Summer Projects for 2009
Robotics and computer vision (Funding pending, 1-2 students)
These students will work alongside the REU students. See the REU robotics description or website for more details. Directed by Prof. Dodds.
“Serious” Games (funding is pending, 2-3 students
We will focus on games for seniors using the wii platform and on educational games on cell phones. Directed by Prof. Sweedyk.
Smarter Educational Software for Sketch recognition (NSF Funded, ~2-3 positions)
Research on this project began in the summer of 2006 and will continue for the next few years. I have funding to hire students in the summer and I am happy to work with students in independent study projects during the year.
The goal of this project is to build a sketch-based simulation system for digital design, to be used in E85. In 2008 we made some exciting progress on our recognition algorithms. This summer we will continue to improve recognition algorithms and explore user interface issues. The goal is to have a fully functional prototype, ready for use in pilot studies or in the classroom by fall 2009. Directed by Prof. Alvarado.
Restrictions: I will be on leave for part of the summer, so I am looking for highly independent students or students who would like to work during only the first half of the summer. Some experience is required.
Recognizing Military Course of Action Diagrams (DARPA funding pending, 2 positions)
The goal of this project is to recognize freely-drawn military course of action diagrams to support military commanders during battle planning. This project will focus mainly on sketch recognition algorithms, but may also include a user interface component. The unique and exciting part of this project is that we aim to achieve a working system with high recognition accuracy (>90%) on freely drawn data. Students will work closely with my colleague Tracy Hammond and her grad students at Texas A&M University. Directed by Prof. Alvarado.
Restrictions: You must be a US Citizen to work on this project. I will be on leave for part of the summer, so I am looking for highly independent students to work on this project. Some experience is required.
Trace Repository (Funding pending, 1-2 students
The SNIA Trace Repository contains over a terabyte of data collected by observing the behavior of real file systems. Harvey Mudd is responsible for the management and enhancement of this repository. Students will work to improve the user interface, develop tools related to traces, and integrate tools contributed by researchers, etc. Directed by Prof. Kuenning.
Summer Staff (Funded, 3-4 students
The computing staff helps system administrator Tim Buchheim run the Computer Science Department's servers, networks, and labs. Summer is the time for big changes and major tasks. Summer is also the time to learn about the CS Department's systems; prior sys. admin. knowledge is appreciated, but not required. Directed by Prof. Kuenning.
Observationally Cooperative Multitasking (funding pending, 2-3 students
Programmers face a multi-core future but no one knows how to take advantage of it. Current techniques for concurrency tend to be primitive, complicated, and error-prone. This project will investigate and implement OCM, a new approach to writing concurrent code without the hassle and bugs of locks. Programmers write code as if each thread has the machine to itself until it explicitly yields control; under the hood, the system detects noninteracting threads and runs them simultaneously. Directed by Profs. O'Neill and Stone.
Work on CSL (Checkable Sequencing Language) (JPL funding pending, 1 position)
CSL is a language under development at JPL, initiated by Profs. Stone and Keller. It involves modeling spacecraft components and commands, for purposes of mission validation. CSL provides a Java-like front-end that dually translates into the primary JPL legacy language and into models for the model-checker Uppaal. The latter enables validation of time-dependent concurrent systems. The work would involve implementation and testing aspects of the language and its interfaces, and possibly working with modeling engineers at JPL to determine additional features. Having taken programming languages and compilers would be a definite plus. Directed by Prof. Keller.
Restrictions: You must be a US Citizen to work on this project.
REU Program Projects
The REU program will take up to 5 students total from HMC but the exact distribution of students among the three projects will be determined after students apply. See our REU website for details.
To participate in the REU, you most be a US citizen or permanent resident, and you must also still be an undergraduate in the fall, so seniors are not eligible.
Intelligent Music Software (Funded, 2-3 students; will also consider non-REU students)
This project involves researching techniques for educational software tools to help students learn to understand music. Currently we have been concentrating on jazz improvisation, where our approach is to aid the student in constructing melodies similar to ones that could be improvised. This year, we will focus on inferring the deeper structure of music using machine-learning techniques, as well as real-time performance aspects. In addition to several papers, the output of this project has been three major versions of the Impro-Visor tool, which is distributed free and has over 3500 users. Participants should be interested in music knowledge representation and retrieval, machine-learning of musical knowledge, real-time execution of music, and real-time human-computer interfaces. See impro-visor.com and our REU website. Directed by Prof. Keller.
Better Garbage Collection (Funded, 3-4 students)
If you like programming languages and operating systems, you'll probably be interested in garbage collection. See the description on our REU web site for details. Directed by Prof. O'Neill.
Computer Vision and Robotics (Funded, 2 students
This project will aim to create a team of robots that can participate in July's Student Robotics Competition, part of the IJCAI conference to be held in Pasadena. We will balance the work between continuing the visual mapping/spatial reasoning done by students in 2008 with new approaches, most likely borrowed from the field of machine learning. The project page on our REU website has more detail. Directed by Prof. Dodds.
Algorithms for Computational Biology
This project will investigate algorithm design and analysis for problems in computational biology. This research has a significant theoretical component (analyzing the computational complexity of problems and developing algorithms and heuristics) as well as some system-building and testing (implementing our algorithms and testing them against existing methods currently used by biologists). We will work closely with biologists throughout the summer. For more details, see the project page on the REU website.
