At 18 o'clock on the August 13th, the two-day CCF-GAIR artificial intelligence and robotics summit ended perfectly. During this period, the world’s academics, industry, and investment community gathered together to present a grand event for related practitioners. Technology and experience sharing sessions. The entire conference agenda includes: artificial intelligence home, robots and drones, smart driving and investors. First, let's review the August 12th AI series of speeches at home.
Academician Xu Yangsheng of the Chinese Academy of Engineering: Issues and Reflections on Intelligent Robots
First of all, Xu Yangsheng, principal of the Chinese University of Hong Kong (Shenzhen), academician of the Chinese Academy of Engineering, robotics and intelligent system expert, based on his more than 30 years of robotic research experience, summed up the technical path, historical evolution and application cases of intelligent robots.
Kingsoft CEO Zhang Hongjiang: smart, big data, cloud
“The pace of artificial intelligence development is so fast that it is inseparable from big data and cloud computing.†CEO of Kingsoft, former president of Microsoft Asia Engineering Academy, academician of the American Computer Association (ACM), IEEE Fellow Zhang Hongjiang in “Smart, Big Data, Cloud ". Twenty years ago, Zhang Hongjiang’s group applied for a patent for a complex distributed architecture but could only operate in a laboratory. Today, 20 years later, it has already left the lab and the entire calculation and processing process is completed in the cloud. , can be presented on our mobile devices for users to use anytime, anywhere.
Oxford Computer Department Head: Symbolism and neural networks in AI should be integrated
The emergence of AlphaGo this year for the first time allowed the world to recognize the advanced level of artificial intelligence at this stage. Then the question of where advanced artificial intelligence has come from, and whether AI poses a threat to humans has been thrown out. In this conference, one of the DeepMind members of the team behind AlphaGo, the head of the computer department of Oxford University, the head of the Oxford-DeepMind Partnership, AAAI, and Michael Wooldridge of the EURAI Fellow, in-depth analysis of strong and weak artificial intelligence, machines Whether it will have the problems of awareness, integration and integration of neural network type AI and symbolism AI, at the same time put forward two kinds of AI solutions, as well as put forward their own opinions on the contradictions people encounter in the field of artificial intelligence research.
Professor Yang Qiang of Hong Kong University of Science and Technology: Several Necessary Conditions for the Success of Artificial Intelligence
Professor Yang Qiang of Hong Kong University of Science and Technology presented the five conditions necessary for the success of artificial intelligence at the CCF-GAIR conference:
Clear business model
High quality big data
Clear problem definition and field edges
Understanding AI's cross-border talent
Calculate ability
We know that artificial intelligence has developed so rapidly in recent years that deep learning has contributed. Professor Yang Qiang is studying a new algorithm model whose operation principle is based on deep learning plus reinforcement learning and migration learning, so that some mature data models can be more easily applied in other fields without having to 0 to 1 training data.
Executive Vice President of Microsoft Research Asia Yong Yong: Computer Vision From Perception to Cognition
Computer vision is an important branch of artificial intelligence. How to let computers understand images has always been a major problem in the field of research. During the conference, executive vice presidents of Microsoft Asia Research Institute, IEEE, IAPR, and SPIE Fellow dared to explain the profound things in a simple and easy way. They analyzed computer vision with lively and interesting practical applications, and at the same time, let people directly feel the wonders of CV.
Academician Li Ming of the Canadian Royal Society: How to deal with NLP problems with deep learning
In addition to computer vision, natural language processing is an important part of artificial intelligence. Both Siri and Xiao Bing are natural language processing products, but the user's impression of it is only at the entertainment level. There are still many problems in answering the questions. Fellow of the Royal College of Canada, Professor of the University of Waterloo, ACM, IEEE Fellow Li Ming used a deep learning model different from other methods at the CCF-GAIR conference to make the robot Q&A more accurate. In addition, he also put forward the concept of “semantic distanceâ€, giving NLP researchers a new research idea.
Li Hang, Director of Huawei's Noah's Ark Lab: Our Vision and Practice in Artificial Intelligence
As we all know, companies such as Google, Facebook, Microsoft, and China’s Baidu all have relatively deep accumulation in the field of artificial intelligence. However, many people do not seem to know that, in fact, Huawei has already begun to deploy artificial intelligence and established the Noah’s Ark Laboratory. This time, Li Hang, director of the Huawei Noah's Ark Lab, talked about Huawei's Noah's Ark Lab's deep-growing products and technologies in the AI ​​direction.
From Search Engines to Action Engines, How Google Scientists Made Smart Searches
At present, speech recognition has reached a very high level, but semantic analysis still has a long way to go. Therefore, in the competitive environment with a small difference in overall speech recognition rate, the pros and cons of the semantic analysis algorithm and the usage scenarios that have been adopted have become the most important conditions for NLP products. At the CCF-GAIR conference, former Google scientist Lindkang cut in from the perspective of the NLP, and described in depth the presentation of search engines and voice assistants in the future.