Monday, April 17, 2017
12:00 PM
209 Computer Studies Building
Dr. Reza Rawassizadeh
Dartmouth College
NoCloud Data Analytical Algorithms for Mobile and Wearable Devices


Conducting data collection experiments is a major trend in diverse scientific disciplines. In particular, the ubiquity and affordability of mobile, IoT and wearable devices have enabled us to continually and digitally record our daily activities. Due to resource limitations the data analysis is usually done on a remote host, such as a cloud. One the other hand, these devices are deeply personal and can have presence in our most private spaces. This pervasive presence means that they can recognize behaviors that we do not intend to share with others. The severity of privacy and security risks of wearable, mobile, and Internet of Things (IoT) devices can be significantly higher than traditional privacy risks such as web activities. Therefore, we believe there is a need to focus on applications and data analytical methods that can operate independently on small devices such as wearables and IoT devices, i.e. NoCloud algorithms. In this talk, I will describe two of our resource efficient data mining approaches and one information retrieval facility that could be used on wearable, mobile and IoT devices. First, I introduce a set of scalable methods for identifying frequent behavioral patterns. Second, I describe a set of methods for indexing the mobile data including: spatio-temporal event detection, clustering and contrast behavior detection algorithms for temporal multivariate data. Third, I describe a light query interface that can be run on the smartwatch and translate users’ queries into machine understandable command.



Reza Rawassizadeh is a research associate within the Department of Computer Science at Dartmouth College. He received his BSc in Software Engineering in 2003, Master of Computer Science in 2008, and PhD in Computer Science in late 2012, in University of Vienna, Austria. His research interests include wearable and mobile technologies, and data analytical methods on human-centric information of mobile and wearable devices. He publishes his work in two disciplines: pervasive computing and data mining venues such as “TKDE”, “Sensors”, “CHI” and “Communication of ACM”. In addition to his academic background, he holds several professional certifications and he has worked about six years in industry, including Siemens (the largest engineering and industrial manufacturing in Europe) and the United Nations.