$1 Million EUREKA Grant Aims to Develop System Modeled After Consumer Recommendation Engines

A team of researchers from the University of Pennsylvania’s Annenberg and Engineering schools have been awarded a $1 million grant from the National Institutes of Health (NIH) to develop a new way of evaluating effective anti-smoking appeals.

The work is supported by NIH’s Exceptional Unconventional Research Enabling Knowledge Acceleration, or EUREKA grant. These awards support high risk research endeavors with potentially high payoff. The grants are intended to support studies that can revolutionize science in a field, even though the research carries an unusually high degree of scientific risk.

Joseph Cappella, Ph.D., from the Annenberg School for Communication, will lead the Penn team. Michael Kearns, Ph.D., from the School of Engineering and Applied Science, Department of Computer and Information Science, will bring expertise in computational modeling and algorithmic challenges to the effort. They will develop descriptors for a large number of smoking cessation advertisements, and compile preference data from smokers to develop and test algorithms that can be used to recommend anti-smoking appeals that are effective for individual smokers. The approach is modeled after recommendation systems commonly used by commercial operations like Amazon.com and Netflix where suggestions for products are based on past preferences or matches to related items.

“The design of effective messages to increase healthy and reduce risky behavior has shown incremental progress at best,” said Prof. Cappella. “This proposal radically alters conventional approaches to the selection of effective messages in health behavior change.”

Cappella explains that conventional approaches to health message research, while effective, “advance the science of message design too slowly, are driven by inadequate theory, and require very complex factorial interactions among audience characteristics, message features and the target behavior.”

The outcomes of this research include:  

  • An algorithm for preferences for effective (smoking cessation) appeals;
  • A leap beyond approaches to message design, side-stepping the tedious work in testing messages one-feature-at-a-time;
  • An approach employing methods familiar to anyone ever having bought a book on Amazon or selected a movie via Netflix; and
  • Setting the stage for automatic user friendly recommender systems.

Cappella is a researcher with Annenberg’s Center of Excellence in Cancer Communication Research (CECCR) and a member of the Abramson Family Cancer Research Institute. Among the projects CECCR has undertaken are ones aimed at understanding how smokers seek out, perceive, and process risk information. Prof. Kearns studies machine learning, artificial intelligence, and social networks; he is the founding director of the Penn Engineering’s new major in Market and Social Systems Engineering.