DingTalk Rebuilds AI Ecosystem: Platform Launches Gynecology Specialized Large Model with Over 90% Accuracy

Recently, the "Daukou Gynecology Large Model," developed by Yisheng Jiankang (Hangzhou) Life Science Co., Ltd., achieved an accuracy rate of 90.2% in professional testing. This marks the first high-accuracy, highly practical domain-specific large model successfully trained on the DingTalk AI platform. Once deployed, the model will effectively alleviate the shortage of specialized gynecologists and help more female users manage their health.

DingTalk Powers the Deployment of Industry-Specific Large Models

The launch of the gynecology large model signifies that the DingTalk ecosystem has expanded beyond SaaS, service providers, consulting, and delivery ecosystems to include AI entrepreneurs. As general-purpose AI large models become more widespread, industries across the board are striving to deeply integrate these models with their specific business scenarios, creating industry- or domain-specific large models to drive cost reduction and efficiency gains in operations and management. To address challenges such as complex data engineering, high technical requirements, and data security during the training of proprietary models, DingTalk has built a support system for developing industry/enterprise large models. This system helps companies build large model applications from data annotation through training, inference, and deployment, enabling them to harness the productivity revolution brought by AI.

At 9:21 p.m. on June 30, good news came from the office of Yisheng Jiankang on the 5th floor of the Lvfeng Center in Hangzhou. Their large model—the Daukou Gynecology Large Model—trained on the DingTalk enterprise-specific AI platform, had just completed 100 gynecology-specific test questions. The results showed that the model achieved an accuracy rate of 90.2% in diagnosing six major gynecological symptoms: irregular menstruation, abnormal bleeding, abnormal vaginal discharge, lower abdominal pain, and lower abdominal masses, demonstrating significant practical value.

"The Daukou Gynecology Large Model is like an AI gynecologist, capable of providing pre-consultation and health management services to more women," said Wang Qiangyu, founder of Yisheng Jiankang. "Users simply select their symptoms in the dialogue interface of the 'Qinmi Doctor' app, and they receive professional self-diagnosis results, including primary diagnoses, potential additional diagnoses, recommended tests, suggested treatment plans, and important notes."

Compared to the average 30-minute wait time for traditional online consultations, Qinmi Doctor can generate professional recommendations within seconds, helping women quickly determine whether they need urgent medical attention—especially beneficial for working women and users in remote areas.

Yisheng Jiankang is a life science company specializing in precision diagnostics and health services for women. Its founding team comes largely from well-known internet companies, obstetrics and gynecology medical institutions, and biopharmaceutical firms. Based on technological trends and industry insights, Wang Qiangyu's team believes that training a gynecology-specific large model to create an AI doctor will effectively address the shortage of specialized gynecologists and inadequate medical services, bringing immense industry-wide and societal value to both medical aesthetic institutions and female users.

Just like training a real gynecology expert, a highly specialized "gynecology AI doctor" cannot be easily created using a general-purpose large model. Since launching the development of the Daukou Gynecology Large Model, the Yisheng Jiankang team used an open-source large model as a foundation and trained it with industry-specific data, achieving an initial diagnostic accuracy of around 77.1%. "While 77.1% meets the basic industry standard, it still falls short for medical AI, where health safety is at stake. We must continue to push the boundaries, moving from 'general knowledge coverage' to 'vertical-domain expertise,'" said Wang Qiangyu.

When the performance of the Daukou Gynecology Large Model hit a bottleneck, Yisheng Jiankang shifted the model-training platform to the DingTalk enterprise-specific AI platform. With DingTalk's support, the Yisheng Jiankang team made multi-faceted adjustments in data processing, computing power, and model optimization. Within a month, the two parties boosted the diagnostic accuracy of the Daukou Gynecology Large Model to 90.2%.

The significantly enhanced capabilities of the Daukou Gynecology Large Model will strongly support the rapid growth of Yisheng Jiankang's business. Moving forward, the company will continue to improve the model's performance and accuracy, integrating it into its own product, "Qinmi Doctor," and serving more users through an "AI doctor" intelligent agent. Wang Qiangyu said, "Through this kind of practice, in the future, not only gynecology but also other vertical models, such as dermatology, can be trained and brought into daily life, allowing ordinary people to receive preliminary health guidance at home that rivals that of professional medical institutions."