HONG KONG and SHANGHAI, April 10, 2018 /PRNewswire/ -- Ping An Brain under Ping An Insurance (Group) Company of China, Ltd. (hereafter "Ping An" or "the Group", HKEx: 2318; SSE: 601318) was crowned the winner of "Power Laws: Detecting Anomalies in Usage" at the International Artificial Intelligence ("AI") + Environmental Protection Competition. The achievement is another world-class award that Ping An has won in the area of AI, following their international awards for the smart reading technology of Ping An Medical Imaging Assistant ("PAMIA") and AI-based musical composition technology. Recently, Ping An's technological capabilities in the field of AI have become more mature, laying a solid foundation for the further development of Ping An in the field of smart environmental protection and smart city services.
The International AI + Environmental Protection Competition was jointly organized by Schneider Electric, a global leader in energy management, and DrivenData, an internationally-renowned data competition firm. The highly anticipated competition is hugely influential in the field of environmental protection. The theme of the competition was "Energy Conservation and Environmental Protection". The competition lasted for two months, with over 1,700 teams taking part, including internationally-renowned universities and research institutes such as Nanyang Polytechnic in Singapore and AGL Energy in Australia.
The competition aimed to provide energy-saving solutions in the field of environmental protection. The data used in the competition came from 187 unknown buildings from all over the world, with more than 40 million pieces of real electricity consumption data accumulated over the course of 7 years. The greatest challenge was finding unmarked anomalies in electricity consumption amongst such large volumes of data, as well as accurately identifying anomalies that are most relevant to the final energy consumption from a large number of complex types of electricity consumption anomalies. Participants used this data to build models, predict future energy consumption and monitor abnormal conditions in order to optimize energy adjustment and provide solutions.
During the two-month long competition, Ping An Brain combined semi-supervised learning, unsupervised learning and other AI technologies with specialized electric energy models to create a set of smart environmental protection and energy conservation anomaly detection algorithms, enabling them to claim the top spot just one month into the competition.
Leveraging its leading AI technology, Ping An Technology found significant anomalies in more than 400,000 dimensions and accurately reflected the actual circumstances of electricity consumption. After several rounds of data analysis and verification based on actual circumstances, tens of thousands of anomalies most relevant to the actual energy consumption were accurately identified, and the data model with the highest prediction accuracy was established. Such algorithms could identify abnormal energy usage behaviors and provide solutions for these anomalies. The anomalies in electricity consumption detected in this competition could help these buildings save 15-30% of energy wasted in real life.
The data shows that scientific and highly efficient anomaly monitoring can greatly improve the utilization of energy resources, curb waste, and reduce emissions. The use of advanced technologies such as AI to assist in energy monitoring will improve research and application in the field of smart environmental protection. Through AI + technology, Ping An's "Smart Environmental Friendly Cloud" helped to build a socially-beneficial interactive platform, an environmentally friendly big data platform, and a shared business platform service capability. It is designed to help the government's functional departments in promoting the development of smart cities and green living for the public.
In addition to AI + environmental protection, Ping An has established comprehensive scenarios for AI applications such as smart perception, prediction, risk management and services. At the technical level, Ping An's AI and deep learning technologies have reached world-leading levels. Among which, Ping An's face recognition technology has an accuracy rate of 99.8% and its voiceprint recognition technology has a text-dependent accuracy rate of 99.7%. In disease forecasting, Ping An has achieved precision rates of over 90% for influenza and hand, foot and mouth disease and 92% for chronic obstructive pulmonary disease. In addition, its medical image reading technology ranked first in lung nodule detection and false positive reduction with precision rates of 95.1% and 96.8% respectively according to test results of Lung Nodule Analysis ("LUNA"), an international authoritative assessment in the medical imaging field. Ping An's AI-based musical composition technology won the first place in the international "AI-generated music challenge" contest.
At the application level, Ping An provides AI-based functionalities such as photography and video recording, personal loan risk management, SME loan approval and customer services. Ping An's face recognition technology has been used to identify over 1,000 million people in over 200 scenarios within or outside the Group such as financial services, medical services, daily-life services and security. Ping An's voiceprint recognition technology has been applied to over ten scenarios such as app login, identity verification and blacklist identification. Ping An's micro-expression recognition technology can identify customers' smallest facial actions via remote videos in real time, detect frauds and give risk warnings. The technology has been applied to financial business activities such as Puhui Business's loan approval.
This breakthrough in AI + environmental protection indicates that Ping An is continuing to deepen and extend its efforts in the area of AI and smart city services. Forging ahead, Ping An will continue to pursue innovations in fintech, healthtech and smart city services under a two-pronged strategy of "finance + technology", as well as tap into applications of AI technology in multiple scenarios. While applying innovative technologies to enhance its core businesses, it will also gradually build and optimize the eco-system and platforms, externalize its own core technological capabilities and export innovative technologies, striving to become the world's leading fintech company.
About Ping An
As China's first joint stock insurance company, Ping An Insurance (Group) Company of China, Ltd. ("Ping An") strives to become a world-leading technology-powered personal financial services group. Today, it is an integrated, compact, multi-functional financial services group with services that include insurance, banking, and investment. As of 31 December, 2017, the Group had over 166 million retail customers. At the end of December 2017, the Group's consolidated total assets reached RMB6.49 trillion, while equity attributable to shareholders of the parent company stood at RMB473,351 million. Ping An Life and Ping An Property & Casualty are both ranked as the second largest in China, and Ping An Annuity is the top-ranked in China in their respective areas by premium income.
Its subsidiary Ping An Bank is China's first joint stock bank. The Company's key areas of business include investment, with subsidiaries such as Ping An Trust, Ping An Securities and Ping An Asset Management. Furthermore, Ping An strives to develop internet finance, including Lufax, Ping An Good Doctor, Ping An Haofang, E-Wallet, OneConnect, Ping An Healthcare Management Services, Wanjia Healthcare. It has achieved significant growth in both the scale and user base of internet finance. As of 31 December, 2017, the number of internet users of the Company reached 436 million.
Ping An ranked 16th in Forbes' 2017 Global 2000, and it ranked 39th in Fortune Magazine's 2017 Global 500 Leading Companies. Ping An also ranked 61st in WPP Millward Brown's BrandZTM Top 100 Most Valuable Global Brands.
For more information, please visit http://t.sina.com.cn/pingan, http://t.qq.com/pingan or www.pingan.com.cn.
