Home > News & Events > MITRE Publications > The Edge >

Infoglut is here. People are coping and modifying their information consumption behavior in many ways, both healthy and unhealthy. The problem of too much information seems paradoxical. Life should be better with more information, but how do we handle the overload?

One remedy is to ignore all the new information and go on as if life has not changed. This is, of course, myopic and not even effective, because even traditional information sources (such as books) are being created at a much faster rate than in the past.

Another remedy is to try to read everything. This method has its own obvious pitfalls, such as the tendency to freeze out action. The most popular method, and probably the smartest, is to become a picky eater. How do people typically become picky and find high-quality information? By limiting information to trusted sources: friends, fellow project team members, proven sources, library liaisons, supervisors, and mentors.

Locating high-quality information in this collaborative way seems like a good strategy, but how do these interactions take place? They usually happen in a haphazard way, such as, "That article sounds great, drop it off when you're finished" or "Send me that URL (Universal Resource Locator) when you get a chance." The problem is that this method limits your information horizon to the people close to you.

MITRE, under the Knowledge Management Systems research project, has developed a tool called KEAN (Knowledge Exchange & Annotation eNgine), which automates and facilitates the search for quality information. It provides filters that rapidly drill down to high-quality information in the manner most comfortable to individual users. For example, KEAN can handle all of the following queries:

• What information does Peter think is good?
• What information do people on my team find worthwhile?
• What does everyone think is worth reading?
• What have I found to be useful in the past?

So, if your favorite source of trusted information is your librarian, checking on the information he or she has found to be of high quality would be the query you use most often. If you wanted to see what information everyone thought was good, it is just a few clicks away. KEAN also supports temporal, classification, and keyword filters that help find useful and relevant information more quickly. So, complex queries such as, "Show me all high-quality information on collaborative computing in the past two weeks," are supported. In addition, an important value added by KEAN through its support of annotations is that you can not only find information, but you can see why that information was useful to each individual. The picture shows a screenshot from a KEAN query.

KEAN uses the URL as the pointer to all information. This limits the sources KEAN can filter to those available on the web, but the assumption is that almost all content in the future will be web accessible. KEAN gathers and stores both quality and classification information at the time someone reads the information at a particular URL. A pop-up Java submission form gathers as much information as the reader is willing to give. Incentives for entering information into the tool come from many sources. Knowledge stewards who are MITRE experts on a certain topic are paid to provide fresh information and review submitted information. Individuals find KEAN useful as a better URL (bookmark) management tool, providing classification (no more reorganizing bookmarks into folders), a personal annotation, temporal access ("I want to see just my recent bookmarks"), and ease of transfer (no more cutting and pasting URLs into e-mail). Research has shown that a collaborative tool will not be used unless it is first useful to the individual. There are also corporate policies and initiatives that encourage information sharing.

The classification, date, submitter, keyword, annotation, and quality information filtered by KEAN are all examples of metadata (data about data; in this case, information about a URL). KEAN works its magic by running on top of a metadata engine powered by a relational database. The metadata engine is flexible enough so KEAN users can add future, unanticipated metadata to the filters provided. For example, a KEAN user may want to add filters on the power of information (has this information been useful and changed how you work?) or the accessibility (what level of knowledge is needed to understand the information?).

In the future, KEAN will run on top of a distributed object architecture that will allow much greater flexibility in how information can be filtered, accessed, processed, and viewed. If you don't like how KEAN displays information or orders the information displayed, you can build a custom view by simply tweaking a few display parameters in a component or by building your own custom component that operates on the same data.

Instead of a single place for gathering information and metadata, the architecture will support multiple components in higher leverage locations, such as one already built that plugs into Microsoft Word. The Word plug-in component makes it very easy to share information and also provides a reminder to share by popping up on file-save operations. Custom agent components that learn your personal preference over time could monitor metadata and information submitted, forwarding the information that matches your preference. Existing search engines could plug into or filter the metadata in the database.

KEAN is just one example of an information management and collaborative tool that MITRE has developed. In the Knowledge Management Systems research project and other projects, MITRE is working towards developing a complete knowledge management architecture that supports and accelerates both the individual and organizational learning process.

---

For more information, please contact Daryl Morey using the employee directory.

---