Evaluation of BERT and ALBERT Sentence Embedding Performance on Downstream NLP Tasks
This work addresses the need for efficient sentence embeddings in NLP tasks, but it is incremental as it builds on existing Sentence-BERT and ALBERT methods.
The paper tackled the problem of finding optimal sentence embeddings for BERT and ALBERT models by modifying them with siamese/triplet structures and adding an outer CNN network, resulting in improved performance for ALBERT on semantic textual similarity benchmarks while maintaining competitiveness with fewer parameters.
Contextualized representations from a pre-trained language model are central to achieve a high performance on downstream NLP task. The pre-trained BERT and A Lite BERT (ALBERT) models can be fine-tuned to give state-ofthe-art results in sentence-pair regressions such as semantic textual similarity (STS) and natural language inference (NLI). Although BERT-based models yield the [CLS] token vector as a reasonable sentence embedding, the search for an optimal sentence embedding scheme remains an active research area in computational linguistics. This paper explores on sentence embedding models for BERT and ALBERT. In particular, we take a modified BERT network with siamese and triplet network structures called Sentence-BERT (SBERT) and replace BERT with ALBERT to create Sentence-ALBERT (SALBERT). We also experiment with an outer CNN sentence-embedding network for SBERT and SALBERT. We evaluate performances of all sentence-embedding models considered using the STS and NLI datasets. The empirical results indicate that our CNN architecture improves ALBERT models substantially more than BERT models for STS benchmark. Despite significantly fewer model parameters, ALBERT sentence embedding is highly competitive to BERT in downstream NLP evaluations.