Dialogue with Ping An Group Chief Scientist Xiao Jing: We need to focus on actual pain points, AI is not just a show of skill

Author | Fu Jing

On December 21, the 2020 National Artificial Intelligence Competition came to a successful conclusion in Shenzhen.

This competition is hosted by the Shenzhen Municipal People's Government, co-organized by Shenzhen Science and Technology Innovation Committee, Pengcheng Laboratory and the New Generation Artificial Intelligence Industry Technology Innovation Strategic Alliance, and co-organized by Ping An Technology, Huawei, Tencent, Intellifusion and Shenzhen Bus Group.

Ping An Technology, one of the co-organizers of this competition, is a technology solution expert and high-tech core of Ping An Group. Driven by the dual wheels of "finance + technology", it has a huge business system and is focused on providing end-to-end intelligent technology services to institutions, enterprises, and governments. It leads ecological partners to reach more and broader business scenarios and continuously provides input and data for AI technology upgrades.

Ping An Technology was established in 2008 and began to experiment with technological innovation in the financial sector. In March 2015, Dr. Xiao Jing, winner of the Wu Wenjun Artificial Intelligence Outstanding Contribution Award, the highest honor in China's artificial intelligence field and an AI academic expert, joined the company, laying a solid foundation for Ping An Technology's underlying artificial intelligence technology and applications. Today, based on artificial intelligence, Ping An Technology has covered five major ecosystems: finance, healthcare, automobiles, real estate, and smart cities.

During the competition, Leiphone.com had an in-depth conversation with Dr. Xiao Jing, chief scientist of Ping An Group.

As the main co-organizer of the competition and a judge of the "AI+Pedestrian Re-ID" track, Ping An Technology participated in the planning of the competition at all levels, including track design and competition rule formulation, and arranged experts from the algorithm team and implementation team to cooperate with Pengcheng Laboratory to participate in the review work throughout the process.

This competition has three tracks: "AI+Wireless Communication", "AI+Remote Sensing Image", and "AI+Pedestrian Re-identification". The "AI+Pedestrian Re-identification" track is built on the basis of the previous competition. Compared with last year, it is more difficult and requires contestants to combine real and complex scenes such as public transportation, design algorithms and train them based on large-scale data sets, so as to solve more severe problems such as perspective, lighting, and occlusion.

Dr. Xiao Jing said:

The upgrade of this track focuses on improving the technical capabilities of pedestrian re-identification. It is hoped that the contestants will give full play to their creativity in the competition, propose more accurate and robust pedestrian re-identification algorithms, continuously seek innovation and breakthroughs in new technologies, and focus more on industrial applications to help apply pedestrian re-identification technology to a broader market.

It is worth noting that Ping An Technology attaches great importance to attracting talents. For contestants with outstanding abilities, Ping An Technology provides a green channel for formal employee recruitment and project internship opportunities. The selected candidates will have the opportunity to directly participate in the most challenging artificial intelligence projects.

Of course, the cooperation between Ping An Technology and the organizer of the competition, Pengcheng Laboratory, in terms of industry-university-research-application is not limited to this. Dr. Xiao Jing gave two examples:

• Smart transportation: By combining Pengcheng Lab’s digital retina technology and Ping An Technology’s intelligent vision technology, the existing smart city system can be improved.

• Smart finance field: On April 8, 2020, the two companies jointly launched the "Federated Learning White Paper V2.0" with a number of companies and institutions to jointly build industry technology application references and guidance for data security and privacy protection.

Dr. Xiao Jing said that in the future, Ping An Technology will also work with Pengcheng Laboratory to carry out multi-faceted cooperation in the fields of finance, medical care, transportation, smart cities, etc., jointly promote the application and promotion of artificial intelligence technologies, products and services in key industries, and cooperate to cultivate high-quality, comprehensive innovative talents, effectively promoting the output of original and disruptive major technological innovation results.

As a wholly-owned subsidiary of Ping An Group, which has full financial licenses and a wide range of business product lines, Ping An Technology, which started out as a technology company, has embraced the wave of artificial intelligence and become a cross-border player driven by the dual wheels of "finance + technology".

Compared with general fintech companies, fintech subsidiaries can rely on the core advantages of financial institutions such as data, scenarios, experts, algorithms, and funds. By introducing technological professionals and combining business pain points and demands, they can output leading technology solutions for the industry, truly combining business and technology and maximizing business value.

Dr. Xiao Jing said:

Within Ping An Group, Ping An Technology not only plays a role in maintaining the security and stability of the group's core systems, but is also responsible for supporting the IT needs of all vertical businesses, enabling each subsidiary to enhance the market competitiveness of its core businesses and incubate innovative technology businesses.

In recent years, Ping An Technology has actively empowered five major ecosystems, including financial services, healthcare, automotive services, real estate finance, and urban services, through the "AI platform + ecosystem" model.

In the financial field, Ping An Technology has built the "Ping An Brain" intelligent engine to help various business units improve their intelligent construction. For example, the world's first intelligent damage assessment system for auto insurance images can achieve damage assessment in seconds by taking a photo and uploading it. This not only reduces the risk of fraud, but also improves the user experience.

In the medical field, Ping An has created a full-process intelligent solution covering predictive prevention and control, screening diagnosis, auxiliary treatment, patient management and other aspects, which has been put into practical use in many medical institutions across the country. Leifeng.com learned that as of the third quarter of 2020, Ping An's COVID-19 intelligent film reading service has covered more than 1,500 medical institutions across the country, with a total of more than 40 million films read.

Undoubtedly, Ping An Technology's strong artificial intelligence technology support and digital transformation capabilities are behind the creation of standardized industry solutions for multiple business fields and practical scenarios.

Dr. Xiao Jing said that in the field of artificial intelligence, Ping An Technology has four dimensions of unique advantages and competitiveness: leading technological capabilities, profound business understanding, rich practical scenarios, and deep and coordinated technology and business integration. For example, in terms of technology, Ping An has established 8 research institutes and 57 research laboratories; in terms of technology and business integration, internal business and technology experts work together to deeply integrate technology and business to form intelligent solutions, and simultaneously improve the construction of intelligent technology platforms and solution platforms to form the ability to expand AI applications on a large scale.

On October 21, 2019, the winners of the 9th "Wu Wenjun Artificial Intelligence Science and Technology Award" were officially announced. Dr. Xiao Jing won the Wu Wenjun Artificial Intelligence Outstanding Contribution Award. This award is also known as "China's highest award for intelligent science and technology" and represents a symbol of honor in the field of artificial intelligence.

Dr. Xiao Jing graduated from the University of Science and Technology of China with a bachelor's degree and was recommended to the Institute of Automation of the Chinese Academy of Sciences for a master's degree. He then went to Carnegie Mellon University to study for a doctorate in computer science, where he studied under Professor Takeo Kanade, a member of the National Academy of Engineering of the United States, and focused on computer vision. After graduation, Dr. Xiao Jing worked at Epson and Microsoft, leading the team to conduct algorithm research and development in multiple fields until he joined Ping An in March 2015.

As an AI expert, Dr. Xiao Jing talked about the four elements of AI in the conversation:

The four elements of data, computing power, algorithms, and scenarios complement each other. Only by closely integrating technology and business and creating and implementing comprehensive intelligent solutions can the true value of artificial intelligence be realized and actual business pain points can be effectively solved, rather than simply showing off skills.

In recent years, artificial intelligence has attracted a large number of investors as a hot topic, but it is also facing the situation of "difficulty in scene implementation". Ping An Technology focuses on commercial fields such as finance, education, buildings, security, and retail. In the view of Dr. Xiao Jing, AI empowering traditional industries faces a series of challenges, and the intelligence of traditional industries should not be a simple technology application.

He noted:

Different enterprises have different capabilities and mechanisms in technology and management, and the progress of intelligent operation will be different. For example, large enterprises tend to build their own capabilities, while small and medium-sized enterprises are more suitable to introduce ready-made complete solutions. However, the overall intelligent trend of the industry should be vigorous. Its development focus is not on the pure pursuit of technical indicators, but on the breakthrough of actual business value.