Sign in to view source links and access this dataset
Description
ILSUM-2.0 is a multilingual news dataset for India, containing at least 10,000 articles per language. It extends the ILSUM-1.0 dataset by adding Bengali to the existing Hindi, English, and Gujarati languages. The dataset, created by ILSUM and last updated in July 2024, is built from articles and headlines from several leading Indian newspapers.
Use Cases
Train abstractive summarization models based on the article-headline pairs.
Benchmark multilingual summarization performance across Hindi, English, Gujarati, and Bengali.
Study cross-lingual transfer learning for low-resource languages like Bengali.
Analyze journalistic writing styles and headline conventions across major Indian newspapers.
Strengths
Contains at least 10,000 articles per language, providing substantial data for each.
Adds Bengali language support to the original ILSUM-1.0 dataset, expanding its scope.
Built from several leading national newspapers, suggesting a degree of editorial quality.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown beyond the per-language minimum, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
ILSUM
Collection Method
Built using articles and headline pairs from several leading newspapers of the country.
Freshness
Last updated 2024-07-24 06:37:15; freshness should be verified.
Geography
India
License is unknown; terms of use must be verified before application.