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

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.

A Survey of MWE Identification Experiments: The Devil is in the Details

“…Such cases are rare in the corpora, and as such do not greatly impact the data. One paper (Walsh et al, 2022) attempts to address this problem of overlapping or shared-token expressions by modifying the BIOstyle encoding, while another paper (Taslimipoor and Rohanian, 2018) appends multiple categories separated by a semicolon, similar to the CUPT-style encoding.…”

“…Analyses tended to take one of two forms: example-based analysis reporting individual instances where the model performed better or worse than usual (Klyueva et al, 2017;Walsh et al, 2022), and automatic metrics aggregated across particular properties or phenomena. Among the focused metrics, some papers pay special attention to discontinuities (Björne and Salakoski, 2016;Moreau et al, 2018;Berk et al, 2018a;Rohanian et al, 2019) and seen/unseen MWEs (Maldonado et al, 2017;Zampieri et al, 2018;Taslimipoor and Rohanian, 2018). Some studies analyse the model's features and modules via ablation experiments (Scherbakov et al, 2016;Tang et al, 2016;Stodden et al, 2018;Pasquer et al, 2020a).…”

“…One exception was made for the SHOMA system paper, available only on arXiv, but listed in the PARSEME ST 1.1 paper and website(Taslimipoor and Rohanian, 2018).…”

Predicting Compositionality of Verbal Multiword Expressions in Persian

“…In recent years, deep learning has demonstrated remarkable success in sequence tagging tasks, including MWE identification (Ramisch et al, 2018;Taslimipoor and Rohanian, 2018). RNNs and ConvNets have shown significant progress in this area.…”

The Role of Semi-productivity in Multiword Expression Identification: Why can BERT Capture novel MWEs?

“…However, such techniques have been proven effective only when dealing with very specific MWE classes [5]. In addition, it should be noted that the models that performed the best in the last two editions of the PARSEME shared task: SHOMA [24] and MTLB-STRUCT [23], both based on deep learning, achieved the best score also on unseen MWEs [17,18].…”