Abstract: Today’s intelligent voice assistant, can help us complete some routine actions in daily life. Similarly, in the enterprise, intelligent question answering robots are playing the same role.
This article is shared from the Huawei Cloud community “[billion code leading, Yunju Jinling] Huawei Cloud MVP Cheng Yun: Knowledge-based Transformation, Ultimately to empower the line”, author: We are all cloud experts.
“Hey Sir, how’s the weather today?”
“Little Kitty, help me turn on the TV.”
Today’s intelligent voice assistants can help us complete some routine actions in daily life. Similarly, in the enterprise, intelligent question answering robots are playing the same role.
The pace of machine learning has never stopped
Cheng Yun, MVP of Huawei Cloud, has been studying computer software development since university. After graduation, he went to China’s largest listed communication equipment company to engage in system development and management.
In daily work, Cheng Yun found that the products faced with a large number of the same consulting problems, resulting in duplication of work. He wondered if he could build an intelligent question-answering robot for b-side companies, “like the internal Siri of the enterprise,” which would free up a lot of human resources and create more value.
At that point, Cheng was approached by an ARTIFICIAL intelligence startup called Yunwen, whose founder told him that what YunWen wanted to do was use robots instead of humans to answer repeated questions. The two hit it off, and Cheng yun chose to join Yunwen as CTO.
Since then, Cheng has devoted himself to the research and development of intelligent question-answering robots. After a period of hard work, he and his team have made good progress in intent recognition, named entity extraction, machine reading comprehension and other aspects of NLP. Based on this, they won the cooperation with Jiangsu Electric Power, and the two sides jointly developed the electric power industry’s first set of online customer service system “Electric Doctor”.
Dr. Electric mainly through the form of a question and answer, intelligent understanding, accurate positioning of the knowledge points asked by users, through the interaction of questions and answers with users, 24 hours real-time online customer demands, to provide users with knowledge points about electricity management, business, electricity inspection, customer service and other aspects.
During the construction of “Doctor of Electricity”, Cheng Yun first helped build the service hotspot analysis system and assisted manual to solve the sorting and classification of work orders. Secondly, solve the external customer consultation problems to release manpower; Finally, the intelligent knowledge center is developed to build the overall knowledge management and consumption platform of the power marketing center, and the intelligent knowledge management system and knowledge service scene are gradually built.
From scratch in the center of the intelligent knowledge construction process, the cloud to lead the r&d team with maintenance personnel to analyze the business knowledge management framework, carding electric files and knowledge system, build the tens of thousands of electricity encyclopedic knowledge, support 40000 + users daily knowledge consumption, Dr Honing electricity to the precise effect of C end user service, improve service, He has also co-authored many inventions in the fields of natural language processing and knowledge mapping.
“Machine learning never stops. It doesn’t stop at questions and answers.” Cheng yun said that the work mode of intelligent customer service developed by him and his team has gradually shifted from supporting users and customer service to supporting decision-making. Through textual semantic analysis and monitoring of multi-dimensional interactive data such as service records, customer evaluations and work order contents, we can gain insight into customer demands, explore business opportunities, find product defects and so on, and find more business value from the data level.
Electrical maintenance, but also to see industry experts experience
The question answering robot solves the problem of intelligent customer service, but there is a key module in the transformation of power enterprises, that is, intelligent management of power equipment. Power equipment continuously generates data for a long time, with large data volume, multiple types and serious data island, resulting in the value of the generated data can not be reflected. Expert experience in the use, maintenance and overhaul of power equipment is often scattered in maintenance manuals, work orders, repair records and other documents, resulting in low efficiency of industry expert experience inheritance, long training period for new employees, and affecting the production efficiency of enterprises.
In the power industry, there is a complex causal relationship between the appearance of equipment failure and the nature of the failure. The industry experience determines the decision of fault analysis, maintenance schedule and maintenance method.
All these need to be solved from the whole life cycle of knowledge, including knowledge demand perception, knowledge acquisition and warehousing, knowledge consumption and application, knowledge training and learning.
Three years ago, a summer cloud and team hand in hand together to build the wisdom of the electricity power knowledge map project officially started, the project is named after “institute of electrical drops,” with a marketing business application system and mobile terminal two service channel, for the provincial electric power marketing staff to provide knowledge, intelligent query service file.
Nowadays, “e-school” has gradually expanded into a three-dimensional intelligent interactive training platform integrating intelligent learning, intelligent examination and intelligent training. The data performance is remarkable. Only the intelligent knowledge center has nearly ten thousand monthly active users and nearly ten thousand daily intelligent knowledge consumption.
Knowledge-based transformation is not a slogan
“Knowledge-based transformation is not a slogan, but ultimately a first-line business scenario. We have an absolute advantage in NLP and knowledge mapping technology innovation, and companies have a lot of demands in business innovation.”
Cheng Yun believes that the direction of knowledge-based transformation in the future is to penetrate into various industry segments. By building the graph knowledge base of various industries, the intelligent search and question answering are more accurate, knowledge reasoning and personalized knowledge recommendation are supported, and the knowledge can truly empower the front-line personnel.
Cheng Yun and Huawei Cloud became close friends in the process of industrial knowledge-based transformation. Talking about the cooperation with Huawei Cloud, he said that one of the difficulties in the construction of many government and enterprise projects is the problem of data information security. At this time, I contacted Huawei Cloud, and the two sides hit it off. Through huawei Cloud’s professional enterprise data protection overall solution, we take data protection measures for government, enterprise and industry encryption applications and sensitive information. On the other hand, combined with the EI capability of Huawei cloud, we will continue to dig into more possibilities of product application in various industries.
Finally, around the knowledge-based transformation of the enterprise, cloud also Shared their three plans, “one is to constantly for intelligent with the latest technology to improve existing products, the second is the more in-depth research on mining business scene, focusing on solving the practical problems more, three is the most advanced technology to enable intention scene, have the courage to make bold attempt to breakthrough.”
Click follow to learn about the fresh technologies of Huawei Cloud