This project belongs to the field of computer communication and artificial intelligence technology. The integration of artificial intelligence and the real economy is already a national strategy. In addition to the integration, expression, identification, understanding and search applications of various types of information resources for needs and applications, the core issue needs to pay attention to the standardization of artificial intelligence technology in the actual implementation of advanced technologies. problem; Standardization work plays a fundamental, supporting and leading role in the development of artificial intelligence and its industry. Through the development of standardization, the research and development of artificial intelligence technology can be made more oriented to actual business needs, form a knowledge accumulation and reuse mechanism, and save advanced technologies. The cost of practical application. As practitioners of artificial intelligence technology, telecom operators have urgent needs for natural language processing technology in their various businesses. However, there are the following problems in the practical application of this technology: 1) The accumulated experience of corpus and models in industry fields cannot be shared with the same industry or related industry applications, resulting in high individual application development costs and complex business models;2) The operation, maintenance and optimization costs of individual artificial intelligence systems are high, and there is no knowledge accumulation and reuse mechanism. This project conducted application research on artificial intelligence technology in multiple industries. On this basis, it explored and proposed a basic capability development architecture for artificial intelligence for telecommunications multi-service fields, and completed relevant standards and specifications. 2. Main technological innovation contents 1. Standardized development required for multi-service applications of artificial intelligence technology in telecommunications. Facing the application needs of artificial intelligence in multiple business fields, we abstract the common needs of artificial intelligence technology represented by natural language understanding, and build the standard system required for the output of basic artificial intelligence capabilities with ldquo; open platform for speech recognition and semantic processing capabilities rdquo; as the core. Based on the results of this project, he took the lead in developing a national standard in the fields of artificial intelligence and natural language processing technology, and served as the main editor to complete seven national standards; took the lead in establishing an ITU international standard and jointly proposing two international standards; completed and approved a batch of CCSA industry standards; as the main editor, he participated in the compilation and completion of 7 standards for the National Artificial Intelligence Overall Group. 2. Semantic understanding and technological innovation. Combined with specific business scenarios, based on Semantic Web technology and ontology theory, conduct research on semantic feature extraction technology of short/small texts, and design business-oriented semantic understanding, interaction rules, templates, etc. Applied in: 1) WeChat customer service, which has completed the design of more than 20 scenarios and realized the human-computer interaction application of Shanghai Telecom's WeChat customer service;2) SMS customer service, which has established seven types of understanding scenarios and configured 2155 services;3) IPTV automatic Voice control: Designed more than a dozen categories of semantic understanding interactive scene templates, including on-demand, review, movies, TV series, and channels. 3. Innovation in speech recognition technology. 1) In response to the quality inspection business needs of the call center, the batch speech recognition and transcription accuracy rate of more than 82% of compressed format recording files has been achieved; 2) Aiming at the speech recognition results for information entities in professional fields, such as people's names, place names, professional nouns, etc., the joint application process of speech recognition and semantic understanding is designed, and the corresponding entity speech recognition and interaction logic modules are developed, realizing the use of secondary entity semantic understanding. Speech recognition applications have improved the accuracy of named entity recognition in the business field. It has been applied in Shanghai Telecom's automatic switchboard, realizing duplicate names, homophones, Fuzzy query confirmation and interactive logic and other functions. 3. Application promotion and benefits are applied to Shanghai Telecom customer service, Haobai, IPTV voice control, WeChat customer service, SMS business hall, NOC knowledge base, natural language understanding, voice customer service quality inspection, voice big data mining, call center human-computer interaction, public opinion analysis and other fields, the economic benefits related to this project exceed 500 million yuan. 4. Nearly 20 patents have been applied for for this project. Currently, it has obtained 8 authorized patents and published ten papers.