HP and MIT: Redefining industrial-academic research

In 2000 HP and MIT embarked on a bold 5-year, $25 million experiment to engage in a far-ranging, forward-looking alliance "to create the future of digital information systems."

Three years later, the alliance encompasses a broad range of collaborative projects in wireless networking, pervasive computing, molecular inkjet printing, quantum computing, and digital asset management.

Both sides agree the experiment is paying off and is well on its way to setting a new standard for industrial-academic partnerships.

One high-profile project that has already enjoyed considerable success is DSpace, a groundbreaking digital library system.

Developed by HP Labs and MIT Libraries, DSpace is a free, open source repository for digital intellectual output aimed at ensuring broad access to iimportant scholarship materials far into the future – not just paper or books, but also data files, images, collections of field notes, audio and video clips. Much of that data exists on personal computers,Web sites and on departmental servers. That may be acceptable now, but there's no guarantee those materials will still be accessible as file formats change, as people leave the university or even if someone's hard drive crashes.

Interest in the technology has been high. During the year that DSpace has been in production at MIT, more than a dozen other research institutions have installed it, and another 150 are now evaluating the system.

Another cornerstone of the alliance is Project Oxygen. HP is one of the five founding partners of this human-centered pervasive computing project hosted by CSAIL (MIT's Computer Science and Artificial Intelligence Laboratory).

Compaq (since acquired by HP) played a role too: The mobile Oxygen devices are built around the iPAQ pocket PC. As wireless-enabled products emerge from Oxygen, no company should be better prepared to market them than HP.

According to HP Labs Director and Senior Vice President Dick Lampman, the alliance has been a valuable addition to HP's research efforts.

"The best barometer of real progress is that people are energized.," he says. "The thing that really makes it work is the engagement of the individual scientists and engineers in both communities. Both teams stimulate each other."

On the other side of this coast-to-coast collaboration, the feeling is mutual.

"The HP researchers are of a high quality, and they bring a valuable industrial perspective. One of the key things is that we have a real engagement with these researchers," says MIT Dean of Engineering Tom Magnanti, who co-chairs the alliance steering committee with Lampman. "That's hard to do in [industrial partnerships]. It's a role model for engagements with other industries."

The close cooperation between the two organizations was no accident; there was a conscious effort to forge a new, more intimate type of partnership. Teams would be formed "by invitation," and there would be a total blurring of boundaries to ensure a full exchange of ideas and information. Cultural, organizational, or legal barriers were pushed to the background.

"We were really trying to set up a more ambitious agenda and set this to a new level," says Lampman. "In most of our other engagements with universities, it has been driven by a specific project objective.When you have that model, you run the risk that once a project runs its natural lifecycle, the level of engagement will fall off. We were looking for a partnership that would outlive any individual project."

Although HP and MIT are involved with other multi-million dollar industrial-academic contracts, this particular alliance goes far beyond mere investment.

"This is a major strategic alliance," says Magnanti. "Many of our other industrial arrangements are of much smaller scale. With HP, there's not just one lab involved and there are longer-term projects."

One reason the alliance has worked so well is that an effort is made to foster regular travel between research centers and encourage sabbaticals, internships, workshops, and focus groups.

"Having people at MIT regularly helps us to find a connection point," says Lampman. Another advantage is that many of the researchers on both sides of the SFO-to-Logan divide speak a common language.

"The single most important thing for a successful collaboration is a sense of chemistry," says Gregory Wornell, the EECS (Electrical Engineering and Computer Science) professor who, together with HP's Susie Wee, is heading up one of the marquee efforts of the alliance -- the $750,000 per year HP Wireless Networking Invent Center@MIT.

"A lot of the people on the HP side who helped make this happen came out of MIT a few years ago," says Wornell. :"They haven't forgotten the needs of students and the nature of faculty-student interaction."

Although based at MIT, where six MIT faculty and 25 students participate, the wireless networking project also involves nine researchers from HP, some of whom bring to the table expertise from recent collaborations with NTT DoCoMo in developing wireless streaming media technologies. Some of the connections are forged even closer to home, says Wornell, who reports that the alliance has also brought MIT researchers together with each other.

"This was the first time we've ever had a multi-laboratory collaboration with wireless" he says (The project integrates research from CSAIL, Research Laboratory of Electronics, Laboratory for Information and Decision Systems, and Microsystems Technology Laboratories). In addition, says Wornell, each January the alliance's "researchers in residence" program permits a group of graduate students to spend two weeks in Palo Alto to work with their HP counterparts.

"What's unique is that it's not an internship program," says Wornell. "They're going there for the industrial research interaction; it sets a different tone. They all stay in an apartment complex together, so it's a chance for them to get to know one another, and they come back with a stronger sense of community."

Partnership benefits both sides

As with any good partnership, each side brings something unique to the table.

'There are certain things that are hard to do on campus which you can do in an industrial research setting, especially in developing technology demonstration platforms," says Wornell. "Those are very time- and labor-intensive projects that require a lot of expertise, and we've been able to make use of those at HP."

As for what HP researchers gain from their experiences at MIT, Wornell points to the future. "There are things that happen on campuses that provide a window to what the future may hold, new concepts emerging out of advanced research," says Wornell. "They're on the periphery now, but some of those will migrate to the bulls-eye."

Flexible structure makes change easy

Because it's not tied down to one particular technology, the HP-MIT Alliance also enjoys the luxury of adaptability. Over the last few years, some of the initial collaborations have dropped by the wayside, including projects on telecommunications convergence, speech recognition, and ebusiness.

Although these are still valuable research areas, says Lampman, sometimes the two sides can't "converge" on an approach.When that happens, he says, it's often better to simply move on to something else rather than waste time hammering out a compromise.

"The question is, can we align the objectives?" says Lampman. Discarded projects have been quickly replaced with new ones, and there has also been increased funding for particularly promising projects. "We are looking to develop focal points," says Lampman.

New project in information management

One of the newer alliance projects is an addon program for DSpace called SIMILE (Semantic Interoperability of Metadata in onLine Environments). SIMILE uses the DSpace platform as a testbed to explore information management and collaboration using the World Wide WebConsortium's metadata standards, including the Semantic Web.

"DSpsace was a very strong opportunity and has spread quite widely, so SIMILE emerged as a logical follow on," says Lampman.

Another new project is the $1 million/year collaboration in printed electronics research with HP Labs and HP's Imaging and Printing Technology Platforms Group located in Corvallis, Oregon. Working with MIT researchers at EECS, the Media Lab, and the Microsystems Technology Lab, the group is exploring "microtechnologies, printing electronics and how you might use ink jet technology for other areas besides printing," says Magnanti. "What's significant about this is that we've moved from the research area into one of the product divisions."

Layers without lawyers

According to Lampman, the ability to react to new developments such as this so quickly rests upon the legal and contractual work hammered out up front in the early days of the alliance.

"Having that infrastructure in place helps a lot," he says. "It made it easy for [the Printing Technology] division to move quickly. Without that, the level of process and delay that it would take would be a real inhibitor." The core of these agreements rests upon a commitment to open standards and an intellectual property contract that specifies license-and royalty-fee-free, non-exclusive use of project IP and the right to negotiate exclusive licenses.

"IP is a very complex issue in every [industrial-academic] interaction," explains Lampman. "By spending the time up front, it dramatically reduces friction and delays.What we move away from is a situation in which every single engagement becomes a new experience with contracts, legal issues, and IP issues."

Future-looking research

Although some of the research generated by the HP-MIT Alliance may appear in products within a few years, other investments reflect the long-term thinking of a major computer company that has been around for over 60 years and expects to be around for another 60. For example, although hot wireless gadgets may emerge from the Oxygen research over the next few years, the main point is to re-imagine the computing landscape a decade or so in the future.

Another cutting-edge joint project is the Nano-assembled Quantum Computer collaboration, in which scientists such as HP's Stan Williams and MIT's Neil Gershenfeld of the Media Labs' Center for Bits and Atoms are developing a new type of large-scale quantum computer.

The design improves upon previous designs by combining liquid-state NMR (nuclear magnetic resonance) and solid-state semiconductor approaches.

New opportunities

As the two entities look forward to the likelihood of renegotiating their contract for another five years, HP is also looking beyond CSAIL and other computing-oriented MIT labs to exploit the full range of what MIT has to offer.

"Opportunities continue to emerge," says Lampman. "One of the ones where we're increasing our investment [at HP Labs] is in the life-sciences area, and obviously MIT would be a terrific partner in that."

The fact that HP is so committed to such future-looking research says something not only about HP, but about the unique nature of the alliance. "A lot of companies pay lip service to idea of long term research, but when it comes down to it they don't have the stomach for what it takes,” says Wornell. “It takes a long time to mature, and it takes a sense of vision to know where you're pointing."

In the meantime, researchers on both sides of the alliance are looking for a host of new "disruptive" technologies to emerge from the arrangement.

"Strategic funding like this helps people get together who wouldn't normally cross paths," says Wornell. "Out of that come all sorts of great ideas."

Adds Lampman, "Both sides have a big stake in finding this digital future, and working together substantially increases our chances of success."

by Eric S. Brown