Exploring Multilingual Text Data Distillation
This work addresses efficiency in training deep learning models for multilingual text tasks, but it is incremental as it builds upon existing techniques.
The authors tackled the problem of data distillation for multilingual text classification, which is challenging due to the discrete nature of text and poor generalization to new architectures, by proposing language-model-based techniques that improve cross-architecture generalization and language-specific fairness.
With the rise of deep learning, large datasets and complex models have become common, requiring significant computing power. To address this, data distillation has emerged as a technique to quickly train models with lower memory and time requirements. However, data distillation on text-based datasets hasn't been explored much because of the challenges rising due to its discrete nature. Additionally, existing dataset distillation methods often struggle to generalize to new architectures. In the paper, we propose several data distillation techniques for multilingual text classification datasets using language-model-based learning methods. We conduct experiments to analyze their performance in terms of classification strength, and cross-architecture generalization. Furthermore, we investigate the language-specific fairness of the data summaries generated by these methods. Our approach builds upon existing techniques, enhancing cross-architecture generalization in the text data distillation domain.