In the past two years, artificial intelligence (AI) has become a hot topic in the industry. As a technology leader, IBM provides a one-stop service from cloud, hardware and cross-industry consultants to help companies invest more effectively in internal data. Drive AI to land in the actual application scenario.
In fact, IBM currently has more than 1,400 applications in the field of artificial intelligence applications, including GPU acceleration, image analysis, data encryption, semantic interpretation and other innovative technologies. Last year, there were practical applications in Taiwan's major industries, especially in the financial and medical industries, including the financial robots that IBM Watson and E.SUM Bank cooperated to analyze and identify the Chinese conversation content of users and respond to the corresponding conversations. ; In addition, it is imported into a medical institution to assist gene sequences through image recognition.
From a hardware perspective, IBM is also the most powerful support role for AI applications. IBM launched a new generation of POWER9 processors designed for artificial intelligence in December last year. The maximum I/O bandwidth is 9.5 times that of Intel x86 processors, and the memory capacity is 2.6 times. The number of high-performance cores is twice that of x86, and the memory bandwidth is 1.8 times that of x86. It can improve the training time of nearly 4 times deep learning framework and deal with various AI tasks that require a lot of intensive operations.