亚信科技

On May 16th, AsiaInfo (stock code: 01675.HK) held a product launch conference for the “YuanSi” industry big model in Beijing, which officially unveiled of the company's new general AI and cognitive enhancement platform, TAC MaaS, as well as three industry big models and eight cognitive enhancement tools. This announcement signified that AsiaInfo has entered the fast track of empowering vertical industries based on “full-stack AI”. 

Figure: Launch Site 

Wen Ku, Chairman of the China Communications Standards Association, Gao Nianshu, Executive Director and Chief Executive Officer of AsiaInfo, Wei Kai, Deputy Director of the AI Research Institute of CAICT, and Song Qi, Vice General Manager of China Telecom Cloud Network Operation Department, Liu Hong, Deputy General Manager of China Mobile Information Technology Center, and Lou Yu, Deputy General Manager of China Unicom Digitizer Department, attended the conference and delivered speeches. Leaders, experts, and hundreds of professionals from Institute for AI Industry Research, Tsinghua University, China Telecom Data Development Center, China Unicom Software Research Institute, Henan Mobile, Alibaba Cloud, Baidu AI Cloud, and Intel also witnessed this moment together.  

Figure: Guests Delivering Speeches 

 Only a Great Big Model Can be Tailored to Align with the Specific Needs of Various Industries! 

“We do not develop a general big model; rather, we are focusing on creating exclusive big models tailored for various industries.” 

In the speech, Gao Nianshu identified three key challenges impeding the widespread adoption of big models: the discrepancy between the applicability of general big models to vertical industries, the lack of unified management leading to redundant construction, and the shortage of methods and tools for rapid implementation. After 31 years of practice in numerous projects, large and small, AsiaInfo has established itself in real-world scenarios in vertical industries such as telecommunications, government affairs, energy sources, and transportation. Taking into account factors like “industry-specific, data security, continuous iteration, and comprehensive cost,” AsiaInfo has focused on three levels - “platform, model, and tools” - to establish an XGPT product system for industry cognitive enhancement, which involves building one big model capability foundation, training M industry-specific models for specific industries, and providing N out-of-the-box cognitive enhancement tools.  

Figure: Gao Nianshu Delivering a Speech 

Gao Nianshu stated that AsiaInfo will continue to actively participate in the development and improvement of artificial intelligence technologies, constantly conduct experiments, and promote the innovative applications of big model technology in enterprise businesses. In addition, AsiaInfo will collaborate with partners on the “product + service” model to accelerate the empowerment of enterprise applications and jointly explore the limitless possibilities of artificial intelligence. 

Figure: Wen ku, Gao Nianshu, Wei Kai, Song Qi, Liu Hong, Lou Yu, Ouyang Ye, Liu Yunxin, Zhang Yu Attended the Joint Release Ceremony 

Leveraging “Full-stack AI” to Create New Productivity 

During the product demonstration time, Dr. Ye Xiaozhou, Chief Scientist of AsiaInfo, and Chen Sheng, Deputy General Manager of the Mobile Support Business Department, provided detailed introductions to one general AI and cognitive enhancement platform, three industry big models, and eight cognitive enhancement tools. 

Figure: Product System Published in this Release 

One General AI and Cognitive Enhancement Platform: TAC MaaS 

YuanSi · General AI and Cognitive Enhancement Platform TAC MaaS is an intelligent MaaS foundation and a one-stop enterprise-level AI platform developed by AsiaInfo. It provides full-stack AI technology and out-of-the-box MaaS (Model as a Service) services, effectively bridging the gap between general big models and industry applications, and addressing issues such as siloed operations and maintenance and low resource utilization during AI development and service delivery. 

Three Industry Big Models 

YuanSi · Programming Big Model performs very well in various industry programming tasks as it can support over 10 programming tasks tailored to industry-specific scenarios, such as code comment generation, code summarization, and unit test generation. 

YuanSi · Autonomous Network Big Model provides intelligent capabilities of network's full life circle, including real-time network status analysis, network fault identification, and network complaint handling, facilitating the evolution of communication networks towards higher-level autonomy. 

YuanSi · Intelligent Operation and Maintenance Big Model provides comprehensive intelligent operation and maintenance services such as customer complaint handling, operation and maintenance work order processing, knowledge-based Q&A, and decision execution. 

Eight Major Cognitive Enhancement Tools 

YuanSi Huijian · Business Intelligent Handling Assistant (ChatCRM) is a comprehensive benchmark AI+ application leveraging big model technology that includes intelligent capabilities such as answering questions, assisting in handling, self-service handling, recommendations, reminders, document generation, and casual chats. It paves the way for future unmanned business hall in terms of technical capabilities. 

YuanSi Huijian · Autonomous Network Assistant (AN CoPilot) has been integrated with the dedicated big model for autonomous networks, network cognitive enhancement services, and an intelligent toolkit for network operation and maintenance. It serves as an accelerator and catalyst for the evolution of autonomous networks towards higher levels. 

YuanSi Huijian · Business Intelligence Enhancement Analysis Toolkit (ChatBI) is designed and positioned as a comprehensive intelligent data analysis assistant for one-stop data retrieval, visualization, and utilization. By leveraging big language models, enhanced analytics, and domain knowledge integration, it enables business personnel and data analysts to rapidly acquire, analyze, and present data, providing intelligent analysis services for managers. 

YuanSi Huijian · Knowledge Chatbot (ChatBots) refers a multi-engine knowledge dialogue software tailored for industries such as communications, transportation, finance, government, and enterprises. It can be widely used in various scenarios such as intelligent customer service, marketing recommendations, and knowledge Q&A, supporting enterprises in digital transformation, cost reduction, and efficiency enhancement. 

YuanSi Huijian · Turing Programmer (TuringCoder) possesses three capabilities: cross-database Code Search, Code Explain, and Code Complete based on big model technology. It efficiently generates and optimizes code from natural language, while also providing functions such as BSS engineering Q&A, unit test and generation, code style checking, code log cleaning, and code security checks, enhancing development efficiency, unleashing the creativity of developers, and supporting enterprises in digital transformation. 

YuanSi Huijian · Intelligent Operation and Maintenance CoPilot (Ops CoPilot) provides enterprise operation and maintenance expert services based on a dedicated big model for operations and maintenance, helps personnel quickly identify faults, efficiently obtain handling strategies, and automatically generate self-healing scripts for faults. 

YuanSi Huijian · Task Orchestration Automation Tool (RPATaskGPT) has been deeply integrated with big models and automation technology that it can automatically understand user intentions through conversational interactions and intelligently orchestrate the generation of robotic processes to complete task execution and provide feedback. This helps to create a new generation of “zero-threshold, fully automated, ultra-intelligent” digital employees. 

YuanSi Insight · Data Management CoPilot (DMG CoPilot) focuses on providing enterprises with intelligent data management services. It injects intelligence into data management activities in areas such as intelligent data development, intelligent governance, and intelligent operations, achieving a comprehensive improvement in management efficiency. 

Figure: Dr. Ye Xiaozhou and Chen Sheng Introducing Related Products

The “White Paper on AI Native Technology Reconfiguration Software Products” Has Been Released with Significant Impact

As industry big model drives the rapid transition of software systems in various industries from the “+AI” intelligence injection model to the “AI+” reconfiguration model, to comprehensively and effectively address the challenges faced by enterprises in digital intelligence transformation, AsiaInfo, Tsinghua University, and Intel have jointly released the “White Paper on AI Native Technology Reconfiguration Software Products”, which systematically expounds on the perspectives and practices of building AI-native applications based on industry big models.

Figure: Welcome to Scan the Code to Download the “White Paper on AI Native Technology Reconfiguration Software Products”

Partners Provide Insights from Their Own Experience and Share their Practical Knowledge in AI

During the keynote speech session, representatives from the three major telecommunication operators, Institute for AI Industry Research, Tsinghua University, Alibaba Cloud, Baidu AI Cloud, and Intel, based on their respective business practices, shared their latest research achievements and application practices in the field of artificial intelligence. They commended AsiaInfo's exemplary contributions to the field of AI and expressed their anticipation for future collaboration.

Figure: Guest Delivering a Speech

After the conference, Dr. Ouyang Ye participated in a joint interview with 16 media outlets, including Communications Weekly, National Business Daily, and BK Media, and engaged in in-depth discussions on the development tendency of big model technology, business application scenarios and effects, and evaluation indicators for industry big models.

Figure: Dr. Ouyang Ye Was Interviewed by the Media

The era of industry digital intelligence has arrived, and big models have come out on top of this trend. AsiaInfo's industry big models are designed for specific industries and are built on three core values: “Our mission focuses on enhancing industry understanding, building tool ecosystems, and providing model-as-a-service service”, providing enterprises with a one-stop suite of AI solutions. In the future, AsiaInfo will collaborate closely with industry-university-research partners to enhance its “product + service” system and continue to provide efficient and user-friendly “full-stack AI” to government and enterprise customers across various industries.