[Global Technology Report] On March 29, Alibaba Cloud announced the ET medical brain and announced that it officially entered the medical AI field. It is understood that after more than a year of research and training, artificial intelligence ET has a number of medical capabilities and can assume the role of a doctor's assistant in areas such as patient virtual assistants, medical imaging, precision medical treatment, drug efficacy excavation, new drug development, and health management.
The ET medical brain research and development uses a large number of deep learning technologies and uses massive amounts of data as an example to train a machine to accomplish a specific task. That is, the computer can improve medical skills by learning case data. Because you can sleep 24 hours and handle thousands of tasks at the same time, ET's learning progress is much faster than humans.
Ali Wan, an artificial intelligence scientist, said: "In the testing of multiple diseases, we are aware that the machine not only can complete the task, but also in some aspects is as good as the human doctor. ET is fully qualified to become a doctor's assistant."
In addition to drastically improving the doctor's work efficiency, the ET Medical Brain also tries to fundamentally defeat cancer. By cooperating with BGI, ET analyzes the DNA sequence of a large number of lung adenocarcinoma cases to find the key gene mutations that cause the disease.
“We set the ET medical brain as an open artificial intelligence system. In addition to Alibaba Cloud's artificial intelligence scientists being involved in research and development, we would like to absorb external sophisticated algorithms and medical experience so that ET can grow faster. Senior physician.†Wan Wanli said.
The scientist of Zhejiang Deshang Yunxing is the developer of the ultrasound diagnostic algorithm for thyroid nodules. They use deep learning to process ultrasound images, and at the same time, they add concepts such as rotation invariance to modern mathematics, making the system more “smartâ€. The algorithm is now integrated into the ET medical brain and piloted in multiple hospitals.
With the aid of computer vision technology, this algorithm can rapidly scan and analyze the thyroid B-ultrasound, circle the nodular region, and give a judgment of benign and malignant, which greatly saves the doctor's diagnosis time. In general, the accuracy of human doctors is 60%-70%, and the accuracy of the current algorithm has reached 85%.
On the same day, Aliyun announced the joint launch of the Tianchi Medical AI Series with Intel and LinkDoc. The first quarter challenged the diagnosis of early lung cancer.
The Aliyun Tianchi platform will provide massive desensitization and high-resolution chest CT scan data. The player needs to obtain the nodule feature through the original CT image training model algorithm, and finally realizes the intelligent judgment of the nodule area of ​​the image picture.
In fact, Alibaba Cloud has begun to explore the medical field before ET medical brains. Last year, Shanghai Huashan Hospital took advantage of Alibaba Cloud's computing power to replace some clinical trials with digital models and live experiments simulating mice to speed up the development of specific drugs.
Also in the same year last year, BGI, Aliyun and Anhui Medical University jointly announced that they had completed analysis of 1000 human exomes in 21 hours, 47 minutes and 12 seconds. 40 years ago, if humans wanted to sequence the whole genome of E. coli, it would take 1000 years.
The same day Aliyun also announced the recent progress of ET in the industrial field. At present, Alibaba Cloud Artificial Intelligence ET has become a good helper for humans in various fields such as transportation, justice, industrial manufacturing, and medical care.