Can Computers Read Our Eyes? Using Eye-tracking Input for Speed Reading Application

Date: 
Tuesday, February 7, 2017 - 11:00
Location: 
TH 434
Presenter: 
Nada Attar
Abstract: 
The effectiveness of reading applications depends on a number of factors, including the reader’s cognitive processing and reading speed, as well as the interface design. We discuss the use of eye movement data analysis as a tool to improve reading interface applications and to gain further insight into readers’ cognitive processes as reading speeds vary. Our first experiment tested the possibility to improve reading interfaces using eye fixation data and successfully demonstrated faster reading. We designed a second experiment to measure human cognitive processing at different reading speeds, focusing on pupil size and eye fixation as indicators. We analyzed the data as functions of reading speed and compared it to performance in subsequently administered comprehension questionnaires. Our results show a strong correlation between pupil size and the subject’s comprehension under different speed rates. In conclusion, eye fixation data can be used to improve the reading interfaces, while pupil data can assess understanding in lieu of comprehension at various reading speeds. Our approach could be an effective tool for designing better reading interfaces and applications.
Bio: 

Nada Attar earned her Ph.D. in Computer Science from the University of Massachusetts Boston in December 2016. Her research interests span designing and testing interfaces that enhance user performance. Her focus is on using machine learning to filter psycho-physical data and exploring neural networks to understand data patterns that reflect the user’s efficiency. During her Ph.D., she was a Fellow at the Harvard University in the Visual Learning Laboratory and designed a mobile interface for dyslexia. She made new paradigms to estimate the cognitive state of users within a short interval of a task. Nada was also a lecturer in computer science at UMass Boston and worked for IBM T. J. Watson Research Center in New York in their Customer Experience and THINK labs. She completed her Master's in Computer Science at Tufts University. Nada has a patent, several publications including press, and she is the founder president of Computer Science Graduate Students Association (CSGSA) at UMass.