SemRel2024: A Collection of Semantic Textual Relatedness Datasets for 13 Languages
Cardiff University · Bayero University, Kano-Nigeria · University of Alberta · Ahmadu Bello University · Northeastern University · Research, Microsoft · Mohamed bin Zayed University of Artificial Intelligence and Amazon · KU Leuven · Bahir Dar University, Universität Hamburg · Uppsala University · U Hamburg · University of Melbourne · International Institute of Information Technology Hyderabad, India · Mohamed bin Zayed University of Artificial Intelligence and University of Houston · International Institute of Information Technology Hyderabad · Kotebe University of Education · Capital One AI Foundations · Universität Hamburg · National Research Council Canada
PDF 由论文原始站点提供，PaperCompass 不保存论文文件。DOI 10.18653/v1/2024.findings-acl.147 ↗
摘要
Exploring and quantifying semantic relatedness is central to representing language and holds significant implications across various NLP tasks. While earlier NLP research primarily focused on semantic similarity, often within the English language context, we instead investigate the broader phenomenon of semantic relatedness. In this paper, we present SemRel, a new semantic relatedness dataset collection annotated by native speakers across 13 languages: Afrikaans, Algerian Arabic, Amharic, English, Hausa, Hindi, Indonesian, Kinyarwanda, Marathi, Moroccan Arabic, Modern Standard Arabic, Spanish, and Telugu. These languages originate from five distinct language families and are predominantly spoken in Africa and Asia – regions characterised by a relatively limited availability of NLP resources. Each instance in the SemRel datasets is a sentence pair associated with a score that represents the degree of semantic textual relatedness between the two sentences. The scores are obtained using a comparative annotation framework. We describe the data collection and annotation processes, challenges when building the datasets, baseline experiments, and their impact and utility in NLP.