Fabian Bustamante



Russell Joseph



Dongning Guo



Bryan Pardo

The beginning of an excellent CAREER

Four EECS faculty members receive prestigious National Science Foundation awards

Four assistant professors in the McCormick School’s Department of Electrical Engineering and Computer Science (EECS) — Fabian Bustamante, Dongning Guo, Russell Joseph, and Bryan Pardo — have received 2007 Faculty Early Career Development (CAREER) awards from the National Science Foundation.

The CAREER program offers the NSF’s most prestigious awards for new faculty members. It recognizes and supports the early career-development activities of those teacher-scholars who are most likely to become the academic leaders of the 21st century. CAREER awardees are selected on the basis of creative career-development plans that effectively integrate research and education within the context of the mission of their institution. The minimum CAREER award is $400,000 over a five-year period.

Five McCormick faculty members also received CAREER awards in 2006: Hooman Mohseni and Seda Memik, both assistant professors of EECS; Guillermo Ameer, assistant professor of biomedical engineering; Bartosz Grzybowski, assistant professor of chemical engineering; and Pablo Durango Cohen, assistant professor of civil and environmental engineering.

For the EECS department, the high concentration of recent CAREER award winners is a strong indicator of future success. “This level of achievement for young faculty is indicative of the exciting progress that the department has made over the past two years,” says Dean Julio M. Ottino. “The CAREER award recognizes outstanding teaching and research, and is an important recognition for junior faculty.”

—Kyle Delaney

Fabian Bustamante
Fabian Bustamante researches the design, deployment, and evaluation of large-scale distributed systems in both wide-area and mobile networks. His primary focus is globally distributed services designed following a cooperative, self-organizing model. Examples of such systems include peer-to-peer data storage, group communication infrastructures, and multiplayer games. Cooperative data storage services, for example, aggregate the often underutilized disk space and network bandwidth of existing desktop computers to provide a self-scaling, universally accessible storage system.

“Most large-scale systems — such as sets of hundreds, thousands, or even a million interconnected computers — are designed to regularly and independently measure their environment for adaptation,” says Bustamante. “As these systems grow in popularity, redundant measurements will result in an unsustainable degree of monitoring and restrict the variety, number, and span of these services. I believe that any solution to systems problems in this context must be carefully crafted so as not to impose unbearable demands on the environment or the systems administrators at risk of curtailing further growth.”

Bustamante will use his CAREER award to define and explore a new model, called “3R,” for the design and implementation of distributed systems. 3R focuses on reducing aggregated control and administrative overhead by strategically reusing and recycling environmental information gathered by ubiquitous services such as content distribution networks and some peer-to-peer systems.

Russell Joseph
Russ Joseph’s research focuses on computer architecture and power-aware computer systems, including techniques for monitoring, characterizing, and optimizing performance and power consumption. Making the best use of components such as transistors, logic gates, and memory elements, his group works to develop next-generation processors that will allow for improved computing performance without overheating, wasting energy resources, or failing during critical computations.

Joseph will use his CAREER award to develop microarchitectural and system software models as well as methodology and enhancements to improve the power, performance, and reliability of multicore microprocessors. He aims to identify ways to overcome two important obstacles to future microprocessor implementation: manufacturing defects and variations that prevent silicon from being produced as intended; and in-field degradation and failure, which causes parts of the processor to stop working shortly after production.

Joseph calls this a “new paradigm” of probabilistic architectures, which proposes hardware and software support that enable the design and management of these high-level microprocessors. “We are investigating techniques that allow us to overcome these challenges by modeling and designing resilient hardware and software that monitors, makes a diagnosis, and adapts the processor to recover from variation, degradation, and potential failure,” he says. “This is a rather ambitious goal, and this award allows us to examine this topic in some detail over a five year period.”

Dongning Guo
When the One Laptop Per Child program distributes low-cost laptops to children in the developing world, users will find an interesting approach to networking. Each laptop is constantly in communication with others around it, despite a lack of central infrastructure. This connectivity allows users to chat and share files and networks without being directly connected to the Internet.

The connections create a mobile ad hoc network, an area studied by Dongning Guo. His work concerns the fundamental capacity of communication systems and the design of practical systems for achieving that capacity. Guo’s CAREER award will allow him to explore the relationships between information theory and estimation theory and their application to wireless networks.

“The current state of the art for mobile ad-hoc networks is rudimentary, providing a data rate that is orders of magnitude worse than networks with infrastructure,” Guo says. “It’s important to improve these networks because they are critical for the military and for other applications, such as disaster relief.”

“Our work is unique in that it bridges the engineering and fundamental science of information transmission,” Guo says. “We hope to unveil the boundary between what is physically possible and impossible given all constraints and assuming the best design of a communications network.”

One major project in Guo’s research group is a joint effort with six other research institutions aimed at an overall theoretical foundation for mobile ad-hoc networks. These networks allow wireless terminals — such as laptops or phones — to communicate and self-organize into efficient Internet-like networks without preexisting infrastructure.

Bryan Pardo
As digital music databases continue to grow exponentially, finding ways to automatically index, label, and access multimedia content in meaningful ways is of paramount importance. For instance, online consumers may struggle to identify a song heard on the radio if they don’t know the name of the artist or song. Finding other methods of identifying the music — perhaps by melody or timing — could improve the way consumers, performers, and academics utilize the growing archives of digital music.

Bryan Pardo develops new ways to search for and identify digital content by applying machine learning, probabilistic natural language processing, and database search techniques to auditory user interfaces for human-computer interaction. He takes a broad view of natural language, including timbre and prosody (timing, pitch contour, loudness), with an emphasis on music and speech prosody, and hopes to — per the title of his award proposal — make “Music Documents Accessible in Musical Terms.” Pardo will use his CAREER award to develop source separation and score alignment, two key technologies required to automatically find, label, and manipulate important musical structures in audio recordings of music.

“For scholars, musicians, and even casual listeners, the music document is only the beginning, a tool to initiate the task at hand. Musicians may be interested in remixing a musical recording, scholars may wish to analyze the harmonies in a piece, and others may want to remove the sound of an unwanted cell phone ring from a recording of their daughter’s flute recital.” Pardo says. “Systems able to reliably access audio features, annotate the audio with new information, and integrate the annotated audio with lyrics and musical scores would represent a fundamental improvement in our ability to access and manipulate music documents, allowing a number of new musical applications and interactions for the expert and novice alike.”