A Large Parallel Corpus of Full-Text Scientific Articles

Sentence-aligned full-text scientific articles collected from the Scielo database. The corpus supports three language pairs and includes a small subset of trilingual sentence alignments. It was presented at the Eleventh International Conference on Language Resources and Evaluation (LREC-2018) in 2018.

Use Cases

• Train machine translation models based on the parallel sentence alignments for English, Portuguese, and Spanish.
• Develop cross-lingual information retrieval systems for scientific literature based on the aligned full-text articles.
• Conduct linguistic analysis of scientific writing styles across different languages based on the trilingual subset.
• Benchmark multilingual text processing algorithms based on the structured, domain-specific corpus.

Strengths

• Corpus is sentence-aligned for all three language pairs (English-Portuguese, English-Spanish, Portuguese-Spanish).
• Includes a subset of sentences that are trilingual aligned, which is less common in parallel corpora.
• Based on full-text articles from the Scielo database, providing domain-specific scientific content.

Limitations

• Row count and total corpus size are unknown, which may limit suitability assessment for large-scale training.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Scielo database

Collection Method

Collected and sentence-aligned by the authors.

Time Range

null

Freshness

null

Geography

null