Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark

Ningyu Zhang, Mosha Chen, Zhen Bi, Xiaozhuan Liang, Lei Li, Xin Shang, Kangping Yin, Chuanqi Tan, Jian Xu, Fei Huang, Luo Si, Yuan Ni, Guotong Xie, Zhifang Sui, Baobao Chang, Hui Zong, Zheng Yuan, Linfeng Li, Jun Yan, Hongying Zan, Kunli Zhang, Buzhou Tang, Qingcai Chen

TL;DR

CBLUE introduces the first Chinese Biomedical Language Understanding Evaluation benchmark to address the lack of Chinese biomedical benchmarks. It aggregates real-world data from diverse sources into eight tasks spanning NER, information extraction, normalization, and classification, and provides an open leaderboard with reproducible PyTorch tooling. Across 11 public Chinese PTMs, results show substantial gaps to human performance and highlight task-specific challenges such as long-tail distributions and colloquial biomedical language. By offering detailed annotation protocols, privacy-preserving data handling, and case studies, CBLUE aims to accelerate Chinese BioNLP research and foster community-driven dataset expansion.

Abstract

Artificial Intelligence (AI), along with the recent progress in biomedical language understanding, is gradually changing medical practice. With the development of biomedical language understanding benchmarks, AI applications are widely used in the medical field. However, most benchmarks are limited to English, which makes it challenging to replicate many of the successes in English for other languages. To facilitate research in this direction, we collect real-world biomedical data and present the first Chinese Biomedical Language Understanding Evaluation (CBLUE) benchmark: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis. To establish evaluation on these tasks, we report empirical results with the current 11 pre-trained Chinese models, and experimental results show that state-of-the-art neural models perform by far worse than the human ceiling. Our benchmark is released at \url{https://tianchi.aliyun.com/dataset/dataDetail?dataId=95414&lang=en-us}.

CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark

TL;DR

CBLUE introduces the first Chinese Biomedical Language Understanding Evaluation benchmark to address the lack of Chinese biomedical benchmarks. It aggregates real-world data from diverse sources into eight tasks spanning NER, information extraction, normalization, and classification, and provides an open leaderboard with reproducible PyTorch tooling. Across 11 public Chinese PTMs, results show substantial gaps to human performance and highlight task-specific challenges such as long-tail distributions and colloquial biomedical language. By offering detailed annotation protocols, privacy-preserving data handling, and case studies, CBLUE aims to accelerate Chinese BioNLP research and foster community-driven dataset expansion.

Abstract

Artificial Intelligence (AI), along with the recent progress in biomedical language understanding, is gradually changing medical practice. With the development of biomedical language understanding benchmarks, AI applications are widely used in the medical field. However, most benchmarks are limited to English, which makes it challenging to replicate many of the successes in English for other languages. To facilitate research in this direction, we collect real-world biomedical data and present the first Chinese Biomedical Language Understanding Evaluation (CBLUE) benchmark: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis. To establish evaluation on these tasks, we report empirical results with the current 11 pre-trained Chinese models, and experimental results show that state-of-the-art neural models perform by far worse than the human ceiling. Our benchmark is released at \url{https://tianchi.aliyun.com/dataset/dataDetail?dataId=95414&lang=en-us}.

Paper Structure

This paper contains 100 sections, 2 equations, 2 figures, 32 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. CBLUE Overview
  4. Design Principle
  5. Tasks
  6. CMeEE
  7. CMeIE
  8. CHIP-CDN
  9. CHIP-CTC
  10. CHIP-STS
  11. KUAKE-QIC
  12. KUAKE-QTR
  13. KUAKE-QQR
  14. Data Collection
  15. Annotation
  16. ...and 85 more sections

Figures (2)

  • Figure 1: Analysis of the named entity recognition and information extraction datasets. (a) illustrates the entity (coarse-grained) distribution in CMeEE and the impact of data distribution on the model's performance. We set entity type Body with the maximum number of entities to 1.0, and others to the ratio of number or F1 score to Body. (b) shows the relation hierarchy in CMeIE.
  • Figure 2: We conduct error analysis on datasets CMeEE and QIC. For CMeEE, we divide error cases into 6 categories, including ambiguity, need domain knowledge, entity overlap, wrong entity boundary, annotation error, and others (long sequence, rare words, etc.). For KUAKE-QIC, we divide error cases into 7 categories, including multiple triggers, colloquialism, ambiguity, rare words, annotation error, irrelevant description, and need domain knowledge.