Our research covers all aspects of data science and analytics, statistical
theory, methodology and computing as well as machine learning, including
a rich variety of collaborations with other research disciplines.
We partner closely with our Labs colleagues and experts in computer
science, mathemetics, databases, networking, and software and
Our research cuts across the entire spectrum of statistical application
and theory: time series, spatial statistics, Bayesian analysis,
machine learning, artificial intelligence, queueing theory, and applied
probability. Our research projects are typically motivated by the
massive data sets generated from AT&T's products - over 100M mobile
devices, millions of IPTV customers, the largest Wi-Fi network in
the US, and the network backbone that connects the entire internet.
Current examples of domains where our research is applied include:
Recommender systems are models which predict items
that a customer may like, based on previous purchase or rating behavior. We have been
doing research in these collaborative filtering models to predict movie and television
preferences. This work resulted in our sharing the $1M Grand Prize. Here is a
video describing our approach.
|Customer Experience Modeling.
We use statistical models/classifiers, text analysis, custom survey research, and GIS
data mining to glean insights about our customers' experience with our services, and to
determine drivers of dissatisfaction and churn. Together with internal business partners,
we use the results to design targeted interventions and set company targets.
- Interactive Data Analysis and Statistical Computing in the Cloud with R.
Our rich history
in statistical computing continues to this day with research into the statistical computing tools of the future.
Recent focus is on interactive data analysis tools,
and RCloud, a web-based cloud-computing approach to statistics.
We are a Supporting Institution of the R Foundation and have a
leadership role in the R community through the work of Simon Urbanek.
We are studying data
from mobile phone networks to understand how people live, work,
and move. Specifically, mobile data has the potential to
revolutionize the study of city dynamics -- understanding how
people use resources in a city and allowing us to help design
more sustainable and livable cities. Our work was described in
this cover article in the Communications of the ACM.
Our communications network defines
a rich social network in both the consumer and business spaces.
Our research focuses on extracting information through these
networks by looking at information flow through the network or
viral marketing opportunities. For an example, see our paper on
network based marketing.
Fraud Detection in Telecommunications.
AT&T is the industry leader in protecting our networks from fraud. Fraud occurs in many forms:
hackers breaking into lines, subscription fraud, spoofed cell phones, etc. Each of these requires
large-scale data mining to determine the proper model fro detecting the fraud while minimizing false
positives. Read a paper about our
history that won the Jack Youden Prize.
As one of the largest advertisers on the web, as well as the owner of advertising channels like U-Verse
and yellowpages.com, AT&T is always looking for ways to optimize ad placement. Our efforts here include machine learning
algorithms to estimate ad click through rates and spatial models to aid in local search. See our
paper at ICDM.
Data Quality and Anomaly Detection in Data Streams.
Monitoring vast data streams is a constant problem for large corporations. We model the incoming data as a multivariate data
stream and build models to cluster and alert on those time series. Modeling data quality of the stream draws on expertise in
Statistics, AI, Software engineering and Database Research to develop efficient analyses, algorithms and tools for data auditing,
cleaning and repair.