TY - GEN
T1 - UIUC BioNLP at SemEval-2021 Task 11
T2 - 15th International Workshop on Semantic Evaluation, SemEval 2021
AU - Liu, Haoyang
AU - Sarol, Janina
AU - Kilicoglu, Halil
N1 - Publisher Copyright:
© 2021 Association for Computational Linguistics.
PY - 2021
Y1 - 2021
N2 - 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 in English. 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 Phase 1, our approach yielded 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.
AB - 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 in English. 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 Phase 1, our approach yielded 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.
UR - http://www.scopus.com/inward/record.url?scp=85135820473&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85135820473&partnerID=8YFLogxK
U2 - 10.18653/v1/2021.semeval-1.45
DO - 10.18653/v1/2021.semeval-1.45
M3 - Conference contribution
AN - SCOPUS:85135820473
T3 - SemEval 2021 - 15th International Workshop on Semantic Evaluation, Proceedings of the Workshop
SP - 377
EP - 386
BT - SemEval 2021 - 15th International Workshop on Semantic Evaluation, Proceedings of the Workshop
A2 - Palmer, Alexis
A2 - Schneider, Nathan
A2 - Schluter, Natalie
A2 - Emerson, Guy
A2 - Herbelot, Aurelie
A2 - Zhu, Xiaodan
PB - Association for Computational Linguistics (ACL)
Y2 - 5 August 2021 through 6 August 2021
ER -