HARVARD GAZETTE ARCHIVES

Researchers use medical imaging software to analyze astronomical data in the AstroMed Project. The image above displays the star-forming region IC348 through the medical imaging software 3D Slicer. (Photos by Michelle Borkin)

After a year of hiring, moving into new digs, and generally getting its feet wet, the Harvard Initiative in Innovative Computing (IIC) is ready to forge ahead into the new year, diving into computer-intensive projects that need not just computational firepower, but also innovative thinking.

The initiative is moving ahead on a half-dozen or so projects generated by groups of Harvard faculty in response to a call for ideas last spring. These projects span a broad array of disciplines, from using medical imaging technology to illuminate star creation to producing astonishingly detailed pictures of the mammalian brain to designing a Web portal that offers astronomers new ways to share data.

A close-up of the star-forming region in the Perseus star-forming cloud as seen in the medical imaging software OsiriX.

"We're moving out of the planning mode and into the actual project mode," said initiative director and Astronomy Professor Alyssa Goodman.

Though the initiative has been operating for more than a year, Goodman said it has taken a year to get ready to function at the high level envisioned by project designers.

The initiative has 10 people working for it in various capacities, including consultants, research assistants, engineers, and Senior Research Fellow Felice Frankel, who this summer brought her "Envisioning Science" project to Harvard from the Massachusetts Institute of Technology (MIT).

In addition to bringing new people on board, the initiative launched a new Web site in January, moved into new offices at 60 Oxford St. in March, and continued to develop its goals and mission. A second-generation Web site, highlighting IIC's 2006-07 project portfolio, will launch in October.

IIC Director of Research Programs Tim Clark said that when research requires computational tools or analysis, faculty often draft graduate students or postdoctoral fellows to help rather than go to peer-level researchers and technologists who specialize in computational sciences, largely because there's a shortage of such people at Harvard.

The IIC, he said, exists both to bring world-class computational scientists to Harvard and to create a long-term human infrastructure of highly trained scientific programmers around them. IIC hopes to collaborate with researchers to meet their computational needs in today's data-intensive environment in a way that could not have been possible previously. A brochure distributed by the IIC states the need succinctly: "From sequencing the human genome to simulating the formation of a Solar System, researchers often face an embarrassing mismatch between how much data they can generate and how much they can understand."

But it would be a mistake, Clark said, to think of the IIC as just a place where people tend heavy-duty computers and crunch numbers. The IIC, unlike a supercomputer center, is seeking new and innovative computer-based solutions to today's research problems.

"It's more about super people than supercomputing," Clark said.

The initiative, Goodman said, has five technological focus areas dedicated to different computing needs. One of the areas concerns the visualization of data, and seeks new ways to present and understand scientific information. Frankel's envisioning science project falls into this area. Among the activities planned is a series of "Image and Meaning" workshops, which bring together scientists, graphic designers, and others interested in the visual presentation of scientific information. The cross-discipline collaborations, originally begun when Frankel was at MIT, have proven hugely successful in the past, Frankel said, and one is planned for the Harvard campus in fall 2007.

"It's an opportunity for scientists to have fresh eyes look at their work," Frankel said.

Other technological focus areas of the IIC's activities include distributed computing, which will seek new collaborative tools and computational resources to allow widely separated groups to work together in real time; databases, which will seek new ways to manage and understand the rapidly growing amounts of information gathered by today's scientific techniques; analysis and simulations, which will advance the analysis and presentation of data; and instrumentation, which will seek to develop innovative hardware, such as new data sensors, needed to support computational projects.

This past winter, the IIC put out a call for ideas, generating candidate projects submitted by a dozen teams across the University.

Several of those were selected, including a project that seeks to link social science and life science statistical and computation-management tools, a second that seeks to build a new computational framework for genetics and neuroinformatics, and one that seeks to model animal brains in incredibly fine detail. Additional projects are in the planning stages, including one modeling blood flow in the human body and another seeking ways to harvest currently untapped online data about medical treatment outcomes.

One project that got its start before the call for ideas is "Astronomical Medicine," or AstroMed. AstroMed seeks to apply sophisticated medical imaging technology to data generated by astronomers. The project's rationale is simple: Medical imaging technology is designed to generate sophisticated images from large sets of data, so why not use it on the sky as well as the body?

Though AstroMed's application of medical imaging technology to astronomical data is unique, it has already borne fruit, illustrating new features of star formation areas that were never seen before, Goodman said. Further, it illustrates the innovation and new thinking that Goodman and Clark said they expect to be the hallmark of the IIC. When the AstroMed project is complete, the software created will offer enhanced functionality in the medical imaging community, as well as new capabilities in astronomy, atmospheric science, geology, and other 3-D sciences.

While the IIC's research activities will be important to its mission, education will be another important focus. The IIC already has several students and postdoctoral fellows working on its projects and plans to involve as many young researchers as possible. It also is encouraging faculty interested in sponsoring IIC-related undergraduate theses to e-mail ideas.

IIC-related courses are already listed through the IIC's site, and many more are in the works, Goodman said. New faculty hires associated with the IIC's expansion plans will teach an increasing number of courses in data- and computationally intensive science areas. Goodman herself often offers a course called "The Art of Numbers," which teaches undergraduates how to create clear and effective displays of data, such as maps, charts, graphs, presentations, and Web sites.

For the broader public, the IIC is sponsoring a seminar series, which Goodman says already has a loyal following in just its second year. The series is kicking off this week with a talk by Guy Steele of Sun Labs. Goodman said that longer-term plans for the IIC include a museum in Allston, showcasing data and computationally intensive science, that will be open to the public.

"Today, increasing computational power allows scientists to 'see' new worlds, in much the same way that increasing resolving power in microscopes and telescopes traditionally has," Goodman said.

alvin_powell@harvard.edu

Related links:

Allston planning and consultation advance: Search is on for science building architect