Multilingual Entity Linking Benchmark for Occupations

This document is a technical report that introduces the Multilingual Entity Linking of Occupations (MELO) Benchmark, which consists of 48 datasets aimed at evaluating the linking of entity mentions in 21 languages to the ESCO Occupations multilingual taxonomy. The benchmark is built using high-quality, pre-existing human annotations, addressing the lack of public evaluation benchmarks in this area. The report details the methodology for constructing the benchmark and presents an experimental study that evaluates the performance of both simple lexical models and advanced deep learning models in a zero-shot setup. The findings indicate that while lexical models perform adequately, deep learning approaches generally yield better results, especially in cross-lingual tasks. The document emphasizes the importance of accurate entity resolution in digital HR systems and aims to serve as a valuable resource for future research and innovation in multilingual entity linking.