Dal Alert!

Receive alerts from Dalhousie by text message.

X

MCSc Thesis Defence - ONE-CLASS LEARNING WITH AN AUTOENCODER BASED SELF ORGANIZING MAP

Who:    Deepthi Rajashekar

Title:    ONE-CLASS LEARNING WITH AN AUTOENCODER BASED SELF ORGANIZING MAP

Examining Committee:

Dr. Nur Zincir-Heywood - Faculty of Computer Science (Supervisor)
Dr. Malcolm Heywood  - Faculty of Computer Science (Co-Supervisor)
Dr. Srinivas Sampalli - Faculty of Computer Science (Reader)
Dr. Andrew McIntyre - Faculty of Computer Science (Reader)

Chair:    Dr. Raza Abidi - Faculty of Computer Science

Abstract:

Building techniques that are cognizant of temporal and spatial changes in human behaviour under a one-class learning restriction represents a requirement for many applications. The motivation of this research is to demonstrate the utility of such techniques for the self identification of smart phones. A framework is designed to quantify: (i) the dissimilarity in behaviours among any two users; (ii) the exclusivity of each users behaviour (inclass) from the world (outclass). A central element of the proposed framework is to first identify a discriminating representation for each user. To this end, an autoencoder is employed in which the goal is to identify an encoding that rebuilds the original data with maximum accuracy/least loss. The hypothesis of thesis work is that such an autoencoding step provides an effective mechanism for discovering good data representations prior to the application of a data description technique, such as clustering. Both the autoencoder and the clustering steps are performed relative to a single user. Research shows that relative to the most up to date publicly available smart phone data sets, the resulting (user specific?) behavioural models are capable of uniquely identifying each user under a one-class learning constraint.

Time

Location

CS Room 211