Hierarchical Multi-Instance Multi-Label Learning for Detecting Propaganda Techniques

Anni Chen, Bhuwan Dhingra

Abstract
Since the introduction of the SemEval 2020 Task 11 (CITATION), several approaches have been proposed in the literature for classifying propagandabased on the rhetorical techniques used to influence readers. These methods, however, classify one span at a time, ignoring dependencies from the labels of other spans within the same context. In this paper, we approach propaganda technique classification as aMulti-Instance Multi-Label (MIML) learning problem (CITATION) and propose a simple RoBERTa-based model (CITATION) for classifying all spans in an article simultaneously. Further, we note that, due to the annotation process whereannotators classified the spans by following a decision tree,there is an inherent hierarchical relationship among the differenttechniques, which existing approaches ignore. We incorporate these hierarchical label dependencies by adding an auxiliary classifier for each node in the decision tree to the training objective and ensembling the predictions from the original and auxiliary classifiers at test time. Overall, our model leads to an absolute improvement of 2.47% micro-F1 over the model from the shared task winning team in a cross-validation setup and is the best performing non-ensemble model on the shared task leaderboard.
Anthology ID:
2023.repl4nlp-1.13
Volume:
Proceedings of the 8th Workshop on Representation Learning for NLP (RepL4NLP 2023)
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Burcu Can, Maximilian Mozes, Samuel Cahyawijaya, Naomi Saphra, Nora Kassner, Shauli Ravfogel, Abhilasha Ravichander, Chen Zhao, Isabelle Augenstein, Anna Rogers, Kyunghyun Cho, Edward Grefenstette, Lena Voita
Venue:
RepL4NLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
155–163
Language:
URL:
https://aclanthology.org/2023.repl4nlp-1.13
DOI:
10.18653/v1/2023.repl4nlp-1.13
Bibkey:
Cite (ACL):
Anni Chen and Bhuwan Dhingra. 2023. Hierarchical Multi-Instance Multi-Label Learning for Detecting Propaganda Techniques. In Proceedings of the 8th Workshop on Representation Learning for NLP (RepL4NLP 2023), pages 155–163, Toronto, Canada. Association for Computational Linguistics.
Cite (Informal):
Hierarchical Multi-Instance Multi-Label Learning for Detecting Propaganda Techniques (Chen & Dhingra, RepL4NLP 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-1/2023.repl4nlp-1.13.pdf
PDF Search