Seminar on Computational Learning and Adaptation

  An Interactive Environment for Scientific Model Construction

Pat Langley

Computational Learning Laboratory
Center for the Study of Language and Information
Stanford University
http://cll.stanford.edu/~langley/

Most AI research on scientific model construction aims to automate this process using discovery techniques. In contrast, I describe an interactive environment for model construction that lets the user construct, edit, and visualize scientific models, use them to make predictions, and call on discovery methods to revise them in ways that better fit the available data. The environment relies on a new formalism that embeds mathematical equations, which are familiar to many scientists, within distinct processes, which can encode background knowledge used to constrain model revision. I report initial studies on ecosystem modeling that suggest this environment is more effective than earlier approaches and more transparent to users. In closing, I discuss related work on modeling environments and model revision, then suggest directions for future research.

This talk describes work done jointly with Kevin Arrigo, Stephen Bay, Jaime Fitzgerald, Steve Klooster, Chris Potter, Javier Sanchez, and Dan Shapiro.

Note: This is a practice run for an upcoming talk at the Knowledge Capture Conference, so it will be less polished and shorter than most presentations in this seminar.

Return to the seminar schedule