Advances in Monolingual and Crosslingual Automatic Disability Annotation in Spanish

Abstract

Background: Unlike diseases, automatic recognition of disabilities has not received the same attention in the area of medical NLP. Progress in this direction is hampered by obstacles like the lack of annotated corpus. Neural architectures learn to translate sequences from spontaneous representations into their corresponding standard representations given a set of samples. The aim of this paper is to present the last advances in monolingual (Spanish) and crosslingual (from English to Spanish) automatic disability annotation. The task consists of identifying disability mentions in medical texts written in Spanish within a collection of abstracts from journal papers related to the biomedical domain. Results: In order to carry out the task, we have combined deep learning models that use different embedding granularities for sequence to sequence tagging with simple rules to boost the coverage. Conclusions: Our monolingual experiments demonstrate that a good combination of different word embedding representations provide better results than single representations, significantly outperforming the state of the art in disability annotation in Spanish. Additionally, we have experimented crosslingual transfer (zero-shot) for disability annotation from English to Spanish with interesting results that might help overcoming the data scarcity bottleneck, specially significant for the disabilities.

Advances in monolingual and crosslingual automatic disability annotation in Spanish

“…The DIANN shared task [14] was dedicated to the detection of disability mentions in biomedical research texts in English and Spanish, with the objective of evaluating the performance of various named entity recognition systems in two different languages. In the first position, [39] presented a neural network-based architecture system consisting of a bidirectional long short term memory network (BiLSTM) and a conditional random field (CRF), using static word embeddings for both languages combined with a rule-based acronyms and abbreviation module for the detection of disability-related acronyms and abbreviations, obtaining an F-measure of 0.82 and 0.78 for English and Spanish, respectively. [14] uses a long short-term memory architecture for disabilities, improving the state of the art, with an F-measure of 0.83 and 0.81 for English and Spanish.…”