Statement of Capability

A team developed under the leadership of SAIC and OntologyStream

July 25, 2003

A group of scientists has gathered to define and demonstrate a new class of capabilities. This new class of capabilities will generate measurably high-fidelity ontology from free text and human interventions. While we believe that this is a powerful capability, we don't want to argue this point. It is necessary to clear the air of a critic issue regarding fairness and diversity in the evaluation of capabilities of this type.

We claim simply that the capability we are able to define and deliver is a different capability than what exists anywhere as an intelligence technology. This fact makes it worthy of government interest and funding.

We are making a two-level argument:

· Development in this area of intelligence processing is crucial but is stymied due to institutional closure. Progress of any importance is unlikely to occur without a break with the current procurement pattern, a break that embraces true variety.

· The difference we are offering is large and potentially disrupting. We are working from a base of relevant scientific findings, competent scientists, and effective technology, all of which is virtually unknown in the intelligence field and by the consultants that the intelligence community depends on for procurement advise regarding these capabilities.

The institutional blockage, while not universally recognized, is no secret among those who maintain a broad perspective on the field. Anthony Cordesman (CSIS) pointed out in his recent (July 24^th) testimony to the House Intelligence Committee that current intelligence technology and methods are not consistent with the uncertainty and deception that characterize the threats. With today's state of the art, he says, there is no way that one can develop high fidelity readings concerning weapons of mass destruction. David Alberts (OSD) is arguing, further, that a disruptive transformation of knowledge sharing capability is necessary.

Even those who are mired in the system feel a responsibility to try to break out, but this has been difficult. A most interesting example was the recent NIMA competition intended to select 'innovative' approaches to intelligence handling in various areas. Over 300 research teams submitted proposals to this prestigious program. The first result of interest is that the vast majority of proposals were deemed not innovative.

It might appear to some observers that the people who remain willing to submit proposals to such a program, after years of plugging away on the same ideas, simply have nothing new to say. But even if they did have sometimes to say, they are invested in programs that they would rather continue. From the point of view of regular participants in this community, deviation has little value. The second result of interest is that, while a few proposals were deemed innovative, members of the review committee shamed each other into selecting only the 'best', meaning the ones most could agree to.

The result was that the majority of funding was placed on a 20-year old project that has never come out of the lab! This project's distinctive feature is that it is endless. As long as you entertain the unsupported assertion that a machine with a load of facts will somehow, some day, yield intelligence; you can kid yourself that you are funding innovation. In private remarks and at a conference, on the value of funding Cycorp, John Sowa’s remark on this project is that it needed to be funded, but enough is enough.

While we are expressing a frustration shared by many, we don't want to go overboard and claim that all current work is useless. Distinctively different paths do need to be pursued, and it is good that the current paths have been pursued. We are simply saying that a reasonable scientific strategy would be to taper off from new paths that have had their chance, and that multiple paths should be started and given their chance. But we find that natural science, more than not being in charge, is not even consulted. The communities organized around current paths in intelligence are closed. Others cannot break in, and thus have no financial means to work on the problems.

We are not claiming that it is easy for scientific communities to maintain the variety that is a necessary correlate of innovation. For all the benefits of NIH's 'consensus' committees, it is generally conceded that innovation is suppressed in medical research. It is the judgment of some that many years of work on cancer, for example, have been mostly wasted. Fundamental breakthroughs appear to be coming from outside the mainstream funded communities.

Likewise, the field of mathematics is notoriously unable to recognize innovation, but because this tendency is so well know, normalizing compensations have been installed in the institutions of the field. Some mathematics journals have a rule that, if just one reviewer finds a submission interesting, it is published. Some of the more popular submissions may also be innovative, but they are no more likely to be so than the less popular submissions. Recently experiences with IEEE conferences have show, to some of us, that these conferences are little more that pay and publish activities.

Having said all of that, and that is a lot, it is true that innovative thinking, by definition, will not be broadly understood at first. Yet it is important to invest in it.

To summarize, in the procurement of new intelligence technology research there is no variety maintenance function, except at the margins within a very closed community. Those few within the community who are able to recognize the situation or want to change it are helpless to do anything about it.

As Paul Prueitt has suggested, if intelligence processing didn't matter, scientists could leave it this way and turn to other problems. But intelligence does matter, and might be the one greatest opportunity to pursue effective warfare that avoids collateral damage -- not just to lives and property, but to human rights, privacy, free speech, and other values that we are attempting to protect.

Given the stakes, what is needed is an extraordinary and disruptive intervention to create variety.

We would like to offer our scientific program and community as one genuine alternative worthy of selection. Again, we do not stake our worthiness on direct comparison with existing programs under existing criteria. We are worthy because the difference is distinctive.

We were recognized as such in the NIMA competition. We were included in the small group of fundable projects, and thus must have been of interest to at least one reviewer, but judging from the comments, we were far from a popular choice. An additional interesting fact, attesting to our difference, is that we competed under a category that few others chose, one that NIMA had acknowledged was misunderstood by most teams.

We are not saying that obscurity, by itself, is a qualification for funding considerations. We are working, as we said, from a base of scientific work that does have a following, just not a following within the conventional intelligence community. We also have plenty of technology that is consistent with this science and shows promising results, though again this technology has not broken into the intelligence community. The technology is either on the very leading edge of commercial adoption, such as technology we have evaluated by SchemaLogic Inc, Entrieva Inc, and Recommind Inc, or is beyond what has been productized while being grounded in fundamental mathematics, cognitive and social science, and logic.

In the following months and years, we will continue to develop the specifics of how people and technologies work together and how this new class of capabilities is to be applied to intelligence problems. All we hope to indicate that there is a difference, that the intelligence community has not given this difference a try, and that the persons involved are credible scientists and technologists who believe they can make an innovative contribution to the nation's defense.

Bringing the group together in a series of workshops is the next reasonable step that we request the Office of the Secretary of Defense consider.

( Index )