CSE Bytes and Pieces, Volume 16

Bytes & Pieces

Message From the Department Head

16I S S U E

W I N T E R 2 0 1 3

N E W S L E T T E R O F

T H E D E PA R T M E N T O F

C O M P U T E R S C I E N C E

A N D E N G I N E E R I N G

this issueP.1

P.5

P.10

Greetings from Penn State.

These are exciting times to be in the computing profession. Computer Science and Engineering continues to transform every aspect of our society. Progress in most dis-ciplines relies on the progress of computing. Our faculty, students, and alumni have been and continued to be on the vanguard of this transformation.

We had a very successful year in 2012. Sofya Raskhodnikova was granted tenure and promoted to associate professor, while Yuan Xie was granted promotion to full profes-sor. Computer Science at Penn State was ranked 5th in the world in citations. Com-puter Science research expenditures was ranked 8th in the country.

Our faculty and students are involved in diverse and exciting projects such as analyzing Archbishop Tutu’s genome, designing 3D computer chips, HIV virus, privacy, percep-tion in human and computer vision, crowd scene analysis, health informatics, and quantum computing.

As part of the Global Cyber Learning Factory, our students have been working with international students remotely on developing games on android platforms. Our students are in demand in industry. A recent article in the Wall Street Journal reported that we were ranked 8th in the country by recruiters.

We had Tom Mitchell, E. Fredkin University Professor at Carnegie Mellon University, deliver the Kelly Distinguished Lecture series.

I wish you all an exciting and successful New Year.

Penn State Computer Science Program Fifth in Citations

Penn State’s computer sci-ence program was ranked fifth in the world in cita-tions, according to a sur-vey by Thomson Reuters.

More than 240 programs were included in the Thomas Reuters Essential Science Indicators survey. According to the orga-nization, 970 Penn State computer science papers received a total of 16,872 citations.

The University of California, Berkeley, was ranked number one, followed by the Massachusetts Institute of Technology, Stanford University, and the University of California, San Diego.

“The top five is hard to break into,” said Raj Acharya, head of the computer science and engineering department. “The competition is very intense. Almost every school has a computer sci-ence program, including the Ivy League universi-ties, which are all top-notch programs.”

The Thomson Reuters survey includes ten years of citation data and is updated every two months.

Spotlight on Research

Department News

Alumni News

CSE Bytes & Pieces | winter 2013

spotlight on research

1

Integrated Circuits Go Vertical — Building Three-dimensional (3D) Computer ChipsIntegrated circuit (IC) was invented in 1958 by Jack Kilby, who inte-grated several transistors in a single chip. Seven years later in 1965, Gordon Moore (who was a co-founder of Intel) made a famous prediction (Moore’s Law) that the number of transistors can be inte-grated on a single chip and would double every 18 months due to technology scaling (the transis-tor size shrinks by 0.7 for every technology node). In the past 50 years we have enjoyed the Moore’s Law with increasing the number of smaller but faster transistors on an IC chip with a lower cost and more functionality, building very large scale integrated circuits (VLSI) chips for computers, mobile phones, electronic appliances, etc. For example, the first computer processor was built in 1971 with approximately 2,300 transistors on a single chip using 10 um technolo-gy, running at 1 MHz; while today the modern computer processors are built with 28nm technology, integrating billions of transistors on a single chip and running at several Ghz.

With continued and aggressive technology scaling, however, it becomes more difficult to fabri-cate nanometer transistors as we approach the physical limits. In addition, even though transistors become faster as they shrink, the wires that interconnect the transis-tors have emerged as the dominant

source of circuit delay and power consumption. Consequently, instead of simply shrinking the IC chip in 2D dimension to gain the performance benefits, building inte-grated circuits vertically – or three-dimensional integrated circuits (3D ICs) – becomes an attractive option for overcoming the barri-ers in technology scaling, thereby offering an opportunity to continue performance improvements and to extend Moore’s momentum in the next decade.

Yuan Xie, professor of computer science and engineering, and his research group at Penn State have been at the forefront of many inno-vations of addressing a wide range of challenges of the adoptions of the emerging 3D integration as the main stream technologies for future VLSI chip design.

In a 3D IC, multiple chips are stacked vertically, using various stacking technologies, such as wire-bonded, microbump, or through-silicon-via (TSV). Today, wire-bonded 3D stacked ICs have been widely used in consumer products such as mobile phones or digital cameras, to enable smaller form factor that results in higher pack-ing density and a smaller footprint. However, the emerging TSV-based 3D stacking, where thousands to millions of fast inter-chip connec-tions go through the stacked chips, has the potential to offer the great-

est vertical interconnect density, and therefore is the most promising emerging 3D stacking technology. It offers many benefits for future microprocessor designs. Such ben-efits include: (1) The reduction in interconnect wire length, which results in improved performance and reduced power consumption; (2) Improved memory bandwidth, by stacking memory on micropro-cessor cores with TSV connections between the memory layer and the core layer; providing high memory bandwidth is one of the most criti-cal challenges with the trends of multi-core and many-core micro-processor design, and therefore TSV-based 3D IC seems to be the most promising solution to scale the number of cores on a micro-processor; and (3) The support for realization of heterogeneous inte-gration, which could result in novel architecture designs and new appli-cations. Consequently, the semi-conductor industry has investigated the manufacturing technology since the early 2000s. Companies such as IBM, Intel, Samsung, and TSMC have made announcements on breakthroughs of building TSV-based 3D stacked chips.

To efficiently exploit the benefits of 3D technologies, design techniques and methodologies for supporting 3D designs are imperative. Xie and his research team have pioneered the development of EDA tools and algorithms for 3D IC design since

Penn State | www.cse.psu.edu

2

2004. For example, Xie has collabo-rated with IBM and helped modify IBM’s design infrastructure so that it can be adopted for 3D IC design. A comprehensive cost analysis for 3D chips also discovered that 3D IC design is not always more expensive than conventional 2D IC design. On the contrary, due to the yield improvement for smaller chips stacking together compared to a large 2D chip, 3D IC design can potentially result in a cheaper fabrication cost. Xie’s team is among the first to investigate the design of 3D memory caches, and the design of scan chains for testing 3D chips. The 3D CACTI toolset for cache partitioning introduced by his group has been widely used by other research groups to study the implication of die-stacking on future cache architecture.

Design space exploration at the architectural level is also essential to fully take advantage of the 3D integration technologies to build a high performance, energy-efficient integrated systems, GPUs, system-on-chips, heterogeneous systems, and other integrated systems. Xie and his team are among the first to bring the awareness of this emerg-ing technology to the computer architecture community, resulting in active study of novel architecture designs enabled by 3D stacking. One of the key benefits of 3D ICs is the heterogeneous integration of different technologies. Xie’s group first demonstrated that by integrat-ing emerging non-volatile memory for future chip-multiprocessor design, power consumption can

be saved and performance can be improved. His recent work on using stacked phase-change memo-ry to reduce checkpointing cost in massively parallel systems charters new territory and it can potentially pave the way for the smooth tran-sition from petascale to exascale computing. This research work also attracted $3 million funding from the Department of Energy to support Xie’s team and his col-laborators to study how to lever-age emerging 3D and non-volatile memory for future exascale system design.

In addition to demonstrating the novel 3D designs in general pur-pose computers and super com-puters, Xie and his students also studied how to leverage the massive memory bandwidth provided by 3D ICs for future multimedia sys-tem-on-chip design. For example, they designed a 3D stacked chip with 2 logic layers connected by TSVs and combined with 3-layer of

DRAM stacked memory so that the chip can run a very low frequency (so that power consumption can be saved) while a very high band-width of video encoding can be enabled. Figure 1 shows the chip layout of two logic layers where the embedded processor, the H.264 video encoder, and other IP cores are stacked together as 2-layer logic chips.

Xie has given 17 tutorials at pres-tigious IEEE/ACM conferences and more than 50 invited talks in industry/academia to promote the awareness of the 3D IC technology. He has been involved in exten-sive collaborations with industry partners such as AMD, HP Labs, Honda, IBM, Intel, Qualcomm, and Seagate, to help transition research ideas to industry. Xie has graduated 12 Ph.D. students with two of them as tenure-track profes-sors and others in research labs at leading companies.

Fig. 1. Chip layout of two logic layers where the embedded processor, the H.264 video encoder, and other IP cores are stacked together as 2-layer logic chips.

CSE Bytes & Pieces | winter 2013

spotlight on research

3

Cloud Computing Security ResearchCloud computing is a realization of computing as a utility, where customers submit their computing tasks to a centralized service that provides the resources necessary to execute those tasks. According to the National Institute of Standards and Technology, “[c]loud comput-ing is a model for enabling conve-nient, on-demand network access to a shared pool of configurable computing resources.” Rather than purchasing and maintaining an abundance of hardware resources themselves, customers can “plug in” to the cloud, paying for only the quantity of resources used. This is particularly attractive to those cus-tomers whose resource utilization may vary dramatically or where the costs of hardware and its mainte-nance form a significant fraction of their overall budget.

Despite a promising business model, security is a major con-cern that may limit the impact of the cloud computing paradigm. Customers need assurance that their personal or proprietary pro-cessing can be protected on sys-tems administered by a third party. For example, if a medical billing organization wants to use a cloud system to run their processing, they will have to consider how to pro-tect their processing from remote adversaries and achieve compliance with privacy requirements, such as HIPAA, when such computing is moved to a third party. Key to the safe use of any computing system is the ability to configure security policies that limit adversary access to critical processing. While the cloud vendors aim to make security

configuration simpler, through pre-configured OS distributions and default firewall policies, significant responsibility for security decisions still falls upon the customers [1].

At Penn State’s Systems and Internet Infrastructure Security (SIIS) Lab, co-directed by Trent Jaeger and Patrick McDaniel, we are studying how cloud vendors and customers can work together to configure cloud computations that protect customer processing. This study spans three critical issues in cloud computing: (1) configur-ing cloud computations to prevent known attacks on such pre-config-ured OS distributions; (2) validat-ing the provenance of the software and data that compromise cloud nodes and customer computa-tions; and (3) checking the runtime compliance of cloud computations with the expected configurations.

First, the SIIS Lab is developing tools to perform accurate grey box testing of pre-con-figured operating sys-tem distributions for known vulnerability classes. For example, Figure 1 shows how name resolution vul-nerabilities are found. Processes retrieve many resources from operating systems and middleware by name. Adversaries often have access to shared namespaces, which they can use to redi-

rect victim processes to resources of the adversaries’ choosing. Using our name resolution testing tool, we have detected over 25 previously-unknown vulnerabilities in mature programs, including the MySQL database, Tomcat servlet container, the cups printer daemon, and Ubuntu startup scripts [5]. Further, we are using this knowledge to produce defenses for such attacks and others in systems and programs automatically. For example, we have developed an operating system mechanism to block discovered name resolution attacks efficiently.

Second, the SIIS Lab is developing methods to load software systems in the cloud in a manner that enables validation of the provenance of all software and data. Figure 2 shows a method called the network root-of-trust-installation (netROTI) that

Fig. 1. Name resolution vulnerability detection consists of two phases: (a) launching an attack, and (b) detecting a vulnerability due to the attack and recovering original namespace state.

Fig. 2. Timeline of the network root-of-trust-installation (netROTI) process.

Penn State | www.cse.psu.edu

4

binds the filesystem on a cloud node or in a customer computa-tion to a proof of the installer and the system image whose installa-tion created that file system. While software is generally unchanged at installation, several files, including configuration and data files, may be customized by the installer or retrieved from third parties. The netROTI uses available trusted computing hardware to build a cryptographically strong record that enables remote parties to check the sources for every file in the filesys-tem to detect malicious input. For systems that change rarely, such as the cloud node controllers, this proof may even be used to justify their long-term compliance [3].

Third, the SIIS Lab is exploring methods to measure the runtime integrity of systems after instal-lation via the netROTI. Such methods enable cloud customers to determine whether their cloud instances are enforcing the expected security policies at runtime, includ-ing our new defenses described above. Figure 3 illustrates the high-level protocol cloud custom-ers use to monitor their computa-tions. This protocol enables clients to establish trust in each layer of the verification framework, which will ultimately justify their com-putation’s integrity. The heart of the monitoring framework is an

integrity verification proxy (IVP) that provides loadtime measure-ment (as a netROTI service) and provides runtime monitoring of the client’s computation using memory introspection techniques. To focus runtime monitoring, we measure the security policies being enforced upon the cloud computation (exter-nal and internal). This enables cli-ents to determine whether compu-tations were configured as expected. Since runtime security configura-tion changes are rare, runtime over-head due to monitoring is low [4].This protocol has been integrated into the OpenStack cloud platform. First, clients validate the visible cloud monitoring service, called the cloud verifier. Cloud verifiers moni-tor cloud nodes, including IVPs, which in turn, monitor cloud com-putations as described above. The result is a low overhead mechanism for validating the configuration of cloud computations in real cloud platforms. We are exploring a num-ber of applications of these tech-nologies with a variety of industrial partners.

In the future, we plan to extend black box testing cloud-wide, utiliz-ing our research methods to com-pute adversary access in distributed systems [2]. This work leverages available security policies to iden-tify adversary access to individual

program system calls. Further, we plan to explore methods to ret-rofit legacy code automatically with defenses to certain kinds of attacks that are

difficult or expensive for the system to prevent, such as malicious name attacks. Jaeger has presented several invited talks on cloud computing and trusted computing to promote the need for comprehensive integ-rity in cloud computing and the role of trusted computing building integrity proofs. His work in this area has been generously funded by HP Labs, Samsung, IBM Research, the Air Force Office of Scientific Research, the Air Force Research Lab, and the National Science Foundation.

References

1. Jaeger, T., Schiffman, J.: Cloudy with a chance of security challenges and improvements. IEEE Security & Privacy (Jan/Feb 2010), http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=54031582. Muthukumaran, D., Rueda, S., Talele, N., Vijayakumar, H., Teutsch, J., Jaeger, T., Edwards, N.: Transforming commodity security policies to enforce Clark-Wilson integrity. In: 28th Annual Computer Security Applications Conference (2012)3. Schiffman, J., Moyer, T., Jaeger, T., McDaniel, P.: Network-based Root of Trust for Installation. IEEE Security & Privacy (Jan/Feb 2011)4. Schiffman, J., Vijayakumar, H., Jaeger, T.: Verifying sys-tem integrity by proxy. In: 5th International Conference on Trust and Trustworthy Computing. pp. 179–201 (2012)5. Vijayakumar, H., Schiffman, J., Jaeger, T.: Sting: Finding name resolution vulnerabilities in pro-grams. In: 21st USENIX Security Symposium (2012)Fig. 3. Connection Protocol Sketch.

CSE Bytes & Pieces | winter 2013

department news

5

$18.5 Million NSF Grant to Develop Self-monitoring Health DevicesPenn State, North Carolina State University, the University of Virginia, and Florida International University will collaborate on a national nanotechnology research effort to create self-powered devices to help people monitor their health and understand how the sur-rounding environment affects it, the National Science Foundation (NSF) announced Sep. 6.

The NSF Nanosystems Engineering Research Center for Advanced Self-Powered Systems of Integrated Sensors and Technologies (ASSIST), to be headquartered on NC State’s Centennial Campus, also includes five affiliated universi-ties and about 30 industry partners in its global research consortium. ASSIST will be funded by an initial five-year $18.5 million grant from the NSF.

ASSIST researchers will use nano-materials and nanostructures — a nanowire is thousands of times thinner than a human hair — to develop self-powered health monitoring sensors and devices that operate on small amounts of energy. ASSIST researchers will make devices from thermoelectric and piezoelectric materials that use body heat and motion, respectively, as power sources.

“The ASSIST program offers an

opportunity to utilize core Penn State strengths in materials, nano-fabrication, low-power circuits, and biobehavioral health to advance human health. This is an extremely exciting opportunity for the researchers involved,” said Susan Trolier-McKinstry, Penn State professor of materials science and engineering.

“Currently there are many devices out there that monitor health in different ways,” said Veena Misra, the cen-ter’s director and pro-fessor of electrical and computer engineering at NC State. “What’s unique about our tech-nologies is the fact that they are powered by the human body, so they don’t require bat-tery charging.”

These devices could transform health care by improving the way doctors, patients and researchers gather and interpret important health data. Armed with uninter-rupted streams of heart rate read-ings, respiration rates and other health indicators, sick people could better manage chronic diseases, the elderly could be monitored from a distance and healthy people could make better decisions to keep themselves fit.

For example, personalized expo-sure data for environmental pol-lutants such as ozone and carbon monoxide could help a child suf-fering from asthma avoid an envi-ronmental trigger for an attack. Miniaturized devices the size of a pen or wristwatch will make com-pliance simpler and therefore more likely, resulting in better health outcomes and reduced health costs to society.

The center’s partner institutions will play important research roles. At Penn State, researchers will cre-ate new piezoelectric materials and devices; energy-efficient transistors; extremely low-power sensors; and help understand the correlations between environmental exposure and human health. The Penn State team includes faculty from the Colleges of Engineering, Earth and Mineral Sciences, Education, and Health and Human Development, with Tom Jackson, Penn State pro-

ASSIST aims to produce self-powered health-monitoring devices that can be worn on the wrist. (Credit: NC State University)

Penn State | www.cse.psu.edu

6

fessor of electrical engineering, serv-ing as the center’s research director.

The team from the University of Virginia will develop ways to make the systems work on very small amounts of power, while the group from Florida International University will create sensors that gather biochemical signals from the body, such as stress levels.

The results of that work, coupled with low-power radios developed by the University of Michigan, will be used to process and transmit health data gathered by the sensors to computers and consumer devic-es, such as cell phones, so patients, doctors and researchers can easily digest it. The University of North Carolina at Chapel Hill will pro-vide ASSIST with medical guidance and arrange testing of the center’s technology.

“We have assembled a comprehen-sive team that works together close-ly under a systems-driven approach to tackle this challenging set of global health problems,” Misra said.ASSIST will draw on the exper-tise of industry partners to help guide the center’s work to the mar-ketplace. These partners include companies and agencies involved in nanomaterials and nanodevices, integrated chip manufacturing, software development, bioengineer-ing and health care. ASSIST also has foreign partnerships with the University of Adelaide, the Korea Advanced Institute of Science and Technology and the Tokyo Institute of Technology.

The five-year NSF grant for ASSIST is renewable for an addi-tional five years and follows a two-year selection process by the federal

agency. The grant is among a new group of Engineering Research Center awards that invest in nanosystems.

Participants from the College of Engineering are Suman Datta (EE), Renata Engel (ESM), Tom Jackson (EE), Theresa Mayer (EE), Vijay Narayanan (CSE), Chris Rahn (ME) and Doug Werner (EE). Participants from the College of Earth and Mineral Sciences are Clive Randall (MatSE) and Susan Trolier-McKinstry (MatSE). Also participating are Annmarie Ward from the College of Education, Shedra Amy Snipes from the College of Health and Human Development, and Suzanne Adair from the Graduate School.

– Walt Mills

$1.1 Million NSF Grant Focuses on Better Methods of Analyzing Privacy-Enhanced Data The National Science Foundation (NSF) has awarded a $1.1 million grant to a multidisciplinary Penn State team to develop methods for statistical analysis of noisy and privacy-enhanced data.

Daniel Kifer, assistant professor of computer science and engineering serves as principal investigator on the NSF grant. Stephen Matthews, associate professor of sociology, anthropology, and demography, and Tse-Chuan Yang, a research

associate with the Social Sciences Research Institute, are Co-PIs.Kifer explained that information derived from confidential surveys that are analyzed must be “sani-tized” so that individuals cannot be identified. To do so, “noise” is introduced into the original data for the individuals’ protection. For example, specific information about an individual who suffers from a rare medical condition may be omitted or perturbed in the priva-cy-enhanced data.

The problem, Kifer said, is when the noisy data is analyzed using traditional techniques, it may create false positives or misleading results.

The Penn State effort looks to cre-ate new techniques for extracting more accurate statistical informa-tion from these complicated, noisy datasets. The work will place special emphasis on analyses useful for social science research.

– Curtis Chan

CSE Bytes & Pieces | winter 2013

7

department news

$800,000 NSF Research Grant Seeks to Replicate Human Pattern Recognition in ComputersPenn State and Stanford University will collaborate on a National Science Foundation (NSF) Creative Research Awards for Transformative Interdisciplinary Ventures (CREATIV) grant which can potentially advance the understand-ing of visual regularity perception in both human and machines.

This research integrates theoreti-cal, experimental and algorithmic thrusts to construct a novel concep-tual framework for predicting and understanding the full range of reg-ularity perception, both in humans, by measuring human brain activa-tion and behavior, and in machines, through a computational frame-work for adaptive symmetry detec-tion in computer vision. The ability to detect patterns in natural scenes serves critical biological needs while posing substantial computational difficulties for machine intelligence. Research on human and computer perception of pattern regularity has primarily focused on bilateral symmetry, despite a wide variety of regular patterns beyond reflection. A unique feature of the proposed project is to use symmetry group

theory as an organizing principle for the study of both human and computer perception of patterns. Symmetry group theory, instanti-ated by its subgroup hierarchy, pro-vides a formal and exhaustive cat-egorization of all regular patterns.

The project sits at an interdisciplin-ary nexus between computer sci-ence, psychology, neuroscience, and mathematics. The outcomes of this research could potentially transform the theory of human pattern per-ception and make a quantum leap in robust automatic detection of real world regularities. Because pat-terns are ubiquitous, this research impacts all information processing systems challenged by large digital datasets that are hard to explore manually. Its impact is strengthened further by a systemic outreach to the respective research communities through interdisciplinary work-shops, publications, data sharing, classroom lectures, postdoc and stu-dent training. Applications include anomaly detection in medicine and surveillance data; mobile robot localization in man-made environ-ments; and generic pattern indexing and retrieval.

The NSF CREATIV award is a pilot grant mechanism under the Integrated NSF Support Promoting Interdisciplinary Research and Education (INSPIRE) initiative, which seeks to support bold inter-disciplinary projects in all NSF-supported areas of science, engi-neering, and education research. According to NSF Director Subra Suresh: “INSPIRE is aimed to encourage cross-disciplinary sci-ence. INSPIRE will help to break down any disciplinary barriers that may exist within NSF and encour-age its program managers to use new tools, collaboration modes and techniques in the merit-review process to widen the pool of pro-spective discoveries that may be hidden from or circumvented by traditional means.” The program is intended to attract unusu-ally creative high-risk/high-reward proposals, providing substantial funding that is not limited to exploratory stages. According to the NSF record of 2012 submissions, approximately 400 formal inquiries were received, and 48 full proposals were authorized by program direc-tors and submitted. This INSPIRE

award is partially funded by the Robust Intelligence Program in the Division of Information and Intelligent Systems in the Directorate for Computer and Information Science and Engineering, and the Perception, Action and Cognition Program in the Division of Behavioral Symmetry group theory, instantiated by its subgroup hierarchy, provides a formal and exhaustive

categorization of all regular patterns – a basis for human and machine regularity perception research.

Penn State | www.cse.psu.edu

8

and Cognitive Sciences in the Directorate for Social, Behavioral and Economic Sciences. This CREATIV award will be funded by a three-year $800,000 grant from NSF.

Penn State’s Yanxi Liu, associate professor of computer science and engineering and electrical engineer-ing, will serve as principal investiga-tor (PI) and Rick Gilmore, associate professor of psychology, will serve as Co-PI. Anthony Norcia, pro-

fessor of psychology at Stanford University, will also serve as Co-PI. This award resulted from an ini-tial collaboration among the three participants and was supported by a Penn State Institute of the Neurosciences (PSIN) seed grant.

$450,000 NSF Research Grant to Develop Efficient Computer Vision Algorithms for Crowd-scene AnalysisRobert Collins, associate professorof computer science and engi-neering, has received a $450,000 National Science Foundation award “Distributed Combinatorial Optimization for Crowd-Scene Analysis,” jointly funded by the Division of Information andIntelligent Systems andComputing and Communi-cation Foundations. The goal of this three-year effort is to develop efficient computer vision algorithms based on distributed message passingfor solving crowd-scene analysis tasks such as detec-tion and tracking of closely spaced individuals. These and other vision tasks can be for-mulated as discrete combinatorial optimization problems, e.g. binary linear or quadratic programs, and studying their underlying math-ematical structure is expected to yield insights that allow larger and more challenging problem instances to be addressed. Recent theoreti-cal work proving the correctness of message passing for some classes of binary linear programming prob-lems is being leveraged to develop

practical vision algorithms for crowd scene analysis and extended to develop algorithms for find-ing good approximate solutions to harder problems. The project team is also exploring approximate infer-ence methods based on random-ization and on decomposition of

large-scale problems into collections of interrelated subtasks that can be solved more efficiently and in parallel.

The abundance of video footage from surveillance systems in public spaces has become a driving force for advances in image analysis of crowds. Data collected by auto-mated vision systems can provide

a better understanding of crowd behavior that ultimately may help develop crowd management strate-gies that promote public safety by reducing violence and speeding up evacuations. By continuously moni-toring public spaces, autonomous systems can detect early signs of

dangerous crowd behavior, and can provide law enforce-ment and event management personnel with real-time situ-ational awareness of crowd density and the existence of impediments to normal traffic flow. This work also has implications for architec-tural design of urban public spaces; for example, data generated about the behavior

of pedestrians in public settings can help isolate social factors that make some public spaces attractive while others remain underutilized. Multi-object detection and tracking tech-niques developed in this work have broader applications beyond analyz-ing human crowds, for example to track herds of wildlife, schools of fish, or colonies of cells.

CSE Bytes & Pieces | winter 2013

9

department news

Padma Raghavan Named Distinguished ProfessorPadma Raghavan, a faculty mem-ber in computer science and engi-neering, was named distinguished professor by Penn State President Rodney Erickson. Established by the Office of the President, the title of distinguished professor recogniz-es a select group of professors who have achieved exceptional accom-

plishments in teaching, research, and service.

A faculty member at Penn State since 2000, Raghavan is recognized for pioneering contributions tow-rads the development of scalable parallel sparse linear solvers and energy-efficient parallel scientific computing applications.

In 2007, Raghavan became the founding director of the university-wide Institute for CyberScience at Penn State, organized through the Office of the Vice-President for Research and charged with promot-ing interdisciplinary computation and data-enabled research and education.

Thomas La Porta Named Leonhard ChairThomas La Porta, distinguished professor of computer science and engineering, has been named the William E. Leonhard Chair in Engineering. The Leonhard Chair provides a distinguished faculty member in the College of Engineering the opportunity to continue scholarly excellence through contributions to teaching, research and public service.

A faculty member at Penn State since 2002, La Porta is director of the Penn State Networking and Security Center. La Porta is the principal investigator and director of the $35 mil-lion Communications Network Academic Research Center fund-ed by the U.S. Army Research Laboratory. He also leads a major portion of the Army facility’s

International Technology Alliance Program on network science.

La Porta is the winner of the 2007and 2009 Thomas Alva EdisonPatent Award for his work on cellu-lar networks. He is an Instituteof Electrical and ElectronicsEngineers fellow and a BellLaboratories fellow.

CSE Welcomes New Faculty Member, Paul MedvedevPaul Medvedev joined the faculty this fall as an assistant professor in both the computer science and

engineering department and the biochemistry and molecular biol-ogy department. His research is focused on the interface of biology with theoretical computer science. Specifically, he studies problems

where rigorous algorithms and analysis can have a demonstrated impact in the biological sciences. He is also interested in other areas of research such as phylogenetics, graph theory, computational com-plexity, and on-line algorithms.

Paul states that his current focus is on developing methods to detect and understand large alterations of the DNA landscape that play an important role in the development and progression of diseases. “Penn State is a great place for inter-disciplinary research. The students

especially benefit from the melting pot of different disciplines and the presence of world-renowned leaders in the field.”

Paul received his Ph.D. from the University of Toronto. Before joining the Penn State faculty, he was a post-doctoral scholar at the University of California at San Diego. In 2011, Paul was named one of “Tomorrow’s PIs” by Genome Technology Magazine. He currently holds a Quiggle Career Developmental Professorship.

Penn State | www.cse.psu.edu

10

alumni news

Outstanding Engineering AlumniEric Hennenhoefer and Jan Uhrich were two of the thirteen recipients of the 2012 Penn State Outstanding Engineering Alumni Award. The award is the highest honor conferred by the College of Engineering and recognizes gradu-ates who have received exceptional levels of professional achievements.

Eric T. Hennenhoefer received his B.S. and M.S. degrees in com-puter engineering in 1991 and 1993, respectively. Hennenhoefer is president and CEO of Obsidian Software, Inc. (now acquired by ARM). He focuses on the problem of timely and efficient microproces-sor verification. He also focuses primarily on maintaining the quality and efficiency of RAVEN, working directly with custom-ers in the field. He developed the product requirements defini-tion, architecture, implementation and deployment of RAVEN, Obsidian Software’s flag-ship product. Prior

to founding Obsidian Software, he worked as technical lead at both IBM and AMD, where his responsibilities included defining functional verification methodolo-gies, hardware emulation, formal verification, defining internal CAD tools, and evaluating third-party EDA tools. He holds five patents within the field of microprocessor verification.

Jan Uhrich earned her B.S. in computer science in 1980. Uhrich serves as vice president of the

services and solutions group for Dell Services. She is responsible for Global ProSupport Services as well as Global Product Warranty Services. Her responsibilities include technical support, support operations, field service delivery, service parts, service engineering and readiness, and command cen-ters. She also leads Global Public Services with responsibility for consulting services, managed ser-vices, support services, and vertical solutions.

Prior to this role, Uhrich held numerous executive roles across Dell Services, including enterprise sup-port, managed, consult-ing, training, and deploy-ment services. Prior to Dell, she held manage-ment positions in engi-neering and application consulting at Hewlett-Packard and Apollo Computer. She came to Dell in July 1997 after working 17 years in the computer industry, focused primarily on enterprise products.The 2012 class of Penn State Outstanding Engineering Alumni. Jan Uhrich (1st

row, 3rd from left) and Eric Hennenhoefer (2nd row, 2nd from right).

CSE Bytes & Pieces is published once a year by the Penn State Department of Computer Science and Engineering for alumni, faculty, staff, students, and friends.

Department Head | Raj Acharya

Editor and Designer | Jen Latchford

The Pennsylvania State UniversityDept. of Computer Science and Engineering342 Info. Sciences and Technology BuildingUniversity Park, PA 16802Telephone: 814-865-9505

Copyright, 2012 by the Department of Computer Science and Engineering. Penn

State is committed to affirmative action, equal opportunity, and the diversity of its workforce. This publication is available in alternative media on request.

U.Ed. ENG 13-24

Supporting CSE is a wonderful way to help us continue our efforts to advance today’s students with tomorrow’s technology. Alumni gifts provide much needed support for students, faculty, and the department overall. If you would like to make a contribution, please visit: http://www.cse.psu.edu/alumni-friends/giving or scan this QR code.

Department of Computer Science and EngineeringThe Pennsylvania State University342 Information Sciences and Technology BuildingUniversity Park, PA 16802-6822

Nonprofit Org.U.S. PostagePAIDState College, PAPermit No. 1

www.cse.psu.edu