For Joe-Wong ’11, JP research could blossom into start-up

Two years after graduation, the junior paper of Carlee Joe-Wong '11 on smartphone data usage and pricing is expanding into a start-up company. As a current Ph.D. candidate in the applied and computational mathematics department at the University, Joe-Wong has partnered with companies in the cellular industry to conduct consumer trials that could change the way data is priced for smartphones and mobile devices.

In the spring of 2010, Joe-Wong began independent junior research with electrical engineering professor Mung Chiang to develop a time-dependent pricing algorithm for broadband usage.

“You have to do optimization for the prices. You have to do machine learning of the usage behavior by the individuals and apps. You have to look at feedback control in different time scales,” Chiang said of the project. “For an undergraduate to take on something new, that was very challenging, both in the math involved and the creative concepts involved.”

In the course of the project, Joe-Wong developed a pricing algorithm that could be used by operators to reduce the amount of data usage during peak usage periods and reduce congestion. Other industries, such as transportation and utilities, adjust prices depending on the time of day, but the idea had never been applied to the smartphone industry, Chiang explained.

The completion of Joe-Wong’s junior paper did not mark the end of this project’s development. Post-doctorate fellows in Chiang’s lab built upon Joe-Wong’s work to create a simulation and implementation plan in the following year. A trial was conducted at the University in 2011 and included participation by 50 faculty and staff to see if the algorithm would be effective, according to Joe-Wong.

Participants in the study were asked to install a mobile application on their smartphones. Half the participants would be able to see how much data they used to keep track of data usage, and the other half would see a time-dependent pricing model on the application, according to which they would be charged. The scientists conducting the trial then analyzed how usage patterns were affected and how much data usage was shifted from the busiest peak periods, she explained.

Using his connections with colleagues in cellular companies, Chiang said he convinced company representatives in early 2012 to conduct their own trials on the basis of Joe-Wong’s junior paper. AT&T, Indian Internet service provider Reliance Communications and Alaskan ISP Matanuska Telephone Association are currently conducting trials.

“This is excellent work. There has been a lot of theoretical work done in the area, but the work that Carlee and Mung have done — that is very clever,” senior vice president of the Networking Research Laboratory at Bell Labs Krishnan Sabnani said. “It’s a very elegant solution.”

Joe-Wong, Chiang and other researchers part of the Datami team, a project of the University’s EDGE laboratory, have been working to launch the project as a start-up company called Datami, of which Joe-Wong is a co-founder. The commercialization process will depend on the results of the various trials, she said.

“At Princeton, I would say it still remains not as common a thing to see a research project going from fundamental research, mathematical research all the way to deployment and then commercialization and creation of a start-up within a short span of three years,” Chiang said.

Get the best of the ‘Prince’ delivered straight to your inbox. Subscribe now »

Joe-Wong predicts that it would take smaller operators a couple of years to implement the algorithms and larger operators at least five years.

“Because this is such a novel idea in the pricing sphere, I think it would probably face some opposition, and it would take some time,” she said.

Joe-Wong’s most recent work, which elaborated on her junior paper and was co-written with Soumya Sen, was published in the academic journal IEEE Communications in November 2012.