SHOMA at Parseme Shared Task on Automatic Identification of VMWEs: Neural Multiword Expression Tagging with High Generalisation

TL;DR

This paper introduces a language-independent neural network model for multiword expression identification, achieving state-of-the-art results across multiple languages by leveraging pre-trained embeddings and deep learning layers.

Contribution

The paper presents a novel neural architecture with convolutional, recurrent, and optional CRF layers that outperforms existing systems in MWE identification, especially on unseen data.

Findings

Achieved a macro-average F1 score of 58.09 across languages.

Outperformed all other systems in the Parseme shared task.

Demonstrated strong generalization to unseen data entries.

Abstract

This paper presents a language-independent deep learning architecture adapted to the task of multiword expression (MWE) identification. We employ a neural architecture comprising of convolutional and recurrent layers with the addition of an optional CRF layer at the top. This system participated in the open track of the Parseme shared task on automatic identification of verbal MWEs due to the use of pre-trained wikipedia word embeddings. It outperformed all participating systems in both open and closed tracks with the overall macro-average MWE-based F1 score of 58.09 averaged among all languages. A particular strength of the system is its superior performance on unseen data entries.