Department of Computer Science and Electrical Engineering

PhD Dissertation Proposal

Learning and Executing Energy Efficient, Context-Dependent
Rules for Privacy and Security Management on Mobile Devices

Prajit Kumar Das

8:00am Friday, 16 May 2014, ITE325b

There are ongoing security and privacy concerns around mobile platforms that are increasingly being used by citizens. For example a newly discovered security flaw in WhatsApp that allows hackers using a malicious app to read chat messages stored on the SD card. The Brightest Flashlight application was reported to have logged precise location and a unique user identifier, which have nothing to with its intended functionality. Current mobile platform privacy and security mechanisms are limited to an initial installation phase permission acquisition method. In addition to that, the permissions are of the all or none form. This means that either the users accept all the permissions requested by the mobile app or they cannot use the app in question. Even if permissions were not structured as such, typically, users do not understand the permissions being requested or are too eager to use the application to even care to read them. These issues are present in all major mobile operating systems. Given the penetration of mobile devices into our lives, a fine-grained context-dependent security and privacy control approach needs to be created.

We propose a framework that will allow us to learn the privacy and security rules for a particular user, on their mobile devices. We do this by employing a simple user feedback mechanism. The rule learning framework consists of a “learning mode” where it observes and learns from user behavior and a “working mode” where it implements the learned rules to protect user privacy and provide security. The rules are represented to the user in plain English using an easily understandable construct. The rules are internally written in a logic based language and using Semantic Web technologies. The antecedents of the rules are context elements that are derived from an ontology using a query engine and an inference mechanism. The main contributions of our work include learning modifications to current rules and learning new rules to control the data flow between the various data providers on the user’s mobile device, including sensors and services and the consumer of such data. The privacy and security rule execution consumes significant energy due to the context detection. We create an energy model that allows us to make energy cost optimizations with regards to rule execution. We use a three-fold solution for achieving the said energy cost optimizations.

Committee: Drs. Anupam Joshi (chair), Nilanjan Banerjee, Dipanjan Chakraborty (IBM), Tim Finin, Tim Oates, Arkady Zaslavsky (CSIRO)

PhD Proposal

Distributed protocols for connectivity restoration
in damaged wireless sensor networks

Yatish K. Joshi

1:00pm Monday, 12 May 2014, ITE325b, UMBC

Decreasing costs and increasing functionality of embedded computation and communication devices have made Wireless Sensor Networks (WSNs) attractive for applications that serve in inhospitable environments like battlefields, planetary exploration or environmental monitoring. WSNs employed in these environments are expected to work autonomously and extend network lifespan for as long as possible while carrying out their designated tasks. The harsh environment exposes the individual nodes to q high risk of failure, which can potentially partition the network into disjoint segments. Therefore, a network must be able to self-heal and restore lost connectivity using available resources. The ad-hoc nature of deployment, harsh operating environment and lack of resources makes distributed approaches the most suitable choice for recovery.

Most solution strategies for tolerating the failure of multiple collocated nodes are based on centralized approaches that pursue the placement of additional relays to form a connected inter-segment topology. While they are the ideal solution for dealing with simultaneous multi-node failures, they need to utilize the entire network state to determine where and how recovery should occur. In addition to the scalability concern of these approaches, controlled placement of stationary relays in remote and inhospitable deployment area may not be logistically feasible due to resource unavailability and would not be responsive due to the delay in transporting the resources to the area. Space exploration is an example of those WSN applications in which placement of stationary relays is not practical.

In this proposal, we tackle the problem of connectivity restoration in a partitioned WSN in a distributed manner. We consider multiple variants of the problem based on the available resources and present novel recovery schemes that suit the capabilities and count of existing nodes.

Committee: Drs. Mohamed Younis (Chair), Dr. Charles Nicholas, Dr. Chintan Patel, Dr. Kemal Akkaya (SIU-Carbondale)Dr. Waleed Youssef (IBM)

MS Thesis Defense

Recoloring Web Pages For Color Vision Deficiency Users

Vikas Bansal

11:00am Thursday, May 8, 2014, ITE346, UMBC

Color vision begins with the activation cone cells. When one of the cone cells dysfunction, color vision deficiency (CVD) ensues. Due to CVD, users become unable to differentiate as many colors a normal person can. Lack of this ability results in less rich web experience, incomprehension of basic information and thus frustration. Solutions such as carefully choosing colors while designing or recolor web pages for CVD users exist. We first present the improvement in the time complexity of an existing tool SPRWeb to recolor web pages. After that we present our tool which explores the foreground-background relationship between colors in a web page. Using this relationship we propose an algorithm which preserves naturalness, pair-differentiability and subjectivity. In the last part, we add an additional step in to algorithm to ensure that the contrast in the parsed color pairs meets the required W3C guidelines. In evaluation, we found that our algorithm does significantly better in preserving pair-differentiability and produces lower total cost solutions than SPRWeb. Quantitative experimentation of modified algorithm shows that contrast ratio in each replacement pair is more than 4.5 as required for readability.

Committee: Drs. Lina Zhou (co-chair), Tim Finin (ch-chair), Yelena Yesha, Dongsong Zhang

Dissertation Defense

Feature Extraction and Fusion for Supervised and Semi-supervised
Classification: Application to fMRI and LTM Data

Wei Du

2:00pm Thursday, 24 April 2014, ITE 325B

Extracting powerful features from high dimensional noisy data promises to significantly improve the effectiveness of further analysis, especially of classification. Since there is no single feature selection and extraction method or classifier that works best on all given problems, developing effective and efficient feature selection and extraction methods and classifiers for specific applications has became one of the most active areas in the machine learning field. The aim of this dissertation is to develop novel data-driven methods for extracting and selecting the most distinguishing features for performing classification using functional magnetic resonance imaging (fMRI) and laser tread mapping (LTM) tire data.

FMRI data have the potential to characterize and classify various brain disorders including schizophrenia. However, the high dimensionality and unknown nature of fMRI data present numerous challenges to accurate analysis and interpretation. Independent component analysis (ICA), as a data-driven method, has proven very useful for fMRI analysis in extracting spatial components as multivariate features used in classification, and more recently, for the analysis of fMRI data in its native complex-valued form. In this dissertation, we first present a novel framework to extract powerful features from components estimated by ICA, allowing us to remove the redundancy and retain the most discriminative activation patterns from multivariate ICA features. We apply the proposed three-phase feature extraction framework to two real-valued fMRI data sets, and achieve high classification rates in discriminating healthy controls from patients with schizophrenia. Second, due to the iterative nature of ICA algorithms, typically independent components (ICs) are not estimated consistently during different ICA runs, and hence it is not clear which result to use further. We present a statistical framework that utilizes an objective criterion to select the best of multiple ICA runs such that the multivariate ICA features from the best run can be used for further analysis and inference. Using the proposed framework, we study the performance of a novel complex ICA algorithm for fMRI analysis, entropy rate bound minimization, which takes all three types of diversity into account, including non-Gaussianity, sample dependence and noncircularity that are present in the complex-valued fMRI data. We show that CERBM leads to significant improvement in ICs that provide higher classification accuracy, and thus is a promising ICA algorithm for the analysis of complex-valued fMRI data.

Classification using LTM data is another problem we address where we first study the use of highly multivariate solutions such as ICA and then note the advantages using lower-level features for classification. In this case, an important problem is the selection of best set of features for the best classification performance. Additionally, there are a large amount of unlabeled tire data that are easy to collect but only a few of them can be easily labeled by expert. In this dissertation, we propose a novel mutual information (MI) based approach to achieve feature splits for co-training, a practical and powerful data-driven method in semi-supervised learning. Inspired by the idea of dependent component analysis, the proposed MI-based approach presents feature splits that are maximally independent between- or within- subsets, and thus selects and fuses features more effectively than other feature split methods. Experimental results on both simulated study and LTM tire data indicate that co-training with MI-based feature splits yields significantly higher accuracy than supervised classification.

Committee: Profs. Tulay Adali (Chair), Joel Morris, Janet Rutledge, Charles E. Laberge, Vince D. Calhoun (University of New Mexico and the Mind Research Network), and Dr. Matthew Anderson (Northrop Grumman Corp.)

Click for video

Nidhi Mittal, a 2013 graduate of UMBC’s Cybersecurity Master’s in Professional Studies program talks about her experience. In this video Ms. Mital talks about the value of the cybersecurity program’s instructors, who bring with them a wealth of experience in the public and private sector. 

UMBC offers a variety of master’s degree and certificate options. Our cybersecurity graduate programs leverage a student’s experience toward a range of opportunities within the cybersecurity profession. UMBC’s in-person cybersecurity programs are designed to prepare computer science, information systems, and other experienced professionals to fill management and leadership roles in cybersecurity and cyber operations.

MS Thesis Defense

Learning Hierarchical Workflows Using Community Detection

Akshay Peshave

1:00pm Friday, 18 April 2014, ITE 325b

Workflows identified from user event logs and click-stream data are useful as knowledge bases for behavioral analysis and recommendation systems. In this study we identify abstractions or summaries of event logs modeled as user activity flow networks. The abstractions are identified based on structural properties as well as user activity flow dynamics over the network using community detection methods. We apply a fast modularity optimization and multi-level resolution approach to detect hierarchical community structure in user activity flow networks. The detected communities are compared to those detected by the information-theoretic map equation minimization approach to weigh pros and cons of the fast modularity optimization approach in the workflows context. We further attempt to identify the most probable sources and sinks of user activity in individual communities and trim the network accordingly to reduce entropy of the workflow abstractions.

Committee: Drs. Tim Oates (chair), Matt Schmill and Tim Finin

Computer Science and Electrical Engineering
University of Maryland, Baltimore County
Ph.D. Dissertation Defense

A Rapidly Deployable Image Classification System Using Feature Views

Adrian Rosebrock

9:00am Friday, 18 April 2014, ITE 346, UMBC

Constructing an image classification system using strong, local invariant descriptors is both time consuming and tedious, requiring much experimentation and parameter tunings to obtain an adequate performing model. Furthermore, training a system in a given domain and then migrating the model to a separate domain will likely yield poor performance. As the recent Boston Marathon attacks demonstrated, large, unstructured image databases from traffic cameras, security systems, law enforcement officials, and citizens can be quickly amassed for authorities to review; however, reviewing each and every image is an expensive undertaking, in terms of both time and human effort. Inherently, reviewing crime scene images is a classification task. For example, authorities may want to know if a given image contains a suspect, a suspicious package, or if there are injured people in the photo. Given an emergency situation, these classifications will be needed as quickly and accurately as possible. In this work we present a rapidly deployable image classification system using “feature views”, where each view consists of a set of weak, global features. These weak global descriptors are computationally simple to extract, intuitive to understand, and require substantially less parameter tuning than their local invariant counterparts. We demonstrate that by combining weak features with ensemble methods we are able to outperform current state-of-the-art methods or achieve comparable accuracy with much less effort and domain knowledge. We then provide both theoretical and empirical justifications for our ensemble framework that can be used to construct rapidly deployable image classification systems called “Ecosembles”.

Finally, we recognize the fact that image datasets give us the relatively unique opportunity to extract multiple feature representations through the use of various descriptors. In situations where the original dataset is not available for further feature extraction or in cases where multiple feature views are ambiguous (such as predicting income based on geographical location and census data) the Ecosemble method cannot be applied. In order to extend Ecosembles to arbitrary datasets of diverse modalities, we introduce artificial feature views using kernel approximations. These artificial feature views are constructed from a single representation of the data, alleviating the need to explicitly extract multiple feature views. We then apply artificial feature views to a diverse range of non-image classification datasets to demonstrate our method is applicable to multiple modalities, while still outperforming current state-of-the-art methods.

Committee: Drs. Tim Oates (chair), Jesus Caban, Tim Finin, Charles Nicholas, Jian Chen

UMBC’s ACM student chapter invites nominations from the graduate students in the CSEE department for student officer positions for academic year 2014 – 2015 (Fall 2014 to Spring 2015).

ACM (the Association for Computing Machinery) is a premier organization that promotes computing and technology around the US and the world. On the campus, the ACM student chapter is affiliated and supported by the UMBC graduate students association. The goal of the ACM student chapter is to foster interaction between all students, both graduate and undergraduate, in the CSEE department, provide a forum for student interaction, and opportunities for members to expand their knowledge of computing.

The positions available (and their general responsibilities):

These positions are open to graduate students only. If elected, you would be required to signup as an ACM student member. Membership fee is $19 only.

Please email us the position you would like to run for (there will be elections if we get multiple nominations for a position). Alternatively you can nominate any other person for the positions above. In that case, please send their name, email address and which position you would like to nominate them for.

Please send in your nominations by end of day, Monday, April 14, 2014 to acmofficers at lists dot umbc dot edu. Elections will take place the following week (venue, date and time to be announced later).

Computer Engineering graduate student Alexander Nelson received an honorable mention for the National Science Foundation (NSF) Graduate fellowship. Alex’s research is focused on developing innovative cyber-physical systems that can dramatically improve a person’s standard of living in an impactful way. His current and past projects have involved emergency communications, assistive devices and home automation.

His prior research in assistive devices has gained acceptance to the 2013 IEEE Sensors Conference and 2013 Real Time Systems Symposium and contributed to a Microsoft Software Engineering Innovation Foundation award. His current research gained acceptance to a Works-in-Progress session at the 2013 Real Time Systems Symposium. His research co-mentors are Professors Nilanjan Banerjee and Ryan Robucci.

The NSF honorable mention designation is considered a significant national academic achievement and provides access to cyberinfrastructure resources through NSF’s Extreme Science and Engineering Discovery Environment (XSEDE) computing infrastruture. XSEDE is the most advanced, powerful, and robust collection of integrated advanced digital resources and services in the world. It is a single virtual system that scientists can use to interactively share computing resources, data, and expertise.

If you have good software skills and are still looking for a summer internship, check out the 2014 Google Summer of Code program. You can earn $5500 by coding for an open source software project this summer. You will probably work remotely, but in close collaboration with a mentor at one of over 100 participating organizations. To maximize your chances, explore the organizations and find one that needs your skills. Details here; apply by Friday, March 21.