AT&T Labs Research — Leading Invention, Driving Innovation

Katie Kuksenok
AT&T Labs Fellowship Award Winner

An interactive translation system

"Machine translation can carry literal meanings across language barriers faster than any human. However, current approaches struggle with the textual context or interpreting cultural cues, something people do naturally. I hope to leverage the insight gained through human collaboration in investigating new approaches to machine translation."  Read more.

- Katie Kuksenok

Edward McFowland III
AT&T Labs Fellowship Award Winner

Using anomaly detection to ensure data integrity

"Bad data leads to sub-optimal decisions. This problem is especially acute in the third world, which lacks digitization, expertise, and often Internet access. And where it’s imperative that resources, which are very limited, be spent wisely.”  Read more.

- Edward McFowland III

Sean Sanders
AT&T Labs Fellowship Award Winner

Tracking the latest killer app

"Understanding the behavior of your data is essential to accurately classify the trends, patterns, and anomalies that tell you what is happening within the data. Massive datasets require more care in the design or choice of tools and algorithms and more efficient data management."  Read more.

- Sean Sanders

Using cellular data for urban planning

Continuous streams of data from ubiquitous mobile devices provide a way to monitor city dynamics frequently, cheaply, and at an unprecedented scale.

In an article published in the current edition of Pervasive Computing, AT&T researchers show how anonymized and aggregated cellular data can be used to analyze commuting patterns in a New Jersey suburban city. Article requires payment.

Paper on mining twitter data named Best Paper at SWSM

Extracting Descriptions of Problems with Product and Services from Twitter Data by AT&T Researcher Narendra Gupta was named Best Paper at this year’s Social Web Search and Mining Workshop.

The paper describes a new way to mine Twitter data to find customer-complaint tweets related to a company’s products or services. Recently the technique helped generate more accurate alarms, sometimes before a ticket was opened.

FCC recognizes Speech Mashup with two awards

For making it easier for developers to create assistive technologies, the FCC gave two awards to AT&T’s Speech Mashup, which puts speech technologies in the cloud. Jay Wilpon accepted the awards on behalf of AT&T Research. To read more about the Speech Mashup and accessibility, go here.

See the Assistive Technology Project for other AT&T research on accessibility.

Computers and Intractability is most-cited CS book

According to Microsoft Academic Search (currently in beta), a resource for academic and technical publications, Computers and Intractability: A Guide to the Theory of NP-Completeness by Michael Garey and AT&T Labs executive director David Johnson is the most cited computer science book. Published in 1979, this book was the first on the theory of NP-completeness and computational intractability and is now considered a classic.

STEM may be key to upward mobility for Latinos

Median wealth for Hispanic households fell 66% between 2005 and 2009. Alicia Abella, interviewed here by New America Media, sees this alarming statistic as a call for further efforts to promote STEM education among Hispanics and other under-represented minorities. Technology jobs pay well, and their number is growing faster than supply. By 2018, the computer field alone will create 800,000 new jobs, far more than the 24,000 who graduated in this sector in the past three years.

Latest speech translation technology enables Spectra

Spectra is a free iPhone and iPad app that accurately translates in close to real time, allowing for more natural, conversation-like interactions, whether in English, Spanish, Japanese, Chinese, French, German, or Italian.

The speech translation technology behind Spectra, as featured in this ad, was created at AT&T Research. Spectra also incorporates AT&T WATSON^TM and AT&T Natural Voices^®.