Through the big windows of a corner office could be seen the bobbing masts of Marina del Rey. Against this tranquil backdrop, Herb Schorr was explaining to me how he planned to grow his unusual waterside outpost of the University of Southern California into a significant player in the world of corporate research.
“The rules of the research game are changing,” he said. “And we want to change with it.”
Schorr, 70, a native New Yorker with a background in research management at IBM and a taste for modern art -- he was among the first collectors of Jean-Michel Basquiat -- became director of USC’s Information Sciences Institute in 1988. Since then, it has grown into a nearly $70-million enterprise funded mostly by federal grants. But more growth opportunities in research and development are on the corporate side, where ISI hopes to fill the gap between the basic research customarily performed in academia and the product development usually handled by industry.
By positioning ISI as a provider of private research in its core fields of computer science, artificial intelligence and information technology, Schorr aims to build corporate funding to as much as a third of ISI’s revenue, up from less than 20% today. He believes that strategy will enable ISI to grow at a real rate of 4% a year. (“It’s a slow business, not Google,” he cautions.)
Several corporate projects already rank as showcase ventures at ISI. One is a collaboration with Chevron Corp. to develop real-time sensors to send information from oil fields to off-site controllers -- a project Chevron touts as the creation of “smart” oil fields.
The institute has also spun off at least a half-dozen projects as commercial companies; the best-known of these is Language Weaver, whose translation software is regarded as among the best in that infant field.
ISI was founded in 1972 by the late Keith Uncapher, a Rand Corp. executive seeking a way for the university to attract grants from the Defense Department, which was beginning to invest more in civilian, rather than strictly military, applications. Early grants from the Pentagon’s Advanced Research Projects Agency, or ARPA, gave ISI a front-row seat at the creation of the Internet, originally an ARPA project.
Indeed, engineers and computer scientists probably know the institute best as the professional home of Web pioneer Jonathan Postel, who helped develop the Web’s domain name system. This is the scheme that transforms every website’s Internet address -- a series of four numbers separated by periods -- into the plain-language .com and .edu addresses employed by Web surfers today. Postel held the job of director of the Internet Assigned Numbers Authority, which governed that system, until his death in 1998.
Government funding has remained the mainstay of ISI’s budget, but Schorr’s models are institutions such as SRI International, an independent lab that shed its formal ties with Stanford University in 1970, and BBN Technologies, which was founded by two MIT professors in 1948, absorbed into Verizon Communications in 2001 and spun off by Verizon in 2004. Both are considerably larger than ISI, but there’s no indication that they’ve cornered the market in independent research, especially for a lab with good university connections.
For years, industry has been pulling back from basic and mid-stage R&D;, investing chiefly in the commercialization of ready products. In the mid-1990s, the cult of shareholder value made the corporate funding of research centers filled with Nobel laureates pondering deep subjects appear spendthrift and naive, especially given the difficulties faced by big companies trying to exploit the serendipitous output of such think tanks. (Xerox Corp.'s failure to market the personal computer hardware and software invented at its Palo Alto Research Center, the legendary PARC, is the most prominent negative lesson.)
Their solution was, in effect, to outsource more research to entrepreneurial start-ups, universities and independent labs. The upside for researchers was that jobs at institutions with multiple income sources were less vulnerable to corporate downsizings in lean times.
Today ISI’s relationship with USC is a loose one. Its 430 staff members are formally employees of the university, but they have little or no campus duties. Among its PhDs are 28 designated as “research professors” at USC’s Viterbi School of Engineering, meaning they can supervise the doctoral work of graduate students but aren’t eligible for tenure or required to carry a full teaching load.
The institute doesn’t receive any funding from the university -- in fact it contributes (modestly) to the USC budget. But it regards its access to on-campus talent as a selling point to government and corporate clients. USC similarly sees ISI as a source of interesting projects for campus faculty and grad students alike.
Former Viterbi Dean C.L. Max Nikias strengthened the bond by increasing the number of joint grant applications and encouraging ISI to reach out to departments outside its core competencies of computer science and electrical engineering.
“I called ISI the hidden jewel of USC,” says Nikias, who is now the USC provost. “The only thing I had to do was polish it.”
But that raises a question about how USC will manage the tensions that might arise between its academic mission and the commercial preoccupations of corporate grant makers.
“Industry research fits really well with academia,” says David Patterson, a UC Berkeley professor of computer science and president of the Assn. for Computing Machinery. “But development causes problems with universities as it becomes more secret and proprietary. And there’s more money for development than for research.”
Schorr notes that ISI already works with some corporations under non-disclosure agreements, and argues that the institute has learned how to manage such relationships. Nikias is similarly sanguine.
“Most companies understand that there are limits when they have a relationship with a university,” he says. “I’ve found that if you really spell out all the terms of the relationship at the beginning, then you avoid surprises in the future.”
Golden State appears every Monday and Thursday. You
can reach Michael Hiltzik at email@example.com and view his weblog at latimes.com