Inside Intel

When Intel Corp. located a lab at Carnegie Mellon University in 2001 to host joint technology R&D projects, cynics could have been forgiven for thinking the marriage wouldn’t last. Intel, after all, is a $35 billion company, the world’s largest semiconductor manufacturer, a behemoth that could be expected to guard its intellectual property zealously.

Yet the collaboration already has achieved more than anyone could have anticipated, providing a new model for technology development even as it advances a host of cutting-edge projects.

The lab at CMU is one of three launched by Intel — the others are at the University of California and the University of Washington — to align the company with leading researchers on leading campuses. David O’Hallaron, the CMU professor who serves as director of the Intel lab here, says Carnegie Mellon was a natural fit for Intel’s vision.

“Carnegie Mellon always has had this tradition of collaboration, multidisciplinary work, working in large groups on large projects,” O’Hallaron says. “That’s a CMU hallmark. It’s actually worked out like that. As soon as the lab got on campus, immediately there were all kinds of faculty working with them.”

Historically, established tech companies wait until a start-up has proved the marketability of its product before investing in the fledgling firm, acquiring it or licensing its technology. That, however, can take years, a difficult period of growth when even promising companies can fail.

But through the Intel-CMU collaboration, the private sector is in on the ground floor, providing key financing and expertise and having a piece of dozens of potentially hot technologies. As a result, more of those projects will succeed . . . and quicker.

“Intel is interested in any exciting, groundbreaking, rule-changing ways to use computers,” O’Hallaron says. “Those kinds of changes are happening at universities, not at corporate labs, which are more focused on products and the more immediate technology road map.”

Moreover, through an open collaboration research (OCR) agreement, Intel and CMU share ownership of any technology that emerges from their teamwork. That’s a pioneering approach that has attracted other significant players. In addition to their own OCR, Intel and CMU have similar agreements with UPMC, the University of Pittsburgh and the pharmaceutical giant Merck.

“This is a unique new model for research, and that’s broken down all the walls,” O’Hallaron says. “The reaction I get is, ‘That’s too good to be true. What’s the catch?’ There are no limits to what we can do together. From the point of view of a university professor, this Intel lab is like another department on campus, a source of smart people to work with.”

To be sure, Intel reaps its share of benefits as well, harvesting good will among faculty and students.

“It’s very much in their self-interest,” O’Hallaron says. “They develop very close relationships with potential employees and the faculty members who are developing those future employees.”

The larger community should benefit as well, as evidenced by these current lab projects:

INTERACTIVE SEARCH-ASSISTED DECISION SUPPORT (ISADS). The title is clunky, the technology potentially revolutionary. Working with UPMC, Intel and CMU researchers hope to develop a tool that will enable doctors to access all images within the UPMC system to assist in their diagnoses. A physician could utilize such a search upon finding, for example, a spot on a mammogram. Explains O’Hallaron:

“The idea is, you take that mammogram and compare it to all the mammograms in your medical records that have spots, and you find those images with similar spots and see how those were diagnosed. Imagine the power of using all the images — CAT scans, MRIs, X-rays — as diagnostic tools. This stuff is just gathering dust right now. It’s inaccessible. The tool we’re working on enables that research.”

Buoyed by their progress with ISADS, the team is looking at harnessing other potentially vast information pools — “Big Data,” O’Hallaron calls them — that currently are floating free and unassembled.

“We’re looking at all kinds of Big Data applications,” O’Hallaron says. ”We’re developing the techniques that will allow us to manage it, manipulate it efficiently and store it.”

INTERNET SUSPEND/RESUME (ISR), which would enable you to access your electronic files, folders and applications from any computer, whether it’s at work, home, or anywhere else. A server would be the key intermediary, allowing you to upload and download your computer environment in short order.

“The dream would be to hit a couple keys, and your environment would pop up with your windows and applications no matter what physical machine you’re using,” O’Hallaron says. “The biggest obstacle is bandwidth and just the size of the environment.”

CMU has taken over administration of this initiative and won a National Science Foundation grant to support it.

DYNAMIC PHYSICAL RENDERING (DPR). Imagine having a conference call with far-off colleagues whose images appear right there in your board room and even track the movements and gestures your colleagues make all those miles away — and not one of you is wearing special headgear or goggles.

It’s based on a new form of programmable matter called “claytronics” that the lab is developing. The fundamental units of claytronics, known as “catoms,” would bind together to form physical analogs of virtual shapes. O’Hallaron admits that producing a virtual image of your boss from Boston is “very far out.” But with the same technology, the team hopes to create a device even more useful — a 3D fax machine.

“Let’s say you’re designing a new car, you have a model of it, and you want to fax a 3D model of that to someone,” O’Hallaron says. “A collection of catoms would be sitting in the fax machine at the other end, and they’d assemble themselves into the right shape.

“You could see the model, explore it, make a change and fax it back. That’s probably our attainable five or 10-year goal.”

---

Evan Pattak last wrote for Pop City about the Quality of Life Technology (QOLT) initiative at Carnegie Mellon University.

---

Captions

David O'Hallaron

Visualization of Boolian Satisfiability

O'Hallaron at the Intel Office Suite

Robotic Arm

Intel

All photographs copyright Brian Cohen