UIUC_BioNLP at SemEval-2021 Task 11: A Cascade of Neural Models for Structuring Scholarly NLP Contributions

TL;DR

This paper presents a neural cascade system that automatically structures scholarly NLP contributions by classifying sentences, recognizing phrases, and extracting triples, achieving top rankings in SemEval-2021 evaluations.

Contribution

The novel approach combines BERT-based models and rules in a cascade to improve the automatic structuring of scholarly contributions in NLP papers.

Findings

Ranked second in Phase 1 evaluation

Achieved first place in Phase 2 evaluation

System demonstrated strong performance after fixing submission errors

Abstract

We propose a cascade of neural models that performs sentence classification, phrase recognition, and triple extraction to automatically structure the scholarly contributions of NLP publications. To identify the most important contribution sentences in a paper, we used a BERT-based classifier with positional features (Subtask 1). A BERT-CRF model was used to recognize and characterize relevant phrases in contribution sentences (Subtask 2). We categorized the triples into several types based on whether and how their elements were expressed in text, and addressed each type using separate BERT-based classifiers as well as rules (Subtask 3). Our system was officially ranked second in Phase 1 evaluation and first in both parts of Phase 2 evaluation. After fixing a submission error in Pharse 1, our approach yields the best results overall. In this paper, in addition to a system description, we also provide further analysis of our results, highlighting its strengths and limitations. We make our code publicly available at https://github.com/Liu-Hy/nlp-contrib-graph.