Researchers have developed a new algorithm that helps individuals find needed information more efficiently by using both their own history and behavioral cues, as well as those of teammates working together on complex problems, according to a Mar. 17 announcement.
The new approach is based on the concept of social foraging, which is a process where individuals share information to solve problems—a concept well known in animal behavior research. Sandeep Kuttal, co-author of the work and associate professor of computer science at North Carolina State University, said, “This collaborative process of sharing information to solve problems is called social foraging – it’s a well-established concept in animal behavior research. But until now, this behavior hasn’t been incorporated into software systems designed to support problem-solving. We think it holds promise for improving software used for collaborative work – such as software engineering, scientific research, or crisis response.”
Kuttal also said that previous algorithms focused only on the history of individual users. “Our goal is to consider both a user’s interactions and how their teammates interact with the system, in order to improve recommendations for the user.” To address this, researchers created Programmer Flow by Information Scent for teams (PFIS-T), a predictive model that uses both teammates’ activity and explicit communication cues to predict what action a user will take next.
The PFIS-T model was tested with data from 30 software engineers working in ten three-person teams on code-maintenance tasks. The study compared actual team behaviors with predictions made by PFIS-T and an earlier model that relied only on individual histories. Kuttal said, “We found that team cues were even more important than we anticipated. The PFIS-T model performs best for teams whose members communicated frequently.”
Results showed that PFIS-T predicted 81.5% of team navigations and improved accuracy by up to 16.7% over models based solely on individual user history. Kuttal added, “One takeaway here is that PFIS-T could complement AI-based next-step predictors by providing a socially grounded signal about how humans actually navigate complex information spaces. We’re optimistic PFIS-T can improve existing tools and serve as the foundation for developing new ones.”
The paper describing these findings will be presented at the ACM CHI conference on Human Factors in Computing Systems in Barcelona from April 13-17.
