This is a code respository of our WebConf'24 research paper Global News Synchrony and Diversity During the Start of the COVID-19 Pandemic and ICWSM'24 dataset paper A Multilingual Similarity Dataset for News Article Frame.

The code directory of the sections are:

1. Data collection: ner_art_sampling;

2. Semeval Baseline: Semeval_baseline;

3. Transformer model: network_inference;

4. Graph Clustering data: ner_art_sampling/network_data/input/oslom.tsv_oslo_files/partitions_level_0;

5. Regression Analysis: ner_art_sampling/visualize_matched_inference.py.

The overall processing pipeline is: The data collection pipeline for multilingual news article similarity is: