TransPOS: Transformers for Consolidating Different POS Tagset Datasets

TL;DR

This paper introduces TransPOS, a Transformer-based method for merging datasets with different POS tagging schemes, but finds that disjoint labels pose significant challenges and may not be effective for dataset consolidation.

Contribution

It proposes a novel Transformer architecture to address the problem of consolidating disjoint POS-tagged datasets, highlighting theoretical difficulties and limitations.

Findings

The approach diverges from expectations, indicating challenges in merging disjoint datasets.

Disjoint labels may not be effective for dataset consolidation.

Theoretical analysis reveals difficulties in using disjoint POS tags for merging datasets.

Abstract

In hope of expanding training data, researchers often want to merge two or more datasets that are created using different labeling schemes. This paper considers two datasets that label part-of-speech (POS) tags under different tagging schemes and leverage the supervised labels of one dataset to help generate labels for the other dataset. This paper further discusses the theoretical difficulties of this approach and proposes a novel supervised architecture employing Transformers to tackle the problem of consolidating two completely disjoint datasets. The results diverge from initial expectations and discourage exploration into the use of disjoint labels to consolidate datasets with different labels.