SURF-IT Research Projects  


Participants | Research Projects | Schedule of Activities | Poster Session Program | Photos

SURF-IT Research Projects

The following faculty-mentored research projects are available during the 2008 SURF-IT Program. They are divided into their own unique areas of research. Select a link for an overview of the project, associated faculty mentors, project prerequisites, and related publications.

Undergraduate Research Projects Mentored by Calit2 Faculty

    1) “Semantic” Information Extraction from Free Text 

    2) A Centrifugal Microfluidic System for Cell Disruption Using a Pulsed Near-Infrared Laser 

    3) A precision toolbox for rapid, high density/high coverage hydrogen-deuterium exchange mass spectrometry on TOF instruments, covering the data pipeline from spectral acquisition to HDX rate analysis 

    4) A Web-Querying Approach to Web People Search 

    5) Computational Metaphor Identification in Political Blogs: Tools for Blog Readers 

    6) Designing Computer Games that Foster Ethical Behavior in Grades 1 and 2 

    7) Extremely Large-Scale 3D Object Viewing 

    8) Extremely Large-Scale Animation 

    9) Fanfiction University 

    10) Green IT 

    11) Human-Mediated Networking 

    12) PageTurner – An automatic Assistant for Guitarists 

    13) Parallelization of Telemedicine Benchmark for the Xbox 360 Architecture 


 Project #1:  “Semantic” Information Extraction from Free Text
Faculty Mentor:  Professor Sharad MehrotraComputer Science

Description:  The XAR project is the flag-ship project in the “semantics powered information synthesis” that we have been exploring within the NSF-funded RESCUE project at Calit2. The primary focus of the investigation is to explore how semantics (taken very generally) can help in information synthesis tasks; within XAR the focus is on text synthesis tasks such as information extraction from free text. By information extraction from text we are referring to the automated extraction of sets of facts or events from a number of (natural language) text documents; for instance a corpus of online news stories, persons’ Web pages etc.

Over the past 1 ½ years we have:
• Developed an open-source and extensible text synthesis system, XAR (see reference 1 below), that has been made available to many research groups within UCI.
• Developed a framework for the systematic incorporation of semantics in the form of declarative integrity constraints; we have used such integrity constraints for improving the accuracy of information extraction at the level of slot-filling from free text with very positive results.

Student Duties and Expected Outcome:
The proposed SURF-IT project is focused on several aspects of expanding the scope of XAR, involving a healthy mix of algorithmic design, system implementation and experimental evaluation in real-world extraction applications. The proposed work includes the development of
1) Generalizing the XAR extraction model and extraction rule language to handle the extraction of “nested” structures,
2) Expanding the (current) scope of semantics, currently restricted to declaratively specified integrity constraints, to incorporate semantics lifted form the data itself, and
3) Experimental evaluation of the system in large extraction applications.

The expected outcome is a substantially enhanced version of the existing XAR extraction system and results of application to larger extraction tasks. If successful, we see strong possibilities for the work resulting in a research paper and/or system demonstration at a major conference.

Prerequisites: Prior course work in 1) Databases, 2) Information Retrieval
Knowledge of JAVA programming.

Recommended Web sites and publications: 
   Naveen Ashish and Sharad Mehrotra, XAR: Information Extraction from Free Text:
   Naveen Ashish and Sharad Mehrotra, Exploiting Integrity Constraints in Free Text Information Extraction, UCI/Calit2 Unpublished Technical Report 2008 (Please contact for copy):

 Project #2:  A Centrifugal Microfluidic System for Cell Disruption Using a Pulsed Near-Infrared Laser
Faculty Mentor:  Professor Marc MadouMechanical & Aerospace Engineering

Description:  The first step in any study involving genetic analysis is the extraction of the genetic material (DNA or RNA) from living cells. There are many techniques for the disruption (breaking apart) of cells in use today that are based on chemical, enzymatic, mechanical, and physical-chemical principles. The most commonly used methods in biotechnology laboratories rely on chemical and enzymatic principles. The main disadvantages of those procedures include labor intensiveness and the need for additional purification steps. In order to simplify the process of sample preparation, a robust, non-damaging and rapid cell disruption method would be very useful. In this study, the use of pulsed near-infrared (1000 to 1500nm) lasers will be explored for use in combination with a centrifugal microfluidic system for the automated, high throughput disruption of cells. The yield and quality of the released DNA compared to those yielded by reference methods will be used to gauge the relative effectiveness of the system.
The compact disk (CD) as a microfluidic platform offers unique advantages over manual sample preparation systems. They include automated operation and highly parallel processing making it optimal for reducing costs and increasing the precision of the results. Sample preparation systems – from cell disruption to purification for further downstream analyses– have been accomplished on CD-based devices.
The goal of this project is the development of a novel, simple and efficient high throughput, microfluidic cell disruption device utilizing near-IR pulsed laser technology.

Student’s Research-Related Duties and Expected Outcomes:
The student will learn how to design and fabricate microfluidic devices, grow yeast and E. coli cultures, and gain hands-on experience with laser optics.

Prerequisites: Basic engineering and science training

 Project #3:  A precision toolbox for rapid, high density/high coverage hydrogen-deuterium exchange mass spectrometry on TOF instruments, covering the data pipeline from spectral acquisition to HDX rate analysis
Faculty Mentor:  Professor Paul D.. GershonMolecular Biology & Biochemistry

Description:  The fast-developing and technically challenging discipline of identifying and examining all the proteins of a cell, tissue or organism is referred to as “proteomics”, and the most powerful branch of proteomics uses the mass spectrometer (an accurate and sensitive molecular weighing balance). Because of the fast growth and specialization, software development lags quite a long way behind the development of hardware. One branch of proteomics involves the identification of changes in 3D shape or “conformation” of protein molecules in real time, or when perturbed by other molecules. One phrase often used for this kind of work is “experimental molecular dynamics”, and this is the current frontier of the growing field of structural biology (where the structural biology provides a snapshot, and the molecular dynamics the movie). This molecular dynamics work is being pursued, in the Gershon lab, via a proteomics approach in which changes in exposure of parts of molecules to the surrounding water or solvent are assayed. By replacing the solvent water is replaced with heavy water, then weighing parts of the protein molecule after various times of exposure to the heavy water, we can elicit the rate of exchange of heavy and regular protons which are attached to various parts of the protein, and thereby understand which parts of the protein are sequestered within the 3D architecture and which are exposed on the surface. Because the numbers of data points multiply through different steps of the experiment, final datasets are very large (50,000 or more data points), and we have therefore created a unique series of modular programs for templating, collating, filtering, aligning, searching and editing data, in various formats, via a combination of tabular and GUI interfaces, through the pipeline from instrument raw output to the generation, analysis and comparison of the final exchange rates plots. Some modules are complete, while others require development. This protein molecular dynamics work is NIH-funded, but our developing interest in the software environment extends a little beyond that mandate of the NIH-funding, hence our interest in exploring SURF-IT. An initial draft manuscript on the software (entitled as this document) is close to completion. We have had two part time students work on the project to date, one of whom is current but will take a summer sabbatical at Yahoo. The faculty mentor is quite heavily involved in much of the programming work himself. The work really needs to be continued through the summer in order to maintain momentum.

Student’s Research-Related Duties and Expected Outcome:
The student will be programming interactively with the lab P.I. and other lab members. The student will develop and understanding of proteomics, mass spectrometry, molecular dynamics and a molecular biology lab environment, as well as honing programming skills. In terms of output, we would aim for the student to develop/complete two or three modules of the overall program-set during a full-time summer session.

Prerequisites: Intuitive programming skills and capacity for abstract thought are essential. Knowledge of Excel / VBA or an ability to pick it up within three or four days would be typical of what we would expect from a computing student on campus. A background in biology and/or chemistry would be very helpful.

Recommended Web sites and publications: 
   Laura S. Busenlehnera and Richard N. Armstrong (2005). Insights into enzyme structure and dynamics elucidated by amide H/D exchange mass spectrometry. Archives of Biochemistry and Biophysics 433 34–46:

 Project #4:  A Web-Querying Approach to Web People Search
Faculty Mentor:  Professor Sharad MehrotraComputer Science

Description:  Searching for people on the Web is one of the most common query types to the web search engines today. However, when a person name is queried, the returned web pages often contain documents related to several distinct namesakes who have the queried name. The task of disambiguating and finding the webpages related to the specific person of interest is left to the user. Many Web People Search (WePS) approaches have been developed recently that attempt to automate this disambiguation process. Nevertheless, the disambiguation quality of these techniques leaves a major room for improvement.

This project will study a new server-side WePS approach. The approach is based on collecting co-occurrence information from the Web and thus it uses the Web as an external data source. As our initial solution, we have developed a skyline-based classification technique for classifying the collected co-occurrence information in order to make clustering decisions. The clustering technique is specifically designed to (a) handle the dominance that exists in data and (b) to adapt to a given clustering quality measure. These properties allow the initial framework to get a major advantage in terms of result quality over all the latest WePS techniques we are aware of, including all the 18 methods covered in the recent WePS competition.

Student’s Research-Related Duties and Expected Outcomes:
Several new research directions will be explored. First, we want to develop a solution for new types of queries: those that involve context. Surprisingly, this very useful query type has been explored to a very limited degree. Next, we want to work on enhancing the skyline-based classification algorithm that we have. In addition, we want to explore the idea of query reformulations. The idea is that by analyzing the results for some of the queries to the web, it might be possible to detect that the insufficient merge evidence in their result is due to the poor query formulation. In that case query often can be reformulated to get better answers. The next thing to work on is data normalization: the collected co-occurrence data and derived feature will need to be "normalized" for different queries in a formal fashion, which has turned to be a nontrivial challenge for the co-occurrence data. We also would like to investigate the predictor-corrector idea: for WePS, the properties of the different sets of returned top-K webpages (answers for the queries) are turned out to be quite diverse. This makes it difficult to set the parameters of the disambiguation algorithms to get the best result. The idea is to use predictive techniques to estimate the properties of the dataset being analyzed, and the based on these estimations use a proper values for the algorithm parameters to get better results. The clustering itself needs to be modified. Specifically merges should be analyzed to see that such situations like merging of two large clusters based on weak evidence for just two elements of those clusters does not occur, and that merging does not happen for other problem scenarios. We currently are using the dice similarity to generate some of the features whereas a proper principled way should be designed to select the features. We also would like to explore the techniques to use external data to better the disambiguation process. Specifically we will look into using US Census data and onotlogies for that goal. Finally, we want to study ways to properly segment Web pages to retrieve only the most relevant information about the namesake of interest: querying with irrelevant information is likely to add noise to the query result and removing or preventing such noise should lead to quality improvement.

At the end of the project, the student is expected to develop a new WePS system that is faster than our current initial solutions and achieves a higher disambiguation quality. A successful completion of the project will imply that the students will write, or be a large part of writing, one or two research publications on the accomplished research.

Prerequisites: The student should be open-minded and have excellent analytic and programming skills (C++, MS Visual Studio C++, web crawling). High GPA and prior programming experience will be a strong plus. The project involves using several third party software packages, such as Stanford Named Entity Extractor, GATE extractor, mathematical solvers, classifier suits, and the students should be willing to learn how to use such packages.

Recommended Web sites and publications: 
   Project white papers and other publications will be provided to the student after the selection is done.:

 Project #5:  Computational Metaphor Identification in Political Blogs: Tools for Blog Readers
Faculty Mentor:  Professor Bill TomlinsonInformatics

Description:  The communications media are experiencing democratization. Rather than receiving news from large corporations, many individuals now use various forms of new media as their primary source of information. One such medium is political blogs (or weblogs), which contain political news and commentary, often with a very distinct personal voice and readily apparent political affiliation or ideology. As more members of the voting public turn to blogs for their news, the ways in which bloggers frame issues, both explicitly and implicitly, becomes progressively more important. One aspect of that framing is the conceptual metaphors used to describe different situations. For example, saying that candidates "attacked" each other highlights the adversarial way in which one candidate must defeat another. However, an alternate conceptual metaphor would be framing an election as an artistic performance, wherein the goal is not to destroy one's opponent but rather to demonstrate superior skill and technique at a set of difficult tasks.

This project employs computational techniques to identify such conceptual metaphors in political blogs. The main focus will be the development and evaluation of a tool that readers can use to examine conceptual metaphors in the political blogs they read. As the user reads, the tool will indicate potential metaphors, highlighting words in the text that instantiate those metaphors. There will also be a web application allowing users to explore computationally identified metaphors in a number of different datasets, including various blogging communities and political speeches. The goal of these tools is to encourage users to think about what a given metaphor highlights, what it hides, and what other metaphor might highlight different aspects of the same situation. By fostering such critical thinking and creativity about metaphor, this project has the potential to influence discussion and debate among blog readers, bloggers, and possibly even the politicians for whom they vote.

Student's Research-Related Duties and Expected Outcomes:
The student will be responsible for the design and implementation of (1) a tool that integrates with a web browser, enabling users to view instantiations of conceptual metaphors in online content as they read and (2) a web application allowing users to explore and visualize conceptual metaphors in a variety of textual datasets. Furthermore, the student will also be required to evaluate the use of the tool using a variety of methods.

Prerequisites: The student should be familiar with web-based programming, including at least some of Javascript, AJAX, (My)SQL, Firefox Extensions, Ruby on Rails, and/or Python. Student should also have some knowledge of qualitative evaluation techniques, e.g., ethnographic methods, and quantitative evaluation techniques, e.g., survey design and ANOVA statistical tests.

Recommended Web sites and publications: 
   Baumer, E., Sueyoshi, M., and Tomlinson, B. (2008, to appear). Exploring the Role of the Reader in the Activity of Blogging. ACM Conference on Human Factors in Computing Systems (CHI 2008).:
   This paper describes findings from a previous study of blog readers that will help in part to inform the development of the tools in this project.:
   In these two interviews, linguist George Lakoff describes his analysis of some of the metaphors used to frame conservative and liberal ideologies.:
   Bill Tomlinson/Faculty Profile:

 Project #6:  Designing Computer Games that Foster Ethical Behavior in Grades 1 and 2
Faculty Mentor:  Professor Liane BrouilletteEducation

Description:  For this project, targeted at early elementary school, ethics will be defined as “how we treat others.” The project is proposed as a preliminary step in the design of a computer game that will combine visual images and narrative to heighten children’s awareness of others’ needs and to encourage pro-social behavior. A key component of the final game project will be to design stories that involve decision-points where children choose an action, then see the consequences that follow from their choice. Characters will be designed to expose students to favorable portraits of those who are “different” and of individuals who accept people who are “different” through showing tolerance, understanding, and empathy toward members of groups frequently discriminated against.

The project builds on a successful program, which used narratives to encourage ethics and tolerance in older children, that was developed by faculty at the Center for the Scientific Study of Ethics at UCI. In this new project, which is planned as a component of a larger program funded by an Improving Teacher Quality grant awarded by the California Postsecondary Education Commission, those faculty will collaborate with faculty from the Center for Learning through the Arts and Technology. As part of this project, the undergraduate student will participate in initial planning activities related to designing the computer game.

Students’ Research-Related Duties and Expected Outcome: The project proposed for this summer is a preliminary survey of currently available electronic games and websites designed for the target audience (children in grades 1 and 2). Under the guidance of sponsors Kristen Monroe and Liane Brouillette, the undergraduate researcher will search out the most successful of the electronic games and websites designed for the target audience, then analyze the commonalties among these resources. Of special interest will be the user interfaces, computer graphics, and audio utilized to assist young children in understanding the on-line text. In addition to on-line and library research, the undergraduate researcher will interview experts in the field, who will be able to identify the most effective components of successful games. Through this process, the undergraduate researcher will gain an enhanced awareness of the characteristics of successful electronic games and their instructional uses. The final product will be a report summarizing findings.

Prerequisites: Prerequisites: Prior familiarity with issues related to video games and instructional technology is required; some previous experience with a field research project on technology and learning is highly desirable.

Recommended Web sites and publications: 
   Grubb, A. & Hines, M. (Eds.) (2000). Tearing down barriers and building communities: Pedagogical strategies for the web-based environment. Issues in web based pedagogy. Westport, CT: Greenwood Publishing, pp. 366-380.:
   Cole, R. A. (2000). Issues in web based pedagogy. Westport, CT. Greenwood Publishing.:

 Project #7:  Extremely Large-Scale 3D Object Viewing
Faculty Mentor:  Professor Stephen F. JenksElectrical Engineering & Computer Science

Description:  The ability to view and manipulate high-resolution 3D objects on very large tiled displays will be useful to many disciplines, including visualization of biological models, mechanical and civil engineering projects, and CAD models. In order to make the problem tractable, drawing of the dataset needs to be distributed across the display computers that make up the tiled display. This data distribution is a significant challenge and needs evaluation of competitive approaches. This project will develop a viable approach to high-resolution 3D data display, with the end result of being able to display such objects on HIPerWall.

Student’s Responsibilities and Duties:
The undergraduate student researcher will investigate various approaches to 3D object display on tiled displays and evaluate their performance and viability. Once a viable approach is chosen, the student researcher will implement it for the HIPerWall using OpenGL, network code, and C/C++.

Prerequisites: The student researcher should have demonstrated programming ability, strong motivation to develop state-of-the art software systems, and excellent communication skills. Programming experience with OpenGL, C/C++, and networking a plus, but not required.

Recommended Web sites and publications: 
   OpenGL(R) Programming Guide: The Official Guide to Learning OpenGL(R), Version 2.1 (6th Edition) ISBN-13: 978-0321481009:
   OpenGL(R) Programming on Mac OS(R) X: Architecture, Performance, and Integration ISBN-13: 978-0321356529:
   Distributed Rendering for Scalable Displays, Humphreys, G.; Buck, I.; Eldridge, M.; Hanrahan, P.; Supercomputing, ACM/IEEE 2000 Conference 04-10 Nov. 2000 Page(s):30 - 30:
   SOAR Terrain Engine :
   Blender a 3D rendering tool:

 Project #8:  Extremely Large-Scale Animation
Faculty Mentor:  Professor Stephen F. JenksElectrical Engineering & Computer Science

Description:  Extremely large animations (tens or hundreds of megapixels) would be very valuable to diverse disciplines, including biomedical engineering, life sciences, computational chemistry, nuclear physics, and art, because they could show extremely high-resolution time-series data in intuitive ways. Such animations are extremely challenging because of the very large data volumes involved, the expense of decompressing such data, and coordination the display of such data on tiled display systems, such as HIPerWall, with its 200 million pixels of screen area. This project will extend the state of the art in high-resolution animation, with the end result of being able to present them on HIPerWall.

Student’s Research-Related Duties and Expected Outcomes:
The undergraduate student researcher will investigate various approaches to large animation and evaluate their performance and viability. Once a viable approach is chosen, the student researcher will implement it for the HIPerWall using OpenGL, network code, and C/C++.

Prerequisites: The student researcher should have demonstrated programming ability, strong motivation to develop state-of-the art software systems, and excellent communication skills. Programming experience with OpenGL, C/C++, and networking a plus, but not required.

Recommended Web sites and publications: 
   OpenGL(R) Programming Guide: The Official Guide to Learning OpenGL(R), Version 2.1 (6th Edition) ISBN-13: 978-0321481009:
   OpenGL(R) Programming on Mac OS(R) X: Architecture, Performance, and Integration ISBN-13: 978-0321356529:

 Project #9:  Fanfiction University
Faculty Mentor:  Professor Rebecca W.. BlackEducation

Description:  This project proposes the design, implementation, deployment, and initial evaluation of a unique online space, Fanfiction University, aimed at scaffolding adolescents’ development of literacy and composition skills. Fanfiction refers to fan-authored texts based on forms of media and popular culture such as movies, books, music, television, and video games. With the advent of the Internet, Fanfiction has become a literacy phenomenon that is difficult to overlook, with the most popular multi-fandom archival site,, currently housing over a million fictions on servers in the U.S., Europe, and Asia, including over 320,000 Harry Potter fictions alone. More important than sheer numbers though, is the fact that a great number of these fan authors are adolescents, English language learners (ELLs), and/or struggling writers and readers, many of whom claim English as their worst subject. Moreover, these youth are voluntarily engaging in literacy practices, such as writing, reading, and peer-reviewing , that many educators struggle to implement effectively in classrooms. The proposed project will leverage the popularity of Fanfiction to create an online space that encourages novel interactions among authors and readers and enhances their activities via a novel technological system. Fanfiction University will serve a much-needed purpose in the fan community— providing a central location not only for writers to archive their fictions, but also to seek resources and different forms of constructive feedback that are tailored to their needs as individuals, writers, and learners —and will provide a model for an informal learning environment that promotes literacy development in adolescents.

Student’s Research-Related Duties and Expected Outcomes:
This student will be involved in all stages of the deployment of this community-based web site. He/she will help to add functionality to an existing prototype of the site, test the site with various OS/ browser combinations, deploy the site with a group of beta testers, revise the site based on initial feedback, manage larger-scale deployment, and help document the process in a co-authored scholarly publication by the end of summer. The student will develop skills in creating and managing a web-based project, and in scholarly writing.

Prerequisites: The student should be familiar with Ruby on Rails and CSS, and be a proficient writer. Personal interest in anime, Harry Potter, or other aspects of fandom a plus.

Recommended Web sites and publications: 
   Black, R. W. (2005). Online Fanfiction: What technology and popular culture can teach us about writing and literacy instruction. New Horizons for Learning Online Journal, XI (2), Spring 2005.:
   Hartl, M. and Prochazka, A. RailsSpace: Building a Social Networking Website with Ruby on Rails.:

 Project #10:  Green IT
Faculty Mentor:  Professor Bill TomlinsonInformatics

Description:  Green IT is a field that explores the juncture between two growing trends. The first trend involves the growth of environmental concern across many human communities. The second trend involves information technology – the use of digital tools and techniques for manipulating information, and the social phenomena that surround these tools and techniques. The rapid adoption of IT is transforming societies around the world, affecting many different topics from communication between people to the workings of international politics. Green IT brings together these two areas, environmental issues and information technology, and examines the ways in which they overlap and connect to each other.

Many different projects relating to Green IT are under way in Professor Tomlinson's lab, in Calit2, and across the campus. The summer research opportunity engages with Green IT through several smaller projects, all tied together around the same theme.

Student’s Research-Related Duties and Expected Outcomes:
This research project will have three main elements. One part of the summer will be spent helping Professor Tomlinson complete a book on the social aspects of Green IT, including background research, discussion of the book's content, and editing for content and style. A second part will involve assisting in the development in the next stages of the GreenScanner project, a community-generated database of environmental impact reviews of consumer products. The third part will be spent helping to construct a web site tying together the many different Green IT projects currently under way at UCI. Through these projects, the student will develop skills in academic research and writing, as well as web design and programming.

Prerequisites: The student should be a proficient writer, and be familiar with HTML. Familiarity with Ruby on Rails, CSS, and SQL a plus. Students with a personal interest in environmental themes preferred.

Recommended Web sites and publications: 
   Tomlinson, B. 2008. "Prototyping a Community-Generated, Mobile Device- Enabled Database of Environmental Impact Reviews of Consumer Products." In: Hawai'i International Conference on System Sciences (HICSS), Social Spaces Minitrack. Big Island, HI.:
   Bill Tomlinson/Faculty Profile:

 Project #11:  Human-Mediated Networking
Faculty Mentor:  Professor Bill TomlinsonInformatics

Description:  This project involves the development of an infrastructure for a new model of interactivity called human mediated networking (HMN). HMN can be defined as the use of human effort to transfer data between computational systems, rather than expending natural resources to drive the exchange. This process is effectively the inverse of computer-mediated communication: in CMC, computers are used to facilitate communication between people; in HMN, people help computers to communicate.

This summer, the team is building an interactive demonstration of HMN. In this demonstration, a series of monitors are arrayed around an exhibit hall. Several mobile devices are each equipped with a single sensor that can read and display location-specific phenomena, such as audio volume, network strength or the number of nearby Bluetooth devices. Visitors walk around the hall with the mobile devices, gathering sensor information about their current location. Visitors can then wirelessly transfer these readings to a nearby monitor simply by pressing a button attached to the monitor. This updates the monitor's information, allowing it to display recent sensed information even without sensors attached. Participants with appropriate hardware will also be able to download the software and use their own devices, for example with this notebook sensing light levels with its camera. A proposal for this exhibit has been submitted for exhibition at the ACM SIGGRAPH 2008 conference (August, Los Angeles).

The Human Mediated Networking exhibit demonstrates a network made up of collocated mobile devices connected by human users. In addition to increasing robustness, distributing sensing capabilities among a population of mobile devices reduces the need for each member of the population to contain components for each kind of sensing. For example, only a few devices need to have thermometers in order for the entire group to know the temperature. Reducing the number of components used can reduce the size of devices, the cost of devices to the end-user, and provide a more sustainable alternative to purchasing a new mobile device whenever a new capability is desired. Beyond the impact on the design of mobile devices, this exhibit helps to highlight possible directions that human-computer interaction might take in the coming decades.

Student’s Research-Related Duties and Expected Outcomes:
The student will undertake a combination of several different research activities. In Weeks 1-2, s/he will conduct a literature search for previous work relating to this area, drawing on both computing and environmental science publications. In Weeks 3-6, s/he will be involved in the development and implementation of the Human Mediated Networking exhibit. Depending on the student's expertise, s/ he may assist with either the programming of the software or the analysis of the environmental factors of the project. In Week 7, the student will be involved in the presentation of this project at SIGGRAPH 2008 if it is accepted, having the opportunity to demonstrating their work to several thousand attendees. In Weeks 8-10, the student will be involved in the process of writing up the results of this research, with the goal of completing a journal- quality publication jointly authored with Tomlinson and graduate student Joel Ross by the end of the summer.

Prerequisites: The student should have expertise either in computing or environmental studies. If the student's primary focus is computing, skills should include communication technologies (e.g., Bluetooth, IrDA), visualization (2D or 3D) and/or interface/user experience design. If the student's primary focus is environmental studies, skills should include life cycle analysis, economics and/or sustainable development, with some familiarity with statistics a plus.

This summer fellowship will be funded by the National Science Foundation (NSF). Dr. Tomlinson has been awarded a supplement to his existing grant, to support a summer research student. At Dr. Tomlinson’s request the student fellowship will be integrated into the SURF-IT program. The student will be chosen by the same procedures and standards, will be expected to meet the same requirements, and will have access to the same activities as the other SURF-IT participants.

Recommended Web sites and publications: 
   Blevis, E. Sustainable interaction design: invention & disposal, renewal & reuse. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (San Jose, California, USA, April 28 - May 03, 2007). CHI '07. ACM Press, New York, NY, 503-512:
   Huang, E. M., Truong, K. N. (2008) "Breaking the paradigm of disposable technology: Opportunities for sustainable interaction design for mobile phones." To appear in the Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI 2008). Florence, Italy. :
   Bill Tomlinson/Faculty Profile:

 Project #12:  PageTurner – An automatic Assistant for Guitarists
Faculty Mentor:  Professor Lee SwindlehurstElectrical Engineering & Computer Science

Description:  For many concert musicians, the aid of a reliable assistant can prove to be invaluable, as dramatically shown in a recent movie [1]. The person must be well versed in the music played, should be capable of following the music being played on the sheet, and should anticipate the need of the artist by turning the page no sooner and no later than needed. Recently, an electronic page turner has been developed for the piano by MuseBook® [2] which automatically recognizes the portion of a score being played on the piano, and uses an electronic copy of the sheet music on a computer display to show the musician’s progress, and to anticipate when the display should be shifted. Given the very unique acoustic signatures of different instruments and the types of music they are designed to play, the MuseBook® product cannot be used for anything but a piano. In this project, we propose the development of a similar device for guitars using a commodity laptop computer and appropriate hardware and software for sound capture and signal processing. The device would be able to detect and recognize individual notes and chords being played on the guitar, locate the particular sequence of notes and chords within a score or guitar tablature, and display in a continuous fashion the progress of the musician through the piece being played. The amount of anticipation could be tuned to the artist’s preference and would always be adjusted to meet the tempo of the music being played. We anticipate that the project would have the following components:
1. A signal processing stage, in which the instrument’s acoustic signature is calibrated and the music is digitized, analyzed and transcribed.
2. A matching stage, in which the transcription is compared to the music stored in the machine. Note that, as a first cut, we would assume a perfect artist (meaning, no wrong notes, no skipped notes, and no inserted notes). Future versions would be robust and would allow interaction with novices.
3. A display stage, for the music to appear in a continuous stream on the screen.

Student’s Research-Related Duties and Expected Outcomes:
We anticipate the need for up to THREE students to be involved on the project, one dedicated to each of the above stages. Three faculty members will be involved to advise and mentor the students during the course of the project
Professor A. Lee Swindlehurst – Electrical Engineering and Computer Science
Professor Jean-Luc Gaudiot – Electrical Engineering and Computer Science
Professor Nohema Fernandez – Music.
Each student will gain insight in his/her respective domain of focus (signal processing, database searching and retrieval, visualization), and the strongly interdisciplinary tone of the project would train them well for team projects. Finally, the application domain (music) would prepare them for interaction with the non‐engineers who would be their end customer.

Prerequisites: All three students will need a strong understanding of C/C++, and must have taken a relevant course in computer architecture such as EECS 112. Experience with Linux, OpenMP and OpenMPI programming is also required. Some knowledge of basic signal processing theory (equivalent to the content of EECS 150AB) would be helpful, as would familiarity with music theory in general and guitar musical transcription in particular.

Recommended Web sites and publications: 
   La Tourneuse de Page (2006):

 Project #13:  Parallelization of Telemedicine Benchmark for the Xbox 360 Architecture
Faculty Mentor:  Professor Jean-Luc GaudiotElectrical Engineering & Computer Science

Description:  Telemedicine presents a popular topic in today’s medicine based on increasing each medical practitioner’s limited accessibility through high-speed networks, high-performance CPUs, and responsive electromechanical devices. Transporting a medical professional or patient to a common destination entails a high cost in time and resources. This is a strong motivation to advance research for cost-effective ways to provide medical services remotely through the help of sophisticated computer systems. With a way to utilize high-performance networks and computers, medical professionals may ideally treat patients around the globe without physically moving them. With networks and electromechanical devices sufficiently developed, the main constraints on any telemedicine program implementation would be efficiency, reliability, and performance. Thus, a strong focus should be placed on increasing the performance of today’s computers under telemedicine’s criteria of response‐time and overall processing time.

Being in the current trend of multiprocessors, performance gains in either criterion pushes us to study parallel code optimization. Being affordable and potentially quite powerful, the Xbox 360 is a great candidate for such a parallel architectural study with a focus on telemedicine applications, which is why we selected it for this project. Targeting the Xbox 360 tri-core PowerPC-based processor, this project will focus on parallel code utilization and optimization using proper thread work management and current parallel coding tools such as OpenMP and OpenMPI. Using a model telemedicine program, the Telemedicine Benchmark, the Xbox 360 architecture can be optimized for core telemedicine applications such as data encryption, data compression, and real-time services. Ultimately, this study will result in determining whether the Xbox 360 presents a viable architectural platform for telemedicine or even scientific computations in general.

Student’s Research-Related Duties and Expected Outcomes:
The student will be responsible for the optimization of the set of programs representative of any telemedicine program on Xbox 360 architecture using publicly available tools. The end result of the research should be refined versions of the benchmark suite which should run significantly quicker and/or be more reliable.
The student will develop a good sense for parallel programming as well as for the techniques used to analyze computer architectures. Also, the student will learn about current practices, tools, and theories in parallel computing.

Prerequisites: The student needs a strong understanding of C/C++. Also, the student must have taken the relevant course in computer architecture such as EECS 112. Experience with OpenMP and OpenMPI programming will be required. Also, Linux experience will be required.

Recommended Web sites and publications: 
   Jordan, Harry. Fundamentals of Parallel Processing. New Jersey: Pearson Education, Inc., 2003:
   Klutke Peter J., Mattioli Paolo, Baruffaldi Fabio, Toni Aldo, and Englmeier Karl-H., “The Telemedicine benchmark—a general tool to measure and compare the performance of video conferencing equipment in the telemedicine area.” Computer Methods and Programs in Biomedicine 60 (1999): 133-41.

   Lamport, Leslie. “How to Make a Correct Multiprocess Program Execute Correctly on a Multiprocessor.” IEEE Transactions on Computers 46 (1997): 779‐82.

   Williams Samuel, Shalf John, Oliker Leonid, Kamil Shoaib, Husbands Parry, and Yelick Katherine, “The Potential of the Cell Processor for Scientific Computing.” Proceedings of the 3rd Conference on Computing Frontiers, May 03-05, 2005, Ischia, Italy: Association for Computing Machinery, 2005. :