Discovering Constraints for Inductive Process Modeling

Authors

  • Ljupco Todorovski University of Ljubljana
  • Will Bridewell Stanford University
  • Pat Langley Institute for the Study of Learning and Expertise

DOI:

https://doi.org/10.1609/aaai.v26i1.8152

Keywords:

inductive process modeling, learning constraints, receiver operating characteristic (ROC) curve

Abstract

Scientists use two forms of knowledge in the construction ofexplanatory models: generalized entities and processes that relatethem; and constraints that specify acceptable combinations of thesecomponents. Previous research on inductive process modeling, whichconstructs models from knowledge and time-series data, has relied onhandcrafted constraints. In this paper, we report an approach todiscovering such constraints from a set of models that have beenranked according to their error on observations. Our approach adaptsinductive techniques for supervised learning to identify processcombinations that characterize accurate models. We evaluate themethod's ability to reconstruct known constraints and to generalizewell to other modeling tasks in the same domain. Experiments with synthetic data indicate that the approach can successfully reconstructknown modeling constraints. Another study using natural data suggests that transferring constraints acquired from one modeling scenario to another within the same domain considerably reduces the amount of search for candidate model structures while retaining the most accurate ones.

Downloads

Published

2021-09-20

How to Cite

Todorovski, L., Bridewell, W., & Langley, P. (2021). Discovering Constraints for Inductive Process Modeling. Proceedings of the AAAI Conference on Artificial Intelligence, 26(1), 256-262. https://doi.org/10.1609/aaai.v26i1.8152

Issue

Section

AAAI Technical Track: Cognitive Systems