The World Artificial Intelligence Conference (WAIC) kicked off in Shanghai recently, with over 500 companies and agencies participating in the event on the theme of “Intelligent World, Indivisible Community.” WAIC gathers the world’s top scientists to talk about the future challenges and breakthroughs of artificial intelligence. At the opening ceremony, leaders from the industry, academia, and research circles expressed their views on the development of AI, including examining the development of AI from the macro, pointing out the challenges and future trends, and proposing executable from the micro for the further development of AI.
The 2020 edition of the WAIC showcased Shanghai’s endeavour to become a rising AI centre of global influence. The event included many talks, dialogues and announcements from government leaders, top scientists and well-known entrepreneurs. The main takeaway from the event was that AI is good for the common values of mankind that benefit the world and experts proposed new solutions for the development and application of artificial intelligence worldwide.
Sign up for your weekly dose of what's up in emerging technology.
For the business perspective, the event had Baidu CEO Robin Li Yanhong, Tencent Chief Operating Officer Ren Yuxin, Tesla CEO Elon Musk, United Nations Digital Cooperation High-level Group Co-Chairman Ma Yun. From Academia, there was Turing Award winners Andrew Yao and Yoshua Bengio.
Baidu CEO Shares AI’s Three Development Stages
Speaking at the conference, Baidu’s CEO Robin Li said that AI development would go through three stages. As the artificial intelligence technology platform is open, more and more applications can be easily developed, benefiting all aspects of the economy and society. He believes that AI is a technological wave comparable to the industrial revolution, and will definitely change every industry. In his view, AI development will go through three stages.
As part of the first stage, the algorithms have seen rapid iteration and innovation. But technological exploration has not evolved into an industrial or economic benefit. The second stage involves the data transformation of the economy. AI can begin to exert its influence in a wide range of businesses.
The third stage is the intelligence of society. Artificial intelligence will penetrate from the economic field to a wider social field. Robin Li also announced two small goals of Baidu: first, the number of intelligent cloud servers will reach 5 million by 2030. Second, the training of more than 5 million AI professionals in the next five years.
Tencent’s Ren Yuxin Describes Focus Areas of Artificial Intelligence
Tencent COO, Ren Yuxin, shared Tencent’s thinking and exploration in the field of artificial intelligence with development areas. The first is the new growth. The normalization of the epidemic makes it very challenging to resume work and resume production. In order to ensure the smooth flow of logistics, capital flow and information, AI has played a major role. According to Yuxin, AI technology is deeply embedded in all walks of life and has become ubiquitous.
Tencent is also supporting Shanghai to build a global esports capital. Gaming provides an ideal experimental field for AI. In the course of AI development, the iconic AlphaGo was born in the ancient Chinese game scene of Go. After solving the Go problem, multiplayer tactical competitive games like MOBA have become an important scenario for us to study the complex decision-making capabilities of AI. At present, Tencent has tried to use the technical advantages of AI in the gaming field to build a simulation environment for car manufacturers to improve test efficiency and safety of autonomous driving.
Musk: Tesla Is Very Close To L5 Autonomous Driving
At WAIC 2020, Elon Musk said he is very confident in the future L5 of autonomous driving and thinks it will be realized soon. According to Musk, Tesla is very close to 5G autonomous driving, and he is confident that we will complete the basic functions of the L5 level this year.
He believes that there is currently no underlying challenge to achieve L5 autonomous driving, but it is necessary to integrate the system and continue to solve many details. In his view, nothing is more complicated than reality, and any simulation is a subset of the complexity of the real world. Musk pointed out that cognition may be the weakest area of artificial intelligence. But in the future, cognition will perform better.
He elaborated that we can think of AI as a game. As long as there are clear game rules, artificial intelligence will definitely play better than humans in any game, he said. Musk also explained why Tesla should develop more AI chips. He said that there are no systems with reasonable cost and low power consumption on the market.
Andrew Yao Spoke On The New Direction Of Artificial Intelligence
Andrew Yao, A.M. Turing Award Winner- 2000 and Professor and the Dean of Institute for Interdisciplinary Information Sciences at Tsinghua University, shared his insights on the current state of AI coming from academic research. According to Yao, the theoretical research done now will give us tremendous progress in the future. In his speech, he mainly talked about three points. First, academia is very important. The challenges facing AI can now be analysed through theory, find ways to find solutions.
Second, AI is definitely an interdisciplinary industry, Yao stated. According to him, there are many examples to show that some of the huge results achieved in AI are often due to cooperation between disciplines that seem to be completely out of bounds. This may require decades of effort because if there are no research results achieved by scientists in other disciplines, it is impossible for AI to achieve such rapid development.
Yoshua Bengio’s Talk On Machine Learning To Empower Contact Tracing
Yet another A.M. Turing Award co-winner, Yoshua Bengio, presented a talk on how ML can be used for contact tracing. Yoshua Bengio said that the problem with standard data tracking is that it only considers the binary information about whether a person’s nucleic acid test is positive or not. According to Bengio, standard tracing (manual or digital) is binary and brittle. In reality, many clues should be integrated before taking a decision about the risk level of a person. This includes symptoms which are available many days before tests, age, biological sex, prior medical conditions, and risk level of all the contacts. ML can be used to predict the infection status probability distribution, given these clues. This provides early warning signals, well before standard tracing would raise a flag. These signals and other information from the app can empower manual contact tracers to make informed decisions and extend their reach. He proposed that the machine learning on mobile phones can be used to strengthen the tools of manual contact tracing and expand the prediction range.
The video recordings of the various talks and panels can be found here.