A field guide to China’s medical NLP jungle

What it does

This is an awesome-list-style index that catalogs open resources for Chinese medical natural language processing. It gathers links to terminology sets, corpora, pretrained models like ChineseEHRBert and MC-BERT, knowledge graphs, NER datasets, QA collections, and shared-task benchmarks such as CBLUE. Think of it as a field guide to a highly specialized, linguistically distinct corner of the NLP ecosystem.

The interesting bit

Medical NLP in English is already a maze; in Chinese it is fragmented across hospital partnerships, academic competitions, and isolated GitHub repos. The maintainer has done the tedious work of tracking down ICD-9/10/11 Chinese mappings, OMAHA terminology samples, and even traditional Chinese medicine knowledge graphs, turning a scavenger hunt into a single reference page.

Key highlights

• Covers the full pipeline: tokenization tools, word vectors, BERT variants, relation extraction, NER, and QA datasets.
• Includes major benchmarks like CBLUE and shared-task data from CCKS/CHIP competitions.
• Lists concrete resources such as the 1.1-million-dialogue MedDialog dataset and the patient-event graph built from Shanghai hospital records.
• Surfaces niche assets: TCM case knowledge graphs, herb-net ontologies, and pharmacists’ examination question sets.

Caveats

• This is a curated directory, not a unified framework; expect to do your own plumbing.
• Most annotations are brief, single-line Chinese descriptions, so skimming for relevance requires reading comprehension or translation.
• Coverage is broad rather than deep: the repo points to resources but does not compare or benchmark them.

Verdict

Anyone building Chinese clinical NLP pipelines or hunting for East Asian medical corpora should bookmark this. If you are looking for a drop-in Python library, this is just a reading list.