CiYi (词义)

A repo for lexical semantics

MWE Type

PIE Classification

@inproceedings{tan-jiang-2021-bert,
    title = "Does {BERT} Understand Idioms? A Probing-Based Empirical Study of {BERT} Encodings of Idioms",
    author = "Tan, Minghuan  and
      Jiang, Jing",
    booktitle = "Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)",
    month = sep,
    year = "2021",
    address = "Held Online",
    publisher = "INCOMA Ltd.",
    url = "https://aclanthology.org/2021.ranlp-main.156",
    pages = "1397--1407",
    abstract = "Understanding idioms is important in NLP. In this paper, we study to what extent pre-trained BERT model can encode the meaning of a potentially idiomatic expression (PIE) in a certain context. We make use of a few existing datasets and perform two probing tasks: PIE usage classification and idiom paraphrase identification. Our experiment results suggest that BERT indeed can separate the literal and idiomatic usages of a PIE with high accuracy. It is also able to encode the idiomatic meaning of a PIE to some extent.",
}

SemEval 2022 Task 2

Multilingual Idiomaticity Detection and Sentence Embedding

@inproceedings{tan-2022-hijonlp,
    title = "{H}i{J}o{NLP} at {S}em{E}val-2022 Task 2: Detecting Idiomaticity of Multiword Expressions using Multilingual Pretrained Language Models",
    author = "Tan, Minghuan",
    booktitle = "Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)",
    month = jul,
    year = "2022",
    address = "Seattle, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2022.semeval-1.23",
    doi = "10.18653/v1/2022.semeval-1.23",
    pages = "190--196",
    abstract = "This paper describes an approach to detect idiomaticity only from the contextualized representation of a MWE over multilingual pretrained language models.Our experiments find that larger models are usually more effective in idiomaticity detection. However, using a higher layer of the model may not guarantee a better performance.In multilingual scenarios, the convergence of different languages are not consistent and rich-resource languages have big advantages over other languages.",
}

Subtask A

Data Preprocess

python experiments/semeval-2022_task02_idiomacity/subtask_a/create_data.py \
  --input_location ../SemEval_2022_Task2-idiomaticity/SubTaskA \
  --output_location data/annotations/semeval-2022_task02_idiomacity/subtask_a \
  --phase evaluation

Train

bash run_semeval2022_task2a.sh data

Subtask B

Data Preprocess

python experiments/semeval-2022_task02_idiomacity/subtask_b/create_data.py \
  --input_location ../SemEval_2022_Task2-idiomaticity/SubTaskB \
  --output_location data/annotations/semeval-2022_task02_idiomacity/subtask_b \
  --sts_dataset_path stsbenchmark.tsv.gz

Train

bash run_semeval2022_task2b.sh data

Acknowledgement

We recommend the following repos:

• lexcomp
• allennlp-models