These days, it's a cliché to say that the Internet has utterly changed the way we deal with information. The resources it has made accessible to us through a few keystrokes seem infinite. While this power to access information offers enormous benefits, it also comes with inherent problems. Chief among those drawbacks is information overload.

“It's almost axiomatic to say we are on the verge of knowledge overload,” says E. J. Grabert, program manager for the University of Illinois Technology Research, Education and Commercialization Center (TRECC)[http://www.trecc.org]. Grabert observes that, with the increase in volume of available information in recent decades the ways to put that information to practical use have also drastically increased.

Tim Wentling, who leads the Knowledge and Learning System Group (KLSG) [http://learning.ncsa.uiuc.edu/] at NCSA, agrees. “We have access to everything in the world, almost, but we can't read it all, we can't use it all,” he says. “When we make decisions, we need help figuring out what information is useful to us, and the increase in knowledge is just making that process more complex all the time.” And decision-making, says Wentling, isn't the only, or even the biggest, problem: The sheer amount of available information is overwhelming. “We have physical limits, and we can only read so much. We can only go to so many presentations. We can only sit in so many meetings.”

Wentling and his group are thinking hard about how to help humans cope with knowledge overload. They envision a system they've dubbed the “extensible brain” that could help people to more efficiently absorb and process information using a variety of visualization, data-gathering, and other developing technologies. How this system might work depends on research and input from a broad range of disciplines, from cognitive psychology to data management to artificial intelligence to linguistics. “We need to get people from these fields together and pick their brains to find out what can be done,” says Wentling.

Funded by TRECC, the project is a continuation of the research involved in developing TRECC's Enterpreneurs' Knowledge Center [http://kc.ncsa.uiuc.edu:8080/trecc//], which integrates communication tools such as online forums and text chat with a searchable knowledge base and an online learning system to provide entrepreneurs with a variety of online collaborative resources. TRECC's Educators' Knowledge Center [http://kc.ncsa.uiuc.edu:8080/education/], similar in structure but geared for K-12 educators, made its debut in September 2004.

The extensible brain project is still in its early planning stages, but the KLSG has come up with a number of ideas about what kinds of processes it might involve. One such technology would involve what Wentling calls “proactive computing,” in which a virtual “personal knowledge agent” (PKA)–a little like that ubiquitous paper clip in a well-known word-processing program, but more helpful–anticipates what a user might need for a project based on that user's computing behavior, and performs needed tasks before they are even requested.

“Suppose you're making notes on our new project,” says Wentling. “You're using some terms like 'extensible brain' and 'proactive computing' and 'advanced search strategies.' The PKA would sense what you were doing and do its own search in the background. When it completed, it would pop up a little window that says, 'I've just finished the search and have identified 14 articles that might be relevant to what you're doing; would you like to see the titles?' And you say yes, and it displays them to you.”

But why stop at making this information available via laptop or desktop? “You're not always going to be sitting in front of a computer,” says Steve Downey, the project's technical lead. “You could have your information delivered to a PDA or a phone. We want to be able to deliver that information to different devices and even allow you to query your agent from remote locations.” Downey and Wentling compare this kind of functionality to the satellite- and GPS-enabled voice-enabled information and safety services now integrated into some late-model vehicles. However, their system, says Wentling, might go beyond allowing a driver to call for roadside assistance or order a sandwich at a restaurant at the next interstate exit. “Maybe while I'm on my way to pick up that sandwich, this device will do high-speed reading aloud to me, so I save time just by driving the car,” Wentling suggests.

And then there's the semantic web, the idea of a dense, global-scale network of machine-readable information that computer scientists are working to make a reality. Personal knowledge agents would use the semantic web to solve sophisticated problems and complete complex knowledge tasks. “These agents would go out and find information and bring it back to us,” says Downey. “They'd go out and talk to the agents of other people and cleverly solve problems themselves.”

A personal knowledge agent may help considerably in winnowing out all but the most useful information. However, in the age of information overload, it doesn't change the fact that the abilities of human beings to process large amounts of information have an upper limit. How to help humans absorb all this knowledge?

High-speed reading–or compressed speech–is a possibility the KLSG is considering. “We know that humans can hear three times faster than we can speak–you can process audio information much faster,” says Wentling. Combining this understanding with reliable text-to-speech technology that can speed up audio without making the speaker sound like a cartoon animal could also increase knowledge-processing efficiency.

And representing information by means other than text is also something the KLSG wants to explore. Visualizing knowledge may be more efficient–for example, a clickable map of a region to display information about demographics, exports, or travel.

According to Wentling and other members of the KLSG, a lot of the technologies that could be useful for a project like this, including visualization, data management, and text-to-speech, are already out there. Creating an environment that enables them to work together in a useful way is a large part of the challenge the KLSG faces. Says Wentling, “Now we have to put all that stuff together in a way that's really meaningful to the lives of ordinary people.”