of using partially-labeled training data to guide the discovery of known and new types of concepts in a large-scale dataset.

The fellowship begins in August. Google Ph.D. fellowships cover tuition, fees, and a stipend for two years of academic study and research.

Derry Wijaya has been selected as a recipient of the Google Anita Borg Memorial Scholarship for $10,000. Sunaya Sitaram and Lavanya Viswanathan were selected as Finalists of the Scholarship this year and were awarded $1000.

"Dr. Anita Borg devoted her adult life to revolutionizing the way we think about technology and dismantling barriers that keep women and minorities from entering computing and technology fields. Her combination of technical expertise and fearless vision continues to inspire and motivate countless women to become active participants and leaders in creating technology.

In her honor, Google is proud to honor Anita’s memory and support women in technology with the Google Anita Borg Memorial Scholarship. Google hopes to encourage women to excel in computing and technology and become active role models and leaders in the field.

Google Anita Borg Scholarship recipients will each receive a financial award for the academic year. A group of female undergraduate and graduate students will be chosen from the applicant pool, and scholarships will be awarded based on the strength of each candidate’s academic background and demonstrated leadership.

In addition, all scholarship recipients and finalists will be invited to attend a retreat at Google. We know how important a supportive peer network can be for a student’s success. The retreat will include workshops, speakers, panelists, breakout sessions and social activities scheduled over a couple of days."

http://www.google.com/anitaborg/

Silly Phone Game Puts Illiterate Pakistanis In Touch With Potential Employers

Carnegie Mellon and Lahore University Research Project Goes Viral

PITTSBURGH—A silly telephone game that became a viral phenomenon in Pakistan has demonstrated some serious potential for teaching poorly educated people about automated voice services and provided a new tool for them to learn about jobs, say researchers at Carnegie Mellon University and Pakistan's Lahore University of Management Sciences (LUMS).

The game, called Polly, is simplicity itself: a caller records a message and Polly adds funny sound effects, such as changing a male's voice to a female voice (or vice versa), or making the caller sound like a drunk chipmunk. The caller can then forward the message to one or more friends, who in turn can forward it along or reply to it.

Polly may not sound like a research project, but Roni Rosenfeld, professor in Carnegie Mellon's Language Technologies Institute, said it is pioneering the use of entertainment to reach illiterate and low-literate people and introduce them to the potential of telephone-based services. Such phone services could help non-affluent, poorly educated people find jobs, find or sell merchandise, become politically active, create speech-based mailing lists and even support citizen journalism.

But people can't use these services if they don't know how.

Even though most people in Pakistan have access to a phone, many don't understand the technology behind an automated telephone-based service, said Agha Ali Raza, a Ph.D. student in language technology and a native Pakistani. "They expect to talk to a person on the other end of the line," he explained. "When they hear, 'Press 1 to do this,' or 'Press 2 to do that,' they don't press anything; they just start talking."

With Polly, Rosenfeld, Raza and Umar Saif, an associate professor of computer science at LUMS, have shown that if the training is fun, people will not only learn how to use phone-based services, but will eagerly spread the word and even show each other how to use it. Polly was launched in Lahore, Pakistan in May 2012 by giving its phone number to five poor, low-skilled workers. By mid-September, 85,000 people had used it almost half a million times.

Though budget pressures forced researchers to begin limiting calls to Polly in September, the total number of users climbed to more than 160,000 people, including some non-Pakistanis, as of mid-April. Overall, the system has handled almost 2.5 million calls. The project continues to run.

What's more, Polly doesn't just deliver funny messages; it also includes job listings. "We daily scan Pakistani newspapers for advertisements for jobs that are appropriate for low-skilled, low-literate workers, record them in the local language and make them available for audio-browsing during the interaction with Polly," Rosenfeld said. As of mid-April, the ads had been listened to more than 380,000 times and had been forwarded more than 21,000 times.

Raza, the lead author, will present results of the research on May 1 at CHI 2013, the Conference on Human Factors in Computing Systems, in Paris, where the report has received a Best Paper award.

Rosenfeld said the entertainment value of Polly helped it spread rapidly, but can't sustain it over time, noting game play dropped rapidly as its novelty wore off. But adding services, such as the job ads, can keep people calling in.

"We found that users took to the job information in large numbers and that many of them started calling Polly specifically for that service — exactly the result we had hoped for," he said.

The researchers now are determining whether the system can be scaled up to serve a larger population base over an extended period of time and be made more cost efficient. They also hope to use Polly to train people in other countries.

More information, as well as video demonstrations, is available on the project website, http://www.cs.cmu.edu/~Polly/.

The project is sponsored in part by the U.S. Agency for International Development under the Pakistan-U.S. Science and Technology Cooperation Program, the U.S. State Department's Fulbright Program and the Higher Education Commission of Pakistan.

The Language Technologies Institute is part of CMU's School of Computer Science.

Selen Uguroglu, currently a final-year Ph.D. in the LTI, has been awarded the Apple iOS Scholarship for Women in Technology. The scholarship is work $10,000 and provides an internship opportunity.

Andre Martins (Noah Smith + Eric Xing [CMU], and Mario Figueiredo + Pedro Aguiar [IST, Portugal]) received an SCS Dissertation Honorable Mention for "The Geometry of Constrained Structured Prediction: Applications to Inference and Learning of Natural Language Syntax"

People have learned to preserve their privacy by safeguarding computer passwords. But with the rise of voice authentication systems, they also need to protect unique voice characteristics.

Researchers at Carnegie Mellon University's Language Technologies Institute (LTI) say that it is possible with a system they developed. The CMU system converts a user's voiceprint into alphanumeric strings that can serve as passwords.

This system would enable people to register or check in on a voice authentication system — without their actual voice ever leaving their smartphone. This reduces the risk that a fraudster will obtain the person's voice biometric data, which could be used to access bank, healthcare or other personal accounts.

"When you use a speaker authentication system, you're placing a lot of faith in the system," said Bhiksha Raj, CMU associate professor of language technologies. "It's not just that your voiceprint might be stolen from the system and used to impersonate you elsewhere. Your voice also carries a lot of information — your gender, your emotional state, your nationality."

"To preserve privacy, we need systems that can identify you without actually hearing your voice or even keeping an encrypted record of your voice."

Raj and Manas Pathak (CS'09/CS'12), a recent Ph.D. graduate of LTI, have devised a method for converting a voiceprint — a spectrogram that represents the acoustic qualities of speech. On Sept. 21 they will present their work as a keynote address at the Information Security Conference in Passau, Germany.

Because a person's voice never sends the same signal twice, even when repeating the same word or phrase, converting the voiceprint into a single password won't do.

So, the CMU system uses different mathematical functions to generate hundreds of alphanumeric strings. To authenticate the user, the system compares all of the strings with those that the system has on file from the initial registration; if enough of the strings match, the user is authenticated.

The system also adds what the researchers call "salt" — a random string of digits unique to each smartphone — to the alphanumeric strings, providing an additional level of security.

In tests using standardized speech datasets, Raj and Pathak found that their system was accurate 95 percent of the time. The privacy-preserving method is computationally efficient, so it could be used with most smartphones.

But Raj also warned that improving the security of voice authentication systems would be just a first step to protecting privacy overall.

"With increasing use of speech-based services, such as the iPhone's Siri assistant or the personal videos uploaded to YouTube, the issue of the privacy of users' speech data is only just beginning to be considered," he said.

In addition to Raj and Pathak, Jose Portelo and Isabel Trancoso of INESC-ID in Lisbon, Portugal, contributed to this research. This work was supported by the National Science Foundation and Portugal's Foundation for Science and Technology (FCT).

Economist Article
MIT Tech Review

Among the returning Carnegie Mellon University students this week is Elizabeth Davis (CS'14), whose interests in math, science and the French language led her to pursuing a major in computer science with a language technologies minor.

Working side-by-side with faculty member Maxine Eskenazi (DC'73), she created a smartphone app for non-native speakers and will test the app this fall.

"Computer science is relevant in so many fields," said Davis. "And I like that CMU emphasizes the kind of interdisciplinary thinking that brings together interesting ideas from absolutely everywhere."

Choosing CMU was a no-brainer when she learned she wouldn't have to give up any of her interests — which include playing violin and Scottish fiddle. She's even going to France next semester for further study in language technologies.

"It's so interesting to see all the things I like come together like this for me," said Davis, who keeps the rust off her music abilities by playing with CMU's All University Orchestra for non-music majors.

The app she created is designed for non-native speakers to learn which words could cause miscommunication with an audience.

"Say you're going to give a presentation. You load the text of your presentation into the app. Based on your native language it will tell you the words your audience may misunderstand if you don't pronounce them correctly," she explained. "If you know ahead of time that your pronunciation of 'crowds' could sound like 'clouds,' and it would make just as much sense to hear 'clouds' in that context, then you have the opportunity to practice pronouncing that word. Or maybe you'll decide to use a different word in its place."

Eskenazi added, "So, any student going to a conference can use the app to prepare. And because the technology is in the form of an app right now, we are looking forward to getting a lot of people to test it so we can see how they react to using the app for learning and what precision issues there might be."

She and Davis are already planning more uses for the technology.

"The next thing we need to add is speech recognition so that the app can respond as a student practices pronouncing a word," Eskenazi said. "Then perhaps we can add multiple word phrases or expressions. For example, these are the expressions I use. How might people misunderstand what I'm saying?"

Eskenazi says the breadth and expertise in language technologies at CMU has been critical to the success of this project. And she noted that a Summer Undergraduate Research Fellowship (SURF), funded by the Semiconductor Research Corporation (SRC), made the project possible.

"SURGs and SURFs make it possible for students to pursue ideas that have never been researched before," Eskenazi said.

CMU is one of 14 campuses to participate in the SRC Undergraduate Research Opportunities (SRC-URO) program. It is designed to attract and retain students in science and engineering disciplines relevant to technology-based industries and to encourage their enrollment in graduate school by providing sustained undergraduate research experiences and mentoring relationships.

Davis, like many other undergraduates at CMU, enjoys being able to work on something novel.

"I thought what can I lose by trying this? And I got to learn all about what research is like a lot earlier than most people," Davis said. "Research is fun."

Eskenazi added, "Students get a feel for what academia is all about and a better idea of whether they would enjoy graduate school. It also gives them a way to apply what they are learning. By using what you learn, you get deeper knowledge and you have a better chance of being able to use it later in life."

Original Article: http://www.cmu.edu/homepage/society/2012/summer/welcome-back-students.shtml

Curtis Huttonhower, Harvard University, was recently named as a recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE), the highest honor bestowed by the United States government on science and engineering professionals in the early stages of their independent research careers.

Huttonhower, who provided a framework for data mining genome databases, was one of 96 PECASE recipients announced by the White House and was one of 20 recipients nominated by the National Science Foundation. The PECASE program recognizes scientists and engineers who, early in their careers, show exceptional potential for leadership at the frontiers of knowledge.

Add'l CCCBlog Article

Amr Ahmed's dissertation "Modeling Content and Users: Structured Probabilistic Representation and Scalable Inference Algorithms" has been selected as the winner of the 2012 ACM SIGKDD Doctoral Dissertation Award. The award, which constitutes a $2,500 honorarium and a plaque, will be presented during the opening ceremonies at the upcoming KDD Conference in Beijing.

Xu, Y., & Mostow, J. (2012, June 19-21). Comparison of methods to trace multiple subskills: Is LR-DBN best? [Best Student Paper Award]. In Proceedings of the Fifth International Conference on Educational Data Mining, Chania, Crete, Greece. Click here for .pdf file.

Abstract: A long-standing challenge for knowledge tracing is how to update estimates of multiple subskills that underlie a single observable step. We characterize approaches to this problem by how they model knowledge tracing, fit its parameters, predict performance, and update subskill estimates. Previous methods allocated blame or credit among subskills in various ways based on strong assumptions about their relation to observed performance. LR-DBN relaxes these assumptions by using logistic regression in a Dynamic Bayes Net. LR-DBN significantly outperforms previous methods on data sets from reading and algebra tutors in terms of predictive accuracy on unseen data, cutting the error rate by half. An ablation experiment shows that using logistic regression to predict performance helps, but that using it to jointly estimate subskills explains most of this dramatic improvement. An implementation of LR-DBN is now publicly available in the BNT-SM student modeling toolkit.

Andre Martins is the proud recipient of IBM Portugal's Prémio Científico for his dissertation "Turbo Parsers: Dependency Parsing by Approximate Variational Inference"

Sitaram, S., & Mostow, J. (2012, May 23-25). Mining Data from Project LISTEN’s Reading Tutor to Analyze Development of Children's Oral Reading Prosody [Best Paper Award]. In Proceedings of the 25th Florida Artificial Intelligence Research Society Conference (FLAIRS-25), 478-483. Marco Island, Florida. Click here for .pdf file.

Abstract: Reading tutors can provide an unprecedented opportunity to collect and analyze large amounts of data for understanding how students learn. We trained models of oral reading prosody (pitch, intensity, and duration) on a corpus of narrations of 4558 sentences by 11 fluent adults. We used these models to evaluate the oral reading prosody of 85,209 sentences read by 55 children (mostly) 7-10 years old who used Project LISTEN's Reading Tutor during the 2005-2006 school year. We mined the resulting data to pinpoint the specific common syntactic and lexical features of text that children scored best and worst on. These features predict their fluency and comprehension test scores and gains better than previous models. Focusing on these features may help human or automated tutors improve children’s fluency and comprehension more effectively.

Jaime Carbonell and two other University professors (Nadine Aubry and Chris Hendrickson) have been named University Professors, the highest academic accolade a faculty member of CMU can acheive. Read More...

Naoki Orii and Madhavi Ganapathiraju's paper "Wiki-Pi: a web-resource to aid in the discovery of gene function via protein-protein interactions" is in press for PLoS ONE, 2012

Siddharth Gopal and Yubin Kim have each received Peter Jackson Fellowships, which support cutting-edge research in Information Retrieval. Siddharth's fellowship for Spring 2012 supports his research on hierarchical modeling of data using supervised and unsupervised techniques. Yubin's fellowship for Fall 2012 will support her research on text analysis of social media.

The Peter Jackson Fellowships were established by Thomson Reuters Corporation and its employees to honor the memory of the late Dr. Peter Jackson, Chief Scientist and head of R&D of the Thomson Reuters Corporation. Peter was a well-known and highly respected researcher in information retrieval, natural language processing, and artificial intelligence, and an author of three books and 40 papers. He also was an avid musician, a charming dinner companion, and a friend to university research.

Each Peter Jackson Fellowship contributes $17,625 to help cover the costs of a student's graduate education and research. The Language Technologies Insitute thanks Thomson Reuters and its employees for these generous gifts that honor Peter's memory by supporting young researchers.

For more information about Peter and his many accomplishments, please see his website, and several rememberances ([1],[2],[3],[4]) written by his friends and colleagues.

Tarun Sharma, Madhavi Ganapathiraju and Mohamed Thahir's paper "An efficient heuristic method for active feature acquisition and its application to protein-protein interaction prediction" was published in BMC Proceedings, Vol 6, Issue Supplement 7, Article S2, 2012.

Manas Pathak has been awarded a Springer Outstanding Thesis award in the Springer "Best of the Best" series. Manas will receive a 500 Euro prize and will have an appearance of his thesis on "Privacy-Preserving Machine Learning for Speech Processing" in the Springer thesis series.

The National Science Foundation has awarded a $500,000 Phase II Small Business Innovation Research grant to Safaba Translation Solutions LLC, a company spun out of the Language Technologies Institute.

The award was presented for: Wang, Y. C., Kraut, R. E., & Levine, J. M. (2012). To Stay or Leave? The Relationship of Emotional and Informational Support to Commitment in Online Health Support Groups CSCW'2012: Proceedings of the ACM Conference on Computer Supported Cooperative Work (pp. 833-842). NY: ACM.