The

The "Artificial Intelligence Frontier Forum" of the Artificial Intelligence and Robot Innovation Center of the Hong Kong Innovation Academy of the Chinese Academy of Sciences was successfully held

• Publication date: February 5, 2021

On February 2, the "Artificial Intelligence Frontier Forum" sponsored by the Artificial Intelligence and Robot Innovation Center of the Hong Kong Innovation Academy of the Chinese Academy of Sciences (hereinafter referred to as the Artificial Intelligence and Robot Innovation Center) was held online. The "AI Frontier Forum" is the first of a series of international academic exchange activities organized by the AI and Robot Innovation Center. It aims to promote the in-depth development of AI research and promote the exchange of AI technology and talents by inviting global talents to discuss the key frontier theories and technical problems of the future development of AI.

At the special moment of the global fight against the COVID-19, the "AI Frontier Forum" was online in the whole process, gathered with experts, scholars and young talents on the cloud, and discussed the current hot issues in the field of AI, attracting more than 20000 people to watch online. The chairman of the forum is researcher Xu Bo, director of the Institute of Automation of the Chinese Academy of Sciences, the chairman of the procedure committee is researcher Zhang Zhaoxiang, and the deputy director of the Institute of Automation, researcher Liu Chenglin, and researcher Lei Zhen jointly serve as the sub link moderators.

Xu Bo said in his speech that the AI and Robot Innovation Center will fully combine the advantages of the mainland and Hong Kong in different fields, carry out research work around the new generation of AI basic theory, new human-computer interaction technology and equipment, health-oriented and barrier-free intelligence technology, AI open platform technology and other aspects, and help the construction of science and technology in the Greater Bay Area of Guangdong, Hong Kong and Macao. He sincerely invited talented people at home and abroad to join him and work together to create a new chapter in the development of artificial intelligence and robots.

In the keynote speech, Shi Luping, a professor of Tsinghua University, Tao Dacheng, a foreign academician of the European Academy of Sciences and an academician of the Australian Academy of Sciences, Zhang Lei, a chair professor of the Hong Kong Polytechnic University, and James Kwok, a professor of the Hong Kong University of Science and Technology, focused on brain computing and in-depth learning, respectively brought wonderful keynote speeches.

Shi Luping said that the current challenge of AI research lies in the separate development of computer science and neuroscience, and the lack of deep integration of multiple disciplines. The new computing technology, brain computing, which uses the way of human brain to store and process information, will be one of the keys to the future development. He discussed the brain-like computing mode driven by dual-brain fusion and believed that it was still in its infancy and had not yet formed a recognized technical solution. It needed the coordinated development of theory, chips, software, systems and applications. It was the cornerstone of AI competition and could empower all walks of life.

Tao Dacheng has brought insights into credible and in-depth learning. He reviewed the development of deep learning theory and pointed out that privacy and robustness are still important problems in the development of AI. Although the neural network still has some problems, researchers still have expectations for the current in-depth learning, and hope that the model is more robust, less energy consumption, more interpretable, can adapt to the dynamic environment, pay attention to human privacy security and easy to use. In the future, AI will be more compatible with human values, and we should pay attention to the protection of privacy while promoting technology.

Zhang Lei believes that the normalization method is very important for the effective and efficient optimization of the depth neural network, and reasonable use can make the training process more stable. In the report, he shared the two latest developments of the deep neural network learning normalization method, Batch Statistics Regulation for Batch Normalization and Gradient Centralization for Deep Neural Network Learning, which can improve the optimization and generalization of the model.

James Kwok observed that there is a great demand for machine learning knowledge. AutoML can alleviate the pressure brought by the shortage of machine learning talents to some extent. In his speech, he shared the recent progress of automatic machine learning, and specifically introduced the research of effective neural structure search and interactive function search in collaborative filtering using automatic machine learning.

In the academic report, Professor Zuo Wangmeng from Harbin University of Technology, Researcher Liu Xin from the researcher of mathematics and systems science of the Chinese Academy of Sciences, Associate Professor Wu Baoyuan from the Chinese University of Hong Kong (Shenzhen), Associate Professor Liu Yong from Renmin University of China, Associate Professor Yang Meng from Sun Yat-sen University, and Assistant Professor Wang Shiqi from the City University of Hong Kong focused on self-monitoring context modeling and low-level visual application respectively, The new method of Stiefel manifold optimization - distributed computing, AI security, large-scale machine learning model selection theory, weak supervised learning of semantic segmentation and target location, and face detection with high accuracy and strong generalization ability, and other research directions, shared their unique views, and brought a cutting-edge and rich AI feast to the audience.

The Hong Kong Institute of Innovation of the Chinese Academy of Sciences is the first scientific research institution established by the Chinese Academy of Sciences in Hong Kong. Its construction and development are positioned as a world-class new research and development institution, a highland for attracting international high-end talents to innovate, a hub for scientific and technological innovation cooperation and exchange, and an important part of building an international scientific and technological innovation center with global influence in the Guangdong Hong Kong Macao Greater Bay Area, The Institute of Automation of the Chinese Academy of Sciences established the AI and Robot Innovation Center in Hong Kong Science Park. In the future, the Innovation Center will be committed to building a shared and open platform for international scientific and technological exchanges and cooperation. Now it will recruit high-level talents globally, establish high-level and international research teams, and jointly create a bright future for the development of artificial intelligence.