FAME: Feature-Based Adversarial Meta-Embeddings for Robust Input Representations

TL;DR

FAME introduces a feature-guided adversarial meta-embedding approach that effectively combines diverse embeddings, improving robustness and performance across multiple languages and tasks, especially in low-resource scenarios.

Contribution

The paper presents FAME, a novel feature-based adversarial meta-embedding method that handles embeddings of different types and sizes without attention, advancing multi-lingual and domain adaptation capabilities.

Findings

Sets new state-of-the-art in POS tagging for 27 languages

Improves NER and question classification performance across domains

Effective in low-resource settings

Abstract

Combining several embeddings typically improves performance in downstream tasks as different embeddings encode different information. It has been shown that even models using embeddings from transformers still benefit from the inclusion of standard word embeddings. However, the combination of embeddings of different types and dimensions is challenging. As an alternative to attention-based meta-embeddings, we propose feature-based adversarial meta-embeddings (FAME) with an attention function that is guided by features reflecting word-specific properties, such as shape and frequency, and show that this is beneficial to handle subword-based embeddings. In addition, FAME uses adversarial training to optimize the mappings of differently-sized embeddings to the same space. We demonstrate that FAME works effectively across languages and domains for sequence labeling and sentence classification, in particular in low-resource settings. FAME sets the new state of the art for POS tagging in 27 languages, various NER settings and question classification in different domains.