Ping An Brain

Ping An Brain is a comprehensive AI technology service platform that integrates deep learning, data mining, biological recognition and other technologies to provide standard AI products and comprehensive AI solutions. It includes the product modules such as Euler graph (the first intelligent industrial chain analysis platform), Ping An Smart Voice (a product recognizing people by voice), Ping An Crowdsourcing, Ping An Robot, and Crius Smart Recommendation System.

Value Proposition

Professional Machine Learning Platform

With the capabilities of super-computing and simultaneous million-level data processing in seconds, it carries out large-scale, multi-dimensional and multi-level deep learning model training and calculations.

Diverse AI Algorithm Models

Equipped with AI algorithm models such as biological recognition, disease prediction and image recognition model with world-leading accuracy rates.

Standardized AI Products

It provides standard AI products covering financial risk management of enterprises, decision-making assistance in public health, smart financial stores, etc., which can be quickly implemented to add business value.

Comprehensive AI Solutions

Covering six major service sectors of marketing, operation, risk management, decision-making, service, and prediction, it meets the smart requirements for fintech, medical health and urban services.

Product Modules

OLATOP
Crius Smart Recommendation System
Ping An Crowdsourcing
Ping An Smart Voice
Ping An Robot

As the first smart industrial chain analysis platform in the industry, with the supply chain and industrial chain as the inner business logics and reflecting the real business status as the basis for organizing data, it visualizes the entire business operation process involving enterprises, industries, people and events to provide smart and comprehensive solutions for enterprises in risk management, investment, marketing and governmental public services.

It provides smart precision marketing services such as profiling analysis, client grouping and smart recommendation. Equipped with a machine learning platform, it provides clients with integrated machine learning services of data, algorithms, and computing power. With rich big data mining tool components, the system can perform independent modeling analysis; and it provides various open API and real-time interface calls, outputting call results in a custom manner.

The original "OCR + Crowdsourcing" approach allows the digitalization of 99% receipts and reports. By means of accurate image correction, target detection and template self-generalization technologies, OCR can identify a receipt within 0.3s automatically with an accuracy rate of 99%, while manual review and quality inspection completed through the crowdsourcing platform, thus forming a complete electronic solution to complete the picture-to-text input efficiently.
Scan the QR code to download mobile App for free experience of Ping An Crowdsourcing product.

With the 100 million-level voice data training and original fusion algorithm model, it can accurately identify the voiceprint features of speakers and defend recording synthesis, thus reducing the risk of financial fraud; currently, the identification accuracy of the models of Ping An Smart Voice has reached 99.8%, thus saving time for financial identity verification, intercepting fraud and reducing the risk of information leakage. It has been used in more than 100 business scenarios of more than ten business subsidiaries and external clients of Ping An Group.

Combining AI and IoE and equipped with a self-learning business Q&A engine, improves business efficiency and guarantees information security. Dr. An, the smart customer service robot equipped with a self-learning business Q&A engine, has an active customer welcoming success rate over 96%. The small and cute desktop robot has an automatic audio & video recording function which can effectively improve business efficiency.

Success Stories

Lufax

The product has provided comprehensive precison marketing solutions for online financial products to improve the customer conversion rate significantly.

Ping An Property & Casualty Insurance

Auto insurance image data are mined with the deep learning technology to truly enable automatic online damage assessment.

Chongqing Center for Disease Control and Prevention

Through cooperative research and development of AI + big data influenza prediction model, it assists in public health decision-making.

Center for Disease Control and Prevention

It has optimized the influenza prediction model and achieved real-time prevention and control of infectious diseases.

Use Cases

Smart Medical Care

Influenza prediction model: Outputs urban influenza index and disease trend heat maps of different locations based on LBS, thus giving alarm on influenza outbreak trends timely and providing prevention and control suggestions.

Chronic disease risk prediction model: By integrating urban disease data from over 10,000 dimensions, it can predict risks of 45 chronic and frequently-occurring diseases, saving over 50% potential public health expenditures.

Medical image recognition model: Utilizing image recognition and deep learning technologies, it can conduct location, recognition and shape judgment of diseases for medical images of multiple diseases in multiple types.

Smart City

Smart site selection: Through heat map analysis, it displays the comprehensive heat degree of current sites. Focusing on the characteristics of hot sites, it provides optimization suggestions accurately through analyzing location advantages, population flow surrounding data in combination with the characteristics and advantages of current site. It is able to maximize the potential of the location and improve the efficiency of the service reach.

Smart Scheduling

It divides service zones and arranges the position of personnel smartly. With the factors such as time, weather, traffic, personnel and accident considered comprehensively, the insurance survey personnel can arrive at the designated place on time within a limited time.

Smart Customs

It adopts data mining to model and analyze historical data, and identifies the feature combination of new suspicious persons through the decision tree model, so that the association among passengers' features is deeply explored to construct a smuggling group relationship network. In this way, suspicious accomplices can be determined based on the captured smugglers, thus increasing the discovering rate of features of suspicious persons.

Smart Finance

Smart damage assessment: With the deep learning technology, the auto insurance image data can be mined for assessing damages within seconds after taking and uploading photos, as well as auto insurance survey and damage assessment, component matching, model identification, identical vehicle detection, etc.

Early financial warning signals: Through data acquiring, integrating and signals generating, it gives feedback and manages risks timely; and with the 24-hour verification function, it can automatically integrate new data sources on the same day and update calibration models, thus improving risk prediction accuracy.

Smart Environment Protection

It helps local governments integrate environmental data through the construction of environmental big data center and Green Credit Index with the purposes of achieving multi-dimensional integration, multi-level sharing and global view, as well as real-time monitoring of environmental information and multi-dimensional environmental credit rating.

Through the ecological resource management and Ping An Tianeye, it provides comprehensive resource data analysis, resource governmental management and public resource service for government departments, and conducts monitoring of air and space integration to achieve integrated smart analysis of identification, location and prediction.

Through energy-saving IoT, it provides services such as smart sensing, custom scenarios, ultimate energy-saving.