TY - GEN
T1 - Hierarchical Metadata-Aware Document Categorization under Weak Supervision
AU - Zhang, Yu
AU - Chen, Xiusi
AU - Meng, Yu
AU - Han, Jiawei
N1 - Publisher Copyright:
© 2021 ACM.
PY - 2021/8/3
Y1 - 2021/8/3
N2 - Categorizing documents into a given label hierarchy is intuitively appealing due to the ubiquity of hierarchical topic structures in massive text corpora. Although related studies have achieved satisfying performance in fully supervised hierarchical document classification, they usually require massive human-annotated training data and only utilize text information. However, in many domains, (1) annotations are quite expensive where very few training samples can be acquired; (2) documents are accompanied by metadata information. Hence, this paper studies how to integrate the label hierarchy, metadata, and text signals for document categorization under weak supervision. We develop HiMeCat, an embedding-based generative framework for our task. Specifically, we propose a novel joint representation learning module that allows simultaneous modeling of category dependencies, metadata information and textual semantics, and we introduce a data augmentation module that hierarchically synthesizes training documents to complement the original, small-scale training set. Our experiments demonstrate a consistent improvement of HiMeCat over competitive baselines and validate the contribution of our representation learning and data augmentation modules.
AB - Categorizing documents into a given label hierarchy is intuitively appealing due to the ubiquity of hierarchical topic structures in massive text corpora. Although related studies have achieved satisfying performance in fully supervised hierarchical document classification, they usually require massive human-annotated training data and only utilize text information. However, in many domains, (1) annotations are quite expensive where very few training samples can be acquired; (2) documents are accompanied by metadata information. Hence, this paper studies how to integrate the label hierarchy, metadata, and text signals for document categorization under weak supervision. We develop HiMeCat, an embedding-based generative framework for our task. Specifically, we propose a novel joint representation learning module that allows simultaneous modeling of category dependencies, metadata information and textual semantics, and we introduce a data augmentation module that hierarchically synthesizes training documents to complement the original, small-scale training set. Our experiments demonstrate a consistent improvement of HiMeCat over competitive baselines and validate the contribution of our representation learning and data augmentation modules.
KW - hierarchical text categorization
KW - metadata-aware text categorization
KW - weak supervision
UR - http://www.scopus.com/inward/record.url?scp=85102110722&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85102110722&partnerID=8YFLogxK
U2 - 10.1145/3437963.3441730
DO - 10.1145/3437963.3441730
M3 - Conference contribution
AN - SCOPUS:85102110722
T3 - WSDM 2021 - Proceedings of the 14th ACM International Conference on Web Search and Data Mining
SP - 770
EP - 778
BT - WSDM 2021 - Proceedings of the 14th ACM International Conference on Web Search and Data Mining
PB - Association for Computing Machinery
T2 - 14th ACM International Conference on Web Search and Data Mining, WSDM 2021
Y2 - 8 March 2021 through 12 March 2021
ER -