Seeing Meaning

Jethran Guinness, Philip DeCamp, and Deb Roy, with previous work by Amber Frid-Jimenez and Chris Lucas.

We propose a new approach to the problem of information exploration and knowledge discovery in which we conceptualize the computer as a human-like partner that works symbiotically with humans to achieve joint goals.

We are applying this approach to implement interactive visualization systems with emphasis on analyzing the content of a wide variety of literature. Our goal is to develop interactive visual and tactile representations of natural language analysis algorithms to foster fluid and intuitive query refinement and system evaluation. In contrast to the standard computer-as-tool view, we envision the machine as a conversational, interactive agent that moves beyond the current generation of information retrieval and visualization tools, and towards systems which can explain the reasons underlying their actions, and interact with human partners to refine queries through natural conversational interaction.