Exploring Swedish & English fastText Embeddings for NER with the Transformer

TL;DR

This study demonstrates that smaller, well-trained embeddings can outperform larger ones in NER tasks using Transformers, and introduces a new Swedish analogy test set to aid future research.

Contribution

It shows that smaller corpora can produce effective embeddings for NER and provides a new Swedish analogy test set for benchmarking.

Findings

Smaller embeddings can outperform larger ones in NER tasks.

Character n-grams improve Swedish embedding performance.

Swedish and English embeddings achieve better results with less training data.

Abstract

In this paper, our main contributions are that embeddings from relatively smaller corpora can outperform ones from larger corpora and we make the new Swedish analogy test set publicly available. To achieve a good network performance in natural language processing (NLP) downstream tasks, several factors play important roles: dataset size, the right hyper-parameters, and well-trained embeddings. We show that, with the right set of hyper-parameters, good network performance can be reached even on smaller datasets. We evaluate the embeddings at both the intrinsic and extrinsic levels. The embeddings are deployed with the Transformer in named entity recognition (NER) task and significance tests conducted. This is done for both Swedish and English. We obtain better performance in both languages on the downstream task with smaller training data, compared to recently released, Common Crawl versions; and character n-grams appear useful for Swedish, a morphologically rich language.