Former Director's Introduction

Whenever a visitor drops into my office, our conversation usually settles on one or more of three topics:

What applications are we working on.
What is happening scientifically.
What is happening commercially.
To tell the "applications" story, I generally draw a few current slides from a large library that covers the work of our laboratory's 168 members. You see here a representative subset of the slides in that library.
To tell the "scientific" story, from a personal point of view, I explain why I am optimistic about the immediate future, for reasons explained, informally, in a brief essay, why I am optimistic.

Much of the application-oriented work in the laboratory is organized around projects. Many of the projects focus on creating new sorts of connections between and among people, computers, the physical world, and the world of knowledge.
For example, we have developed systems that connect surgeons to their patients with superman-like x-ray vision, and we have developed systems that use natural language to connect people to the tidal wave of information resources provided by world wide web.
Thus, the word connection is a good handle for much of what we do.

In our Enhanced Reality Project, we produce systems that overlay previously recorded information on current images. The lead example works in the world of brain surgery, providing brain surgeons with a kind of supermanlike x-ray vision: they see previously recorded MRI images registered and superimposed on an image currently produced by a TV camera.
Surgeons who have used the system report that they can do their work much quicker, and hence with less trauma to the patient. They also can undertake operations that would have been too risky without the system.

In our Virtual Objects Project, we connect people to simulated worlds through their sense of touch. Medical students, for example, can feel the forces produced by virtual tissues and organs as they execute common diagnostic procedures. Artists can feel forces exerted by virtual paintbrushes on simulated canvasses, and watch paint spread out on a screen as they push harder.

In our Intelligent Gunk Project we aim to embed computers in the physical world's structures, coatings, and materials.
Our original work in this area focused on the development of an experimental bridge in which the compression beams were made nearly ten times stronger by way of a system that senses deflections quickly and exerts tiny corrective forces with piezo-electric actuators.

In our Intelligent Room Project, we use work in natural language understanding and computer vision to pioneer new ideas in human--computer interaction.
The overall idea is to develop software agents that enter into our own world, on our terms, so that we no longer have to work with computers on their terms, using today's mice and keyboards.
Research in human--computer interaction has deep scientific goals as well as high-impact applied goals. We believe, for example, that our goals in human--computer interaction will help us come to grip with issues such as understanding how vision-rooted and language-rooted reasoning work together to produce intelligence.

The InfoLab Project is dedicated to the study and use of language-based systems as interfaces for information access.
We have built a series of information servers, many based on the START natural language system, which are able to retrieve many kinds of media, including text, images, sound and video, in response to questions posed to it in everyday language. One such server answers questions about the U.S. mission in Bosnia. For example, the question "How did the Dayton Peace accords divide Bosnia'' is answered with a map.

Our Intelligent Information Infrastructure Project develops ideas and software in support of the National Information Infrastructure. The key idea is that the National Information Infrastructure should understand the information it is carrying, so as to enable new ways to gather, organize, and transmit knowledge, as well as new ways to operate organizations.
We have built, for example, the publication and distribution system used by the White House Office of Media affairs, in use routinely since January 20, 1993 to distribute various publications nationally and internationally, and an on-line surveying system used to determine the size and character of the audience receiving the documents.

In our Design Associates Project, we aim to use models of structure, function, and causality, together with sophisticated symbolic and numerical computation, to perform a wide range of problem solving and reasoning tasks. The goal is to build systems that work with human engineers to produce better, more innovative designs. Thus the objective is to make a quantum leap beyond the traditional CAD tools that work more at the level of drafting assistants.