Principle of Human-Machine Q&A System

102 0 1 0

With the rapid development of new technologies such as the mobile Internet and artificial intelligence, in the fast-paced and digital era, the self-service willingness of customers will continue to rise, while the traditional channels (telephone, mail) will further decline. At the same time, how to achieve an efficient solution to the growing business and problems? It is can’t rely on a lot of manpower to solve massive technical problems. The advantages of AI technology in solving high-frequency problems and reducing labor costs make intelligent services a hot spot in the industry.

Huawei Enterprise GSC, cooperating with 2012 Lab, Artificial Intelligence Enablement Department strives to build the enterprise intelligent service assistant iKnow, dedicated to solving user problems simply and efficiently.

iKnow has the ability of speech recognition, natural language processing, intent recognition, intelligent search, etc. For the known problems, the system directly gives the answer. For the unknown issues, you can click Live Chat, then our engineers will provide you with global multi-language services in 7*24 hours. Engineers use intelligent platform assistants to improve the efficiency of finding knowledge and quickly solve user problems.

Then, let’s have a look at the iKnow Q&A system.




So, how does iKnow implement smart Q&A?

The Q&A system is divided into two parts: the knowledge base and the Q&A engine.


iKnow analyzes user problems and matches the data in the knowledge base through natural language processing. When matching to the knowledge base, the answers will be ranked by TOP and give the final answer. If the answer in the knowledge base is not matched, the search engine will be used to search the related website content and return the result. If there are no answers the user wants in the knowledge base and website results, click Live Chat, iKnow will provide you a 7*24-hour global multilingual service.

Next, let’s briefly explain the knowledge base:

Knowledge Base: The Q&A system can only answer the knowledge already existing in the knowledge base. Therefore, for the practical effect of the Q&A system, the construction of the knowledge base is largely more important than the technology itself. The knowledge base currently used by the industry is in the form of Q&A form and knowledge graph.


star QA-based knowledge base:

The Q&A system can only answer the knowledge already existing in the knowledge base. So as long as there is enough content in our database, we can support such a Q&A system. For the practical effect of the Q&A system, the construction of the knowledge base is much more important than the technology itself. The coverage, richness, diversity, and even redundancy of the knowledge base are all very important. However, data-supported searches are too simple and only support ordinary fuzzy searches.

To build a knowledge base, you first need to collect as much data as possible about each aspect, which are the source materials for building a knowledge base

A common practice is to unify the various data into the form of a “question-answer” pair (“QA"” for short). This is the QA-based knowledge base.

star knowledge base based on knowledge graph:

Structure a class of knowledge into entities, relationships, entities and entities, attributes, values, and present the answers to users through processes such as knowledge representation, knowledge modeling, knowledge extraction basis (data collection, entity identification, relationship extraction, event extraction), knowledge fusion, knowledge storage, etc. This form of structuring data into entity-relationship-attributes is the knowledge graph. The biggest advantage of the knowledge graph is that it is very powerful in describing data.

Entity, for example:U1900, U1980

Attribute, for example:Functions | Feature, Ports | Protocols, Performance | Capacity

With this definition, you can build a knowledge graph.


As shown in the figure, eSpace U1900 is an entity, and Ports | Protocol is an attribute. The system analyzes the user’s problems, including the processes of discovery, extraction, classification, and completion, convert it into a structured form of query, and then query the answer in the knowledge graph to get all the values of eSpace U1900 ports | protocols

  • x
  • convention:


You need to log in to reply to the post Login | Register

Notice Notice: To protect the legitimate rights and interests of you, the community, and third parties, do not release content that may bring legal risks to all parties, including but are not limited to the following:
  • Politically sensitive content
  • Content concerning pornography, gambling, and drug abuse
  • Content that may disclose or infringe upon others ' commercial secrets, intellectual properties, including trade marks, copyrights, and patents, and personal privacy
Do not share your account and password with others. All operations performed using your account will be regarded as your own actions and all consequences arising therefrom will be borne by you. For details, see " Privacy."
If the attachment button is not available, update the Adobe Flash Player to the latest version!
Login and enjoy all the member benefits

Login and enjoy all the member benefits