Ensembling Finetuned Language Models for Text Classification
This work addresses the need for more reliable and higher-performing text classifiers in NLP applications, but it is incremental as it applies existing ensembling techniques to a new context.
The paper tackled the problem of improving text classification performance by exploring ensembling strategies for finetuned language models, and found that ensembling can boost performance, though specific numbers are not provided in the abstract.
Finetuning is a common practice widespread across different communities to adapt pretrained models to particular tasks. Text classification is one of these tasks for which many pretrained models are available. On the other hand, ensembles of neural networks are typically used to boost performance and provide reliable uncertainty estimates. However, ensembling pretrained models for text classification is not a well-studied avenue. In this paper, we present a metadataset with predictions from five large finetuned models on six datasets, and report results of different ensembling strategies from these predictions. Our results shed light on how ensembling can improve the performance of finetuned text classifiers and incentivize future adoption of ensembles in such tasks.