"It is well-established in the cognitive science and psychology communities that our traits and interests significantly influence our subconscious responses. Inspired by implicit tagging where the meta-data about a multimedia content is derived from the observer's natural response , we propose to infer the user's traits and interests from the eye-tracking data. Eye-tracking data, including fixations, blinks and dilations, captures an automatic and subconscious response, which is influenced by a person's interests , traits [9, 13,34], and attention [3, 11].
"In essence, eye-tracking data is heavily influenced by a person's profile. As such, using machine learning techniques, such as supervised learning, these data can be used to infer one's profile. We are aware that closely monitoring the user's eye-gaze is conspicuous and may lead to privacy and security concerns, per se. We boldly utilize this unconventional media in hope of pushing the boundary of interaction design with a better understanding of the latent user needs. Meanwhile, we expect that those concerns would be effectively alleviated if the system can perform accurate profiling within a reasonably short period of time (i.e., just-in-time), thus mitigating the need for storing any personal information. In other words, the profiling is conducted on-the-fly, and the life-time of personal data is strictly confined to a service session, e.g. the duration of a flight. As compared to conventional methods (e.g., self-reporting, web-activity monitoring and social media mining), the method is unobtrusive and privacy-preserving because it does not keep historical, personal information. In addition, services built with this technology should be deployed with explicit consent of users regarding the usage of their eye-tracking data."