IBM Research, Singapore
Cognitive system is the platform to provide the deep learning capability for different AI solutions. From IBM point of view, it includes those innovations from system hardware, system software, cloud, and deep learning. In this session, we will use PowerAI (IBM’s cognitive system framework) as an example, to introduce the design principle and how it impacts deep learning training stage and inference stage. We will discuss how some of the recent innovations, such as I/O and accelerator could bring the difference. We will also share how to design the cognitive system with container cloud technologies can achieve productivity, scalability and high availability, for both training and inference. There is an option for a demo (in the IBM booth) for a system called VisionBrain (which is an IBM developed Deep Learning based platform for image/video analysis to provide customized deep learning capability for enterprise customer). The system can be demonstrated together with some slides for differentiation, and link to IBM’s PowerAI (https://www.ibm.com/us-en/marketplace/deep-learning-platform)
Yonghua Lin is Senior Technical Staff Member and Senior Manager of Cognitive AI System in IBM Research. She is the founder of IBM Supervessel Cloud (www.ptopenlab.com), which provides open services to communities and industry, e.g. accelerator service, big data, cognitive service and DL/ML computing. She has worked on system architecture, cloud computing and cognitive system research for more than 15 years. And now she is focusing on the AI system for industry. She led the research group develop the large scale AI system on cloud and edge, based on the leading technologies like FPGA virtualization, Deep Learning acceleration and Extreme-Low Bit DNN, etc. Besides, she is driving the innovation on deep learning platform for visual comprehension, called VisionBrain. It has been deployed to provide the image recognition service on IBM cloud. In the past more than 10 years, her work covered all kinds of IBM multicore processors,including IBM network processor, IBM Cell processor, PRISM, IBM POWER 6/7/8, etc. She was also the initiator of Wireless Network Cloud with SDR technology in industry, which is called C-RAN today. She led IBM team built up the FIRST optimized cloud for 4G mobile infrastructure, and successfully demonstrated in ITU, Mobile World Congress, etc. She herself has more than 50 patents granted worldwide and publications in top conferences and journals. She is the chair of IEEE Women in Engineering Beijing Section in 2017.