Optimizing the Performance of Text Classification Models by Improving the Isotropy of the Embeddings using a Joint Loss Function
No Thumbnail Available
Access rights
openAccess
acceptedVersion
URL
Journal Title
Journal ISSN
Volume Title
A4 Artikkeli konferenssijulkaisussa
This publication is imported from Aalto University research portal.
View publication in the Research portal (opens in new window)
View/Open full text file from the Research portal (opens in new window)
Other link related to publication (opens in new window)
View publication in the Research portal (opens in new window)
View/Open full text file from the Research portal (opens in new window)
Other link related to publication (opens in new window)
Date
2023-08-19
Major/Subject
Mcode
Degree programme
Language
en
Pages
16
Series
Document Analysis and Recognition – ICDAR 2023 - 17th International Conference, Proceedings, pp. 121-136, Lecture notes in computer science
Abstract
Recent studies show that the spatial distribution of the sentence representations generated from pre-trained language models is highly anisotropic. This results in a degradation in the performance of the models on the downstream task. Most methods improve the isotropy of the sentence embeddings by refining the corresponding contextual word representations, then deriving the sentence embeddings from these refined representations. In this study, we propose to improve the quality of the sentence embeddings extracted from the [CLS] token of the pretrained language models by improving the isotropy of the embeddings. We add one feed-forward layer between the model and the downstream task layers, and we train it using a novel joint loss function. The proposed approach results in embeddings with better isotropy, that generalize better on the downstream task. Experimental results on 3 GLUE datasets with classification as the downstream task show that our proposed method is on par with the state-of-the-art, as it achieves performance gains of around 2–3% on the downstream tasks compared to the baseline.Description
Keywords
Other note
Citation
Attieh, J, Zewoudie, A, Vlassov, V, Flanagan, A & Bäckström, T 2023, Optimizing the Performance of Text Classification Models by Improving the Isotropy of the Embeddings using a Joint Loss Function . in G A Fink, R Jain, K Kise & R Zanibbi (eds), Document Analysis and Recognition – ICDAR 2023 - 17th International Conference, Proceedings . Lecture notes in computer science, Springer, pp. 121-136, International Conference on Document Analysis and Recognition, San Jose, California, United States, 21/08/2023 . https://doi.org/10.1007/978-3-031-41734-4_8