Ping An Wins Top Ranking in Global Artificial Intelligence Challenge
HONG KONG, June 28, 2019 /PRNewswire/ -- Ping An Insurance (Group) Company of China, Ltd. "(hereafter "Ping An" or the "Group" or the "Company", HKEX: 2318; SSE: 601318)" has received global recognition for its achievements in artificial intelligence (AI) processing of human language in the healthcare field.
The 2019 MEDIQA challenge, sponsored by the international Association of Computational Linguistics (ACL), involved three tests. Ping An Smart City's Natural Language Processing team (PANLP) ranked first in Recognizing Question Entailment (RQE), and seconds in Natural Language Inference (NLI) and medical Question Answering (QA) respectively.
72 teams took part in the challenge, which aims to develop methods, techniques and gold standards for NLP in the medical domain.
"This outstanding achievement demonstrated the world-class healthcare Natural Language Processing technology capability of Ping An," said Dr. Xie Guotong, Chief Healthcare Scientist of Ping An Group. "Natural Language Processing is one of the most difficult challenges in AI. The MEDIQA challenge has increased the interest of many first-class universities and research institutions."
"Recognizing Question Entailment technology has been applied in Ping An's healthcare AI services, such as AskBob, the medical search engine, available in more than 1000 medical organizations in China to provide critical clinical decision support services to enhance doctors' diagnosis and treatment decisions," added Dr. Ni Yuan, Head of PANLP. "The win reflects Ping An's strength in academic competition as well as advanced medical technology capabilities."
Ping An Group's medical healthcare AI covers more than 1,000 common diseases, and about 800 million potential patients benefit from it. The Group expects to invest RMB100 billion (USD15 billion) in technology research and development in the next decade to consolidate its technology leadership in the financial services industry.
NOTES TO EDITOR
The objective of RQE task is to identify entailment between two questions in the context of QA. The objective of NLI task is to identify three inference relations between two sentences: Entailment, Neutral and Contradiction. The objective of QA task is to filter and improve the ranking of automatically retrieved answers.
The evaluations for RQE and NLI tasks are based on Accuracy. The evaluation of QA task is based on the Accuracy, Mean Reciprocal Rank (MRR), Precision, and Spearman's Rank Correlation Coefficient.
For more information, please visit www.pingan.cn
For enquiries, please contact:
Gareth Hewett @ +852 68822027 or gareth.john@pingan.com.hk
Share this article