Microsoft Artificial Intelligence Research Symposium
On March 6, 2018, at the Innocentre in Hong Kong, Microsoft Research Asia (MSRA) hosted the Microsoft Artificial Intelligence Research Symposium in conjunction with Microsoft Hong Kong.
To help Microsoft customers involved in R&D better integrate AI and cloud computing technologies into their work, Microsoft Hong Kong decided it would be useful to inform those customers about Microsoft’s research in AI. Said Cally Chan, General Manager of Microsoft Hong Kong, “It’s important for us to connect with the education and R&D communities to share Microsoft’s knowledge on the topics of AI, big data, and the cloud.”
MSRA and Microsoft Hong Kong decided to host a joint symposium to share the latest research and technologies with academic customers in disciplines including: medicine, biomed and public health, environment, law, and transportation.
The symposium was kicked off by Dr. Winnie Cui, region lead from MSRA. She introduced Microsoft Research and highlighted its major achievements on AI: image recognition rates lower than humans, speech recognition equal to that of humans, and natural language understanding that exceeds human abilities. She also discussed Xiaoice, a chatbot who has 2 million registered users in China alone. Said Cui, “Research plays a crucial role in driving technology breakthroughs at Microsoft, and we believe in the value of research and the value of sharing this groundbreaking research with people. As we make computers perform better in seeing, hearing, talking, and understanding humans, eventually they will be able to better help and serve human beings and our society.”
Several MSRA researchers shared their research findings at the symposium. Knowing many researchers in the audience are doing big data analysis, Dr. Xing Xie reported on recent research progress in big data applications of user modeling, understanding user mobility patterns, and personality trait analysis. Dr. Tao Qin introduced a new framework in AI called Dual Learning, which leverages the primal-dual structure of AI tasks to obtain effective feedback or regularization signals to enhance the learning/inference process. Dr. Frank Soong gave an in-depth discussion on his work on text-to-speech analysis.
Professors from local universities shared their research work as well. The audience appreciated the fact that Microsoft is collaborating with local universities on both leading-edge AI technologies and their application to address societal issues ranging from environment to healthcare.
Along with the research talks, the symposium showcased the spectrum of Microsoft AI services, tools, and platform technologies so that customers in the R&D sector can apply them in their work.