PLOD: An Abbreviation Detection Dataset for Scientific Documents

The detection and extraction of abbreviations from unstructured texts can help to improve the performance of Natural Language Processing tasks, such as machine translation and information retrieval. However, in terms of publicly available datasets, there is not enough data for training deep-neural-networks-based models to the point of generalising well over data. This paper presents PLOD, a large-scale dataset for abbreviation detection and extraction that contains 160k+ segments automatically annotated with abbreviations and their long forms. We performed manual validation over a set of instances and a complete automatic validation for this dataset. We then used it to generate several baseline models for detecting abbreviations and long forms. The best models achieved an F1-score of 0.92 for abbreviations and 0.89 for detecting their corresponding long forms. We release this dataset along with our code and all the models publicly in

PDF Abstract LREC 2022 PDF LREC 2022 Abstract


Introduced in the Paper:

PLOD-filtered PLOD-unfiltered

Used in the Paper:

Acronym Identification
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
AbbreviationDetection PLOD-filtered RoBERTa-large F1-Score (AC) 92.0 # 1
F1-Score (LF) 89.8 # 1
AbbreviationDetection PLOD-unfiltered surrey-nlp/roberta-large-finetuned-abbr F1-Score (LF) 89.8 # 1
F1-Score (AC) 92.2 # 2
AbbreviationDetection PLOD-unfiltered surrey-nlp/albert-large-v2-finetuned-abbDet F1-Score (LF) 87.2 # 2
F1-Score (AC) 90.7 # 1