Update README.md

README.md

@@ -1,3 +1,27 @@
----
-license: cc0-1.0
----
+---
+license: cc0-1.0
+task_categories:
+- translation
+language:
+- en
+- ja
+---
+# Wikidata parallel descriptions en-ja
+- Parallel corpus for machine translation generated from wikidata dump (2024-05-06).
+- Currently we processed only English/Japanese pair.
+- The `jsonl` file is ready-to-train by Hugging Face transformers translation task trainer.
+
+# Dataset Details
+- https://www.wikidata.org/wiki/Wikidata:Database_download
+
+## Dataset Creation
+- As Wikidata description field is noisy and does not represent exact direct translation, filtering is required for machine translation task.
+- We did word count based filtering by using Japanese tokenizer Mecab and did some word-based filtering like `ウィキメディア`
+- We also did deduplication.
+
+## Bias, Risks, and Limitations
+- The data is still noisy and might not be very parallel.
+
+# License
+- CC0
+  - Because all structured data in Wikidata is originally licensed in CC0, this dataset can be also licensed in same license.