Lightly-supervised representation learning with global interpretability

Andrew Zupon, Maria Alexeeva, Marco A. Valenzuela-Escárcega, Ajay Nagesh, Mihai Surdeanu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Scopus citations


We propose a lightly-supervised approach for information extraction, in particular named entity classification, which combines the benefits of traditional bootstrapping, i.e., use of limited annotations and interpretability of extraction patterns, with the robust learning approaches proposed in representation learning. Our algorithm iteratively learns custom embeddings for both the multi-word entities to be extracted and the patterns that match them from a few example entities per category. We demonstrate that this representation-based approach outperforms three other state-of-the-art bootstrapping approaches on two datasets: CoNLL-2003 and OntoNotes. Additionally, using these embeddings, our approach outputs a globally-interpretable model consisting of a decision list, by ranking patterns based on their proximity to the average entity embedding in a given class. We show that this interpretable model performs close to our complete bootstrapping model, proving that representation learning can be used to produce interpretable models with small loss in performance. This decision list can be edited by human experts to mitigate some of that loss and in some cases outperform the original model.

Original languageEnglish (US)
Title of host publicationNLP@NAACL-HLT 2019 - 3rd Workshop on Structured Prediction for NLP, Proceedings
PublisherAssociation for Computational Linguistics (ACL)
Number of pages11
ISBN (Electronic)9781950737109
StatePublished - 2021
Event3rd Workshop on Structured Prediction for NLP, NLP@NAACL-HLT 2019 - Minneapolis, United States
Duration: Jun 7 2019 → …

Publication series

NameNLP@NAACL-HLT 2019 - 3rd Workshop on Structured Prediction for NLP, Proceedings


Conference3rd Workshop on Structured Prediction for NLP, NLP@NAACL-HLT 2019
Country/TerritoryUnited States
Period6/7/19 → …

ASJC Scopus subject areas

  • Software
  • Computational Theory and Mathematics


Dive into the research topics of 'Lightly-supervised representation learning with global interpretability'. Together they form a unique fingerprint.

Cite this