Feel factor: Devices to detect and respond to users' moods
This article was changed April 9 to correct the date of the Boston Marathon bombing.
Computers may not yet have feelings, but the day seems to be quickly approaching when they will be able to detect and respond to how you are feeling.
Cogito Corp., a Boston-based company, is currently working with the Defense Advanced Research Projects Agency to develop Cogito Companion, a smartphone application that gathers usage data and analyzes it for patterns of psychological distress.
The application monitors the phone’s location and time of use, and logs phone calls and text messages. Participants can choose to fill out questionnaires about their mood and can record audio diaries. Cogito’s speciality – automated speech analysis – is performed on the audio diaries. In principle, the same analysis could be applied to phone conversations.
Coincidentally, the application was being tested in April 2013 in Boston when two bombs went off near the finish line of the Boston marathon, giving Cogito a unique pre-and post-disaster data set for understanding posttraumatic stress disorder.
Dell's new research division is also working on developing a variety of computer-embedded sensors intended to detect users' moods.
"We believe that the next step in providing contextual awareness is to provide computers with better cues for actual user intent," said Jai Menon, Dell's chief research officer. "Dell Research is exploring the potential applications of brain-computer interfaces (BCI) as one element of an approach to providing computers with better cues for actual user intent."
According to Menon, the types of sensor data Dell researchers are currently experimenting with integrating include heart rate, perspiration and the like.
"Biosignal sensors may currently have niche or very specialized uses, but we feel that they have the potential to become part of the suite of sensors that would be broadly used in future computer systems," Menon said. The focus of Dell’s experiments is to better understand the capabilities of sensors like electroencephalograms (EEG) to determine if the data is reliable and trustworthy. “How reliably does data from a consumer-grade EEG sensor reflect the actual mental state of a person? Do multi-channel EEG systems provide better fidelity? Do other types of sensors provide better accuracy? Can correlating data from multiple different sensor types provide superior fidelity?”
Menon said Dell is also interested in whether there are patterns of biosensor data that are shared across all users or if a system using biosensor data must be trained for each different user.
Dell’s work is still in its early stages, according to Menon, too early to talk about specific products. But he said he foresees potential applications in healthcare, education, gaming and the workplace.
"If a game could sense the player was bored, maybe it is time to make things more challenging or change the pace," Menon said. "Similarly, sensing frustration in the player, a game may offer a clue for solving a particularly challenge. Games are designed to incite emotion: joy, triumph, amusement, terror, fascination. If the game could detect these states in users, it opens up the possibility to customize and optimize the experience for each player."
In the workplace, if your computer senses you are working intensely, it might divert incoming phone calls directly to voice mail. A teacher might use the embedded sensors to learn how engaged the class is and to change tactics when engagement flags.
"The possibilities are endless," Menon said. "Our core intent is to make the capabilities of biosensors part of the portfolio of everyday capabilities that will bring value to people and improve end-user experience. Biosensors offer the promise of allowing computers to be much more intuitive about needs and mental state of users."
Posted by Patrick Marshall on Apr 08, 2014 at 7:04 AM