LS-Tree: Model Interpretation When the Data Are Linguistic

Authors

  • Jianbo Chen University of California, Berkeley
  • Michael Jordan University of California, Berkeley

DOI:

https://doi.org/10.1609/aaai.v34i04.5749

Abstract

We study the problem of interpreting trained classification models in the setting of linguistic data sets. Leveraging a parse tree, we propose to assign least-squares-based importance scores to each word of an instance by exploiting syntactic constituency structure. We establish an axiomatic characterization of these importance scores by relating them to the Banzhaf value in coalitional game theory. Based on these importance scores, we develop a principled method for detecting and quantifying interactions between words in a sentence. We demonstrate that the proposed method can aid in interpretability and diagnostics for several widely-used language models.

Downloads

Published

2020-04-03

How to Cite

Chen, J., & Jordan, M. (2020). LS-Tree: Model Interpretation When the Data Are Linguistic. Proceedings of the AAAI Conference on Artificial Intelligence, 34(04), 3454-3461. https://doi.org/10.1609/aaai.v34i04.5749

Issue

Section

AAAI Technical Track: Machine Learning