What does a major academic-corporate partnership look like from the inside? The director of Columbia's new high-tech consortium describes the possibilities, synergies, and challenges

Shepherding new media technology from the lab to the public

Dimitris Anastassiou

The Columbia New Media Technology Center is a cross-disciplinary partnership between Columbia and worldwide industry, working together in an integrated environment to develop next-generation advances in multimedia technology. CNMTC looks forward to seeing its research program extend the horizons of engineering research to include local businesses, social programs, and scholars in other disciplines. Meanwhile, we hope to see our growth in New York at large accompanied by expanding possibilities for partnerships with researchers and practitioners in other schools here on the Columbia campus.

Based at Columbia's Fu Foundation School of Engineering and Applied Science, but also involving components of Teachers College and the graduate schools of journalism and business, CNMTC reflects more than a decade of experience coordinating its academic and corporate components. For many years now, I have been devoting my professional life to making this program work; the experience has involved many sacrifices and many rewards. Although new media technology has certain special characteristics that are well-suited to an industrial collaboration, the CNMTC experience offers insights that may also be useful for researchers in other fields who are considering similar efforts.

Multimedia, the next generation
The first-generation multimedia technology, our enabling digital technology, has been developing mainly in California's Silicon Valley. It provides a powerful mechanism for accessing content created by someone else, but it remains difficult for the common user to handle the increasingly sophisticated technology.

The next-generation multimedia technology will build on the first generation to enhance human communication. It will empower people to become producers of multimedia content and convey complex ideas easily and efficiently, using audiovisual signals in addition to text. The emphasis will be on individuals exercising direct creative control over the production of their own content, doing their own "digital storytelling."

In the future, there will be new and powerful tools for multimedia content creation, editing, and content-based search and retrieval. People will not think of pixels, just as text processing does not require a user to think of black or white dots when choosing fonts, sizes, and formats. Similarly, we would like to enable producers of visual multimedia to think only of the meaning and style of the message they want to convey, and thus to cut and paste audio and video objects transparently. The resulting enhanced communication (e.g., the ability to "say a thousand words" using an easily produced image) will lead to increased productivity, a crucial element of economic development.

No one doubts the importance of physical or information highways for the economy. Just as it is necessary to develop efficient cars and trucks to carry cargo at low cost, it is also important to develop vehicles for information that are both sophisticated and easy to drive. The ability of consumers to communicate their content to the world rather than merely being passive recipients of content will lead to social benefits, among them enlivening democracy in a real sense. Our industrial sponsors are particularly interested in helping bring our new media technology to New York's public schools.

CNMTC pursues synergistic partnerships with social non-profit organizations as well as industry. For example, the content-based paradigm will create environments where both users and makers of networked multimedia technology can meet, allowing the broadest possible number of people to participate productively in the emerging digital age. This is the social goal of the Digital Clubhouse Network, a partner of CNMTC, which "uses the power of networked multimedia to develop communities," in the words of its mission statement. In DCN's programs, people use their own life stories, favorite music, and personal imagery as their source of multimedia content. DCN offers educational programs for people who traditionally have had relatively low exposure to high technology, including women, the disabled, and minorities.

CNMTC will provide its evolving technology to the Digital Clubhouse Network, which will offer a nationwide testbed to CNMTC, so that new technology developed at the center will undergo testing and evaluation by actual users. Our industrial sponsors tell us that this collaboration with DCN will greatly benefit the process of technology transfer in general.

Into the Alley and out to the world
Conducting new media technology research in New York City means having proximity to a rapidly emerging industry that may be quick to adopt our research results -- or to employ Columbia students. To hasten this outward movement, CNMTC has entered into partnerships with a large number of local small new media companies. New York's Silicon Alley is an ideal place for the next generation of technology to emerge, because it is fast becoming the world capital of new media businesses, founded on the belief that "content is king" and catalyzed by the explosive growth of the Internet. This shift toward content-based new media technology will also dissolve the perceived dichotomy of content vs. technology, which is said to distinguish the Alley from the Valley.

I share the expectation of many observers that in 1998, several successful high-tech start-up companies will draw attention to Silicon Alley and attract increasing venture-capital interest. To strengthen the connection of our own research program to the local new media industry, CNMTC has opened a second home at the New York Information Technology Center at 55 Broad St. This building is home to numerous new media businesses, including DCN. CNMTC thus occupies a unique role in New York's emerging new media industry.

An increasing number of experts share the vision of a content-based paradigm. For example, the European ACTS industry-university-government collaboration program (which the European Union supports with almost $1 billion annually for research in advanced communication technologies) is moving with full force in that direction. In fact, the United States risks being leapfrogged in this crucial area, despite the sound recommendations of the U.S. National Research Council. The NRC has emphasized the need to enable citizens to interact easily and effectively with the National Information Infrastructure, as described in the book More Than Screen Deep (Washington: National Academy Press, 1997).

The content-based paradigm presents CNMTC researchers with a multitude of opportunities to create technologies that the marketplace will need to adopt. For example, when refined and combined with easy user interfaces, some of the software tools we are now developing could become so widely desired that they will be bundled into the operating systems of future PCs. To facilitate this form of technology transfer, Columbia Innovation Enterprise will have a prominent position at CNMTC's site at 55 Broad Street.

Reality checks and win-win scenarios
Our industrial partners help us ensure that our research is relevant. They tell us when they believe a research area has already served its purpose and it is time to move on to new areas, as well as what these areas might be. All too often, a group of bright academic researchers from prestigious universities, focused in a specific research area, keep accepting each other's scientific papers, attracting other university researchers to the same area. While these papers may be intellectually excellent, unfortunately practical significance does not always correlate with intellectual excellence. Several such self-perpetuating research areas may suddenly collapse.

Industry also tells us which aspects of our research have the greatest potential for technology transfer and commercialization of university intellectual property. This information is extremely helpful for faculty and students, so that the effort devoted to writing invention disclosures and working with attorneys to write patent applications represents time well spent. If an application is successful, the feeling of having contributed to a widely used technical invention, as we have done with the MPEG-2 digital video coding standard, is deeply rewarding, above and beyond the financial benefits to the university and the inventors.

As part of a culture dedicated to industrially relevant research, my colleagues and I often participate in international standardization efforts. The key here is to anticipate the standardization process and to research the challenging issues before that process is initiated. Our students participate actively in collaborations with our industrial partners and typically end up being hired by these firms, often after the firms compete to make the most attractive offer. Our industrial partners tell us that they consider well-trained students to be one of the most important benefits derived from sponsorship of our center.

Sponsorship by a large number of industrial partners with a common vision of next-generation technology has a key advantage: continuity of funding. While government support, however large, is usually limited to a few years, a strong industrial affiliates program typically gains and loses some sponsors each year in an unending cycle. The industrial sponsorship can help attract government funding as well, because many public-sector grants are subject to obtaining matching funds from industrial sources.

The biggest challenge is to structure the program so that it has a "win-win" element for all parties involved: the university, the industrial sponsors, the participating faculty and students, and the center's management team. The program must not be a set of independent research projects, each directed by a sponsoring company that supports, for example, a graduate student and part of a faculty member's time. This approach is bound to quickly drain all the center's resources. Instead, the management should identify a set of important shared benefits to many industrial partners, so that the work of each student and faculty member simultaneously serves the needs of many sponsors. This, of course, is easier said than done.

Another related problem is the time commitment required to make such a program successful. It is extremely difficult to make many industrial sponsors happy, and my experience suggests that it is na‹ve to believe that hiring any technical person to be responsible for this will solve the problem. Unavoidably, taking responsibility for a program means sacrificing much of what a professor is supposed to do. To make things worse, because a successful industrial affiliates program is not common in a university environment, the time spent on it is often unappreciated. Therefore, I think that the only way such a program can remain healthy is its designation as a center, with support from federal grants.

Another challenge involves the internal contradiction between seeking industrial sponsorship for university research and pursuing profit from commercialization of university inventions. The problem is that industrially relevant university inventions almost always come from industrial research sponsorship, but companies will not sponsor a university industrial affiliates program unless they have significant rights on intellectual property coming from their sponsored research. A solution must be found balancing the arguments of both sides. My experience suggests that the best answer is to have both sides cooperate rather than compete. This comes about if the university emphasizes research sponsorship by industrial partners as the highest priority, even if this implies giving up significant intellectual property rights to the sponsors.

Increased research sponsorship results in invaluable benefits to the university: attracting better faculty and students, enhancing the institution's reputation, and catalyzing further research sponsorship by government agencies. These factors will help the university develop more relevant inventions in the future. Therefore, I believe that a cooperative approach in the long run will paradoxically lead to much higher returns to the university from intellectual property, compared with an overly tough negotiating stance. Like components of any network, an academic-industrial center is most productive when it enhances its connectivity.


Related links:

  • Pamela Mendels, "Old Ivy Meets New Media," New York Times, Nov. 24, 1997

  • Randy Whitestone, "New York's Silicon Alley continues to grow," ZDNet News, Oct. 25, 1997
    DIMITRIS ANASTASSIOU is professor of electrical engineering at Columbia and director of the Columbia New Media Technology Center.