Department of Computer Science and Electrical Engineering

Today's New York Times has a "Room for Debate" opinion piece, Computer Science's 'Sputnik Moment'?, on the recent surge in interest in computing majors on US campuses. It asks "Will the influx of students into the field last, and can it raise American educational achievement along the way?" and features eight short essays incuding one by UMBC Professor of Sociology Zeynep Tufekci.

"Computer science is a hot major again. It had been in the doldrums after the dot-com bust a decade ago, but with the social media gold rush and the success of "The Social Network," computer science departments are transforming themselves to meet the demand. At Harvard, the size of the introductory computer science class has nearly quadrupled in five years.

The spike has raised hopes of a ripple effect throughout the American education system — so much so that Mehran Sahami, the associate chairman for computer science at Stanford, can envision "a national call, a Sputnik moment."

What would a "Sputnik moment" entail today? Will the surge of students into computer science last, and could it help raise American educational achievement?"

This complements the NYT article from last weekend, Computer Studies Made Cool, on Film and Now on Campus, about rising computer science enrollments.

” When Keila Fong arrived at Yale, she had never given much thought to computer science. But then last year everyone on campus started talking about the film “The Social Network,” and she began to imagine herself building something and starting a business that maybe, just maybe, could become the next Facebook.

“It’s become very glamorous to become the next Mark Zuckerberg, and everyone likes to think they have some great idea,” said Ms. Fong, a junior, who has since decided to major in Yale’s newly energized computer science program.

MS Thesis Defense

Image Classification and Automated Extraction
of Collocated Actin/Myosin Regions

Ronil S. Mokashi

10:00am Friday, 17 June 2011, ITE 325b

This study illuminates the aspects of cell migration, which is central to many biological processes. To understand cell migration we examine the relationship between local cytoskeletal features and local morphology. We demonstrate this relationship on cells stained for Actin and Myosin We connect the actin/myosin collocalizated structural organization to movements such as membrane protrusions. Membrane protrusions are good indicators of cell migration. Cells can sense the mechanical stiffness or the chemical identity of the surfaces they attach to. We show that these surfaces impact cytoskeletal structure. We develop a classifier to correlate the contextual features extracted from actin/myosin collocalized structure to different cell surfaces.

We also describe a new distance based metric to measure the strength of collocated multi-channel two dimensional data for user selected regions. We provide tools, implemented as plugins for the popular ImageJ toolkit, that are available for download by the general public. These tools allow biologists to specify and score regions of interest by drawing a polygon on their image with a point and click interface. Furthermore, we provide an algorithm that automatically identifies, annotates, and scores an interesting donut shaped region commonly occurring in vascular smooth muscle cells on extra cellular matrix such as dry collagen, wet collagen, fibronectin and monolayer collagen.

Committee:

• Dr. Yaacov Yesha (Chair)
• Dr. Yelena Yesha
• Dr. Michael Grasso

Registration for the Maryland Cyber Challenge and Conference (MDC3) is now open. MDC3 will provide an opportunity for students and professionals to network in a fun environment while participating in exciting games and learning about computer safety and cybersecurity skills.

MDC3 teams up to six players who will compete in one of three categories: high school, college and university, and industry professionals. High school teams will focus on cyber defense techniques whereas college, university and professional teams will compete in a capture the flag match.

Students must be enrolled at a Maryland high school, college, or university. Professionals’ employers must have an office in Maryland and must be either a company or government agency. Teams can register during a day of an orientation session or online if they are unable to attend in person. The next orientation session will held between 4:30pm and 7:00pm on Tuesday, 21 June 2011 at the UMBC Technology Center, 1450 South Rolling Road. People interested in the professional league should come between 4:30-5:30pm and students should come between 6:00pm and 7:00pm.

The sessions will give contestants and coaches insight about the event as well as tips and tricks to prepare for the competition. After registering and orientation, competitors will be able to attend practice challenges during the summer to prepare for the qualifying rounds in September and finals on October 21-22 at the Baltimore Convention Center. Scholarships and prizes will be available for winning participants.

Ph.D. Dissertation Defense

A Security Framework to Cope With
Node Misbehaviors in Mobile Ad Hoc Networks

Wenjia Li

11:00am Tuesday, 14 June 2011, ITE 325b

A Mobile Ad-hoc NETwork (MANET) has no fixed infrastructure, and is generally composed of a dynamic set of cooperative peers. These peers share their wireless transmission power with other peers so that indirect communication can be possible between nodes that are not in the radio range of each other . The nature of MANETs, such as node mobility, unreliable transmission medium and restricted battery power, makes them extremely vulnerable to a variety of node misbehaviors. Wireless links, for instance, are generally prone to both passive eavesdropping and active intrusion. Another security concern in ad hoc networks is caused by the cooperative nature of the nodes. Attacks from external adversaries may disturb communications, but the external intruder generally cannot directly participate in the cooperative activities among the nodes because they do not possess the proper secure credentials, such as shared keys. However, compromised nodes, which are taken over by an adversary, are capable of presenting the proper secure credentials, and consequently can interfere with almost all of the network operations, including route discovery, key management and distribution, and packet forwarding. Hence, it is essential to cope with node misbehaviors so as to secure mobile ad hoc networks.

In this dissertation, we address the question of how to ensure that a MANET will properly operate despite the presence of various node misbehaviors by building a holistic framework that can cope with various node misbehaviors in an intelligent and adaptive manner. The main purpose of this framework is to provide a platform so that the components that identify and respond to misbehaviors can better cooperate with each other and quickly adapt to the changes of network context. Therefore, policies are utilized in our framework in order to make those components correctly function in different network contexts. Besides the policy component, there are three other components, which fulfill the tasks of misbehavior detection, trust management, and context awareness, respectively. To validate and evaluate our proposed framework, we implement our framework based on a simulator.

The specific contributions of this dissertation are: (i) Develop a framework to combine the functionalities of surveillance and detection of misbehavior, trust management, context awareness, and policy management to provide a high-level solution to cope with various misbehaviors in MANETs in an intelligent and adaptive manner; (ii) Utilize the outlier detection technique as well as the Support Vector Machine (SVM) algorithm to detect node misbehaviors, and both techniques do not require a pre-defined fix threshold for misbehavior detection; (iii) Trust is modeled in a vector instead of a single scalar so that it can reflect the trustworthiness of a node in a more accurate manner; (iv) Sense and record various contextual information, such as network status (channel busy/idle, etc.), node status (transmission buffer full/empty, battery full/low, etc.) and environmental factors (altitude, velocity, temperature, weather condition, etc.), so that we can distinguish truly malicious behaviors from faulty behaviors and also more accurately evaluate nodes' trust; (v) Specify and enforce policies in the proposed framework, which makes the framework promptly adapt to the rapidly changing network context.

Committee:

• Dr. Anupam Joshi (Chair)
• Dr. Tim Finin
• Dr. Yelena Yesha
• Dr. Yun Peng
• Dr. Lalana Kagal (MIT CSAIL)

MS Thesis Defense

Estimating Temporal Boundaries For Events Using Social Media Data

Akshaya Iyengar

10:00am Wednesday, 15 June 2011, ITE 325b

Social media websites like Twitter, Flickr and YouTube generate a high volume of user generated content as a major event occurs. Our goal is to automatically determine as accurately as possible when an event starts and when it ends by analyzing the content of social media data. Estimating these temporal boundaries segments the event-related data into three major phases: the buildup to the event, the event itself, and the post-event effects and repercussions.

We describe a technique that estimates the temporal boundaries of anticipated events and helps to monitor changes as events unfold. In our approach we train a multiclass support vector machine (SVM) to classify event data into the aforementioned phases. We then discuss an algorithm for choosing the two class boundaries, such that the total error is minimized. We apply our technique to six events – Hurricane Igor (2010), Superbowl XLV (2011), three games from ICC Cricket World Cup 2011 and the Royal Wedding (2011). We train individual classifies for each of these events. Finally we train a general classifier and compare its performance with the individual classifiers.

The contributions of this research are presenting a set of features for detecting temporal boundaries of events, determining a reasonable value of tradeoff parameter for multiclass SVMs, evaluating the effect of smoothing SVM predictions using sliding window of different sizes and presenting the results of our approach on real event data gathered from Twitter. Our approach can potentially be used to detect the presence and scope of significant sub-events occurring during the course of an event. When applied to natural disasters and man-made disturbances, the derived data can help organizations involved in mediation efforts to track and analyze evolving events.

MS Thesis Defense

Face Recognition using Gabor Jets for Images of Mass Disaster Victims

Kavita Dabke

11:00am Friday, 10 June 2011, ITE 325B

Mass disasters such as earthquakes, tsunamis, floods, landslides, blizzards and other natural calamities affect a large number of people in a short time duration. After such emergencies occur, people affected need medical aid and are admitted into hospitals. In such conditions, it becomes difficult to locate one's family members and friends. Hospitals and medical centers take triage pictures of people getting admitted for their records. The content of these images could be very disturbing for some people to see. Such pictures cannot be posted on notification walls or internet websites for people to identify their missing family members or friends. This thesis addresses this problem by developing methods for searching triage image databases using query images provided by friends or family of missing people. The dataset for this thesis consists of mug shot images of people affected by calamity. These are also called the triage images. The test dataset consist of clean or regular mug shot images of people.

To automate the process of locating missing people, our thesis has a goal of developing a face recognition system based on Gabor Jets to match a clean image to the existing triage images. Here, a clean image means a mug shot image of a person where all features such as eyebrows, eyes, nose, lips, skin, ears, etc. are seen. The system aims at pulling up the exact match from the triage dataset into the top N matches filtered out based on a similarity measure. Face recognition has been studied for clean images, where all features are visible. We have developed a system to work on the domain of triage images by experimenting with existing Gabor Jets-based similarity measures and modifying the algorithm to best fit our needs.

Ph.D. Dissertation Proposal

Improving traffic flow forecasts for road networks
with data assimilation

Shiming Yang

3:00pm Wednesday, 8 June 2011, ITE 325b

Macroscopic models for traffic flow in networks of roads are widely used in analyzing traffic phenomena and for the management and planning of transportation road systems. These models have various simplifying assumptions in order to be tractable. Moreover, we often have only partial and inaccurate knowledge of the model parameters. Consequently, there are modeling errors to be dealt with.

An approach to mitigate our partial knowledge and modeling uncertainties, is to collect measurements of the real traffic system and use computational methods to assimilate them with the model in order to derive more accurate forecasts of the state of the system.

In this proposal, we propose to design, develop, and analyze methods for assimilating measurements from road networks to improve the accuracy of short-term forecasting of traffic flow in road networks. The proposed methods will overcome challenges due to the non-linearity of traffic flow behavior, high dimensionality of the modeled state space, and anisotropic non-Gaussian modeling and measurement error processes.

Committee:

• Dr. Kostas Kalpakis (chair)
• Dr. Milton Halem
• Dr. Yaacov Yesha
• Dr. James Smith

MSEE Thesis Defense

On Gas Detection and Concentration Estimation via
Mid-IR-based Gas Detection System Analysis Model

Yi Xin

2pm Monday, 6 June 2011, ITE 325

Due to recent development in laser technology and infrared spectroscopy, Laser-based spectroscopy (LAS) has been used in a wide range of research and application fields. A particular application of interest is mid-IR laser-based gas detection systems for health and environment assessment. The NSF-ERC Mid-Infrared Technologies for Health and Environment (MIRTHE) project has engineers and researchers from different areas. As a participant in MIRTHE, we study the performance analysis and improvement possibilities of the integrated sensing system.

Herein, we have improved the previously-developed statistical analysis model, and then used our statistical analysis model for a generic mid-IR pulsed-laser gas detection system to predict trace gas detection and concentration estimation performance, and their sensitivity to system parameters. Based on PNNL (Pacific Northwest National Laboratory) data and the Beer-Lambert law, we defined three main spectral peaks of a trace gas for detecting a target gas and evaluate 3-peak joint detection performance in terms of P_D vs. P_FA. For concentration estimation we used the relationship between gas transmittance (beta), molar absorptivity (epsilon), concentration (c), the sample-mean measurement (x_N) from the photo-detector, and number of samples (N) as the basis. Using the standard confidence interval method, we evaluated estimation reliability, and then analyzed estimation errors.

Simulated gas-detection and concentration-estimation results are presented for 17 trace gases at 1ppm and 1ppb concentrations.

Committee:

• Dr. Joel M. Morris (chair)
• Dr. Chuck LaBerge
• Dr. Gymama Slaughter

Ph.D. Dissertation Proposal

Finding Story Chains in Newswire Articles

Xianshu Zhu

1:30pm Thursday 2 June 2011, ITE 325B

Huge amounts of information are shared on the Internet every day, such as online newspapers, digital libraries, blogs, and social network messages. While there are some excellent search engines, such as Google, to assist in retrieving information by simply providing keywords, large volumes of unstructured search results returned by search engines make it hard to keep a clear picture of the evolution of an event. Moreover, in addition to events themselves, people may be more interested in finding out the hidden relationships among different events or causes and effects of an event. However, traditional search engines provide limited support for dealing with these sophisticated search tasks. In this dissertation, we try to enrich search options of existing search engines and organize search results in a more structured and meaningful way.

More specifically, we propose to develop a News Story Reader, with functionality similar to Google maps, that contains the following characteristics: (1) Search results are organized into groups of causes and impacts of events, thus helping web users navigate through the search results in a more directional and efficient way; (2) Enriched search options will allow users to search for correlations between two stories by selecting two articles as start and end points respectively producing a coherent story chain as output; (3) An interactive user interface will provide the functionality to zoom in and zoom out, and add via points to the search result.

In our preliminary work, we start with a relatively simple problem: given a start and an end article we want to find a chain of articles that coherently connect them together. We developed a random walk based algorithm that can find story chains that are coherent and relevant, and with low redundancy. We applied two intelligent pruning methods to reduce the size of the graph so that the algorithm is efficient. Moreover, our next goal is to find hierarchical story chains that can show evolution of stories at different levels of granularity. Thus, we further extended our current algorithm by using random walks on the word-document co-clustering graph with weights biased on name entities to find hierarchical story chains.

The contributions of this dissertation include (1) a News Story Reader system that can help alleviate the information overload problem; (2) design and development of two story chain finding algorithms; (3) exploration of methods that can find story chains on which news articles are connected via causes and impacts; (4) exploration of methods on story chain visualization.

Committee:

• Dr. Tim Oates (chair)
• Dr. Charles Nicholas
• Dr. Tim Finin
• Dr. Sergei Nirenburg

Ph.D. Preliminary Examination

Generating Linked Data by inferring the semantics of tables

Varish Mulwad

9:30am Wednesday 25 May, 2011, ITE 325b

A vast amount of information is encoded in tables on the web, spreadsheets and databases. Considerable work has been focused on exploiting unstructured free text; however techniques that are effective for documents and free text do not work well with tables. In this research we present techniques to generate high quality linked data from tables by jointly inferring the semantics of column headers, table cell values (e.g., strings and numbers), relations between columns, augmented with background knowledge from open data sources such as the Linked Open Data cloud. We represent a table's meaning by mapping columns to classes from an appropriate ontology, linking cell values to literal constants or entities in the linked data cloud (existing or new) and discovering or and identifying relations between columns. The interpreted meaning is represented as linked RDF assertions. An initial evaluation of our preliminary baseline system demonstrate the feasibility of tackling the problem. Based on this work and its evaluation, we are further developing our framework grounded in the theory of graphical models and probabilistic reasoning.

Committee members:

• Dr. Tim Finin (chair)
• Dr. Anupam Joshi
• Dr. Tim Oates
• Dr. Yun Peng
• Dr. L V Subramaniam (IBM Research India)
• Dr. Indrajit Bhattacharya (Indian Institute of Science)