CN101431573B - Method and equipment for implementing automatic customer service through human-machine interaction technology - Google Patents

Abstract

The present invention provides a method for realizing automatic customer service through human-computer interaction technology and equipment for realizing the method. The method includes: receiving a question input by a user in natural language, and converting the natural language through human-computer interaction technology into information for computer identification; and then output feedback information based on the information for computer identification. In addition, the invention also provides equipment for realizing the method.

Description

Method and device for realizing automatic customer service through human-computer interaction technology

technical field

The invention relates to computer network technology and human-computer interaction technology, in particular to a method for realizing intelligent automatic customer service through computer network technology and human-computer interaction technology and equipment for realizing the method.

Background technique

The customer service center characterized by automatic voice and manual service has been developed for many years in China. Businesses of all sizes, from banks and telecoms to SMEs and governments, have already started the service. Customers have gradually changed from being curious about this form of service to becoming accustomed to it. Now, for an enterprise, having a customer service center is only the minimum requirement. For its customers, the customer service center must be able to provide high-quality services. Under the new situation, the customer service center of the enterprise has encountered the following challenges:

1. The knowledge of customer service representative is comprehensive.

With the intensification of market competition, companies continue to launch new products and services. As an important department to contact customers, the customer service center must know all the products and services provided by the enterprise. Therefore, companies usually conduct special training for customer service personnel in the early stage of launching new products. As the speed of launching new products accelerates, the frequency of such training gradually increases (the cost increases accordingly). Since there is more and more information of this kind, and sometimes it has even exceeded the level that customer service personnel can grasp, customer personnel have to temporarily inquire about relevant information or ask other colleagues who know about it at the time of customer inquiry. On the one hand, this temporary information retrieval makes customers wait longer, and on the other hand, the cost of each call increases significantly. In some cases, the customer's call was transferred to different customer service representatives, at which point the customer had to repeat their question over and over again.

2. Substantial increase in the number of users.

With the development of the enterprise market, more customers and potential customers learn about relevant product and service information through the customer service center. At this time, due to the limitation of the hardware system of the original customer center of the enterprise, customers wait in the queue for too long. At this time, enterprises are faced with a choice, increase investment in equipment (apply for more lines, improve the processing capacity of telecommunications equipment) or lose customers. Usually, the investment and reinvestment of the call center is very expensive, and generally takes 3 to 9 months to implement.

On the other hand, with the launch of various products and marketing activities of the enterprise, users will inquire and understand various problems through the customer service center in a short period of time. This intermittent peak puts great pressure on the customer service center. If the enterprise reinvests in the customer service center to cope with this peak pressure, it will lead to equipment and personnel redundancy during off-peak hours. This redundancy is of great importance to the enterprise. Said to be another cost expenditure.

3. Day and night and holiday service.

Due to the accelerated pace of work, it is inconvenient for many users to use the customer service center of the enterprise during working hours, and they usually choose to use the time after get off work or during holidays. As an enterprise, maintaining a long customer service time means an increase in cost, and for the customer service personnel themselves, working on holidays will bring various management problems.

Under the current application status, enterprises generally have no choice but to upgrade the system and recruit new customer service personnel when encountering the above challenges, which brings a series of additional costs. However, upgrading and recruitment cannot completely solve the problem that customer service personnel do not know enough about the product, so some companies invest huge sums of money to purchase and implement complex systems such as knowledge management software, which in turn brings a series of huge costs.

Contents of the invention

The present invention is proposed based on various deficiencies in the current customer service center characterized by automatic voice and manual service, which is used to provide a method and realization of intelligent automatic customer service through computer network technology and human-computer interaction technology equipment for the method. .

The method and corresponding equipment provided by the present invention can be used to realize the construction of a customer service robot system. The customer service robot system or customer service robot is a brand-new concept, which is based on various artificial intelligences such as natural language processing and human-computer interaction. Based on technology, an automatic customer service system using IM (instant messaging tool) or WEB pages as the form of expression, in which users can chat with customer service robots as if they were talking to human customer service, and get specific domain questions in the natural interaction process solution. More specifically, the customer service robot refers to a logical "virtual robot" based on network (including telephone and mobile device) communication, without its own entity, and its function is to provide users with intelligent automatic customer service. This system can be applied to the corresponding external service departments in the government, enterprises and various public institutions. Efficiency, significantly reducing government customer service costs, reducing human errors, and enhancing user experience.

According to the first aspect of the present invention, there is provided a method for realizing automatic customer service through human-computer interaction technology, the method includes: receiving query information based on natural language input by the user; outputting query information based on the query information based on natural language result. Specifically, this step includes performing semantic analysis on the query information input by the user, and performing semantic matching with the content in the pre-stored knowledge base, and obtaining the question matching the query information or its associated answer. step.

Preferably, when the question obtained by the semantic matching has a corresponding answer, the answer is output as a query result. When the question obtained by the semantic matching does not have a corresponding answer but only has associated query information, then output the query information; then receive feedback information based on natural language input by the user for the query information; and then output the query information The feedback information is used as the user input information and the above steps are repeated until the corresponding answer is obtained and the feedback information is instruction information for the user to instruct to end the query.

According to the second aspect of the present invention, there is provided a device for providing intelligent automatic customer service through human-computer interaction technology, the device includes: a human-computer interaction device for receiving query information based on natural language input by a user; a feedback device , for outputting a query result according to the query information based on natural language. Specifically, the device also includes a semantic matching device, which is used to perform semantic analysis on the query information input by the user, and perform semantic matching with the content in the pre-stored knowledge base to obtain a question matching the query information. , the knowledge base includes a plurality of questions with a certain topology and their associated answers; when a question obtained by the semantic matching has a corresponding answer, the semantic matching device outputs the answer as a query result.

Preferably, when the question obtained by the semantic matching has no corresponding answer but only has associated query information, the query information is output by the semantic matching device; Language feedback information; then perform the above-mentioned semantic matching step according to the feedback information, that is, use the feedback information as user input information, perform semantic analysis on it and carry out semantic matching with the content in the pre-stored knowledge base, and obtain the same The query information is matched with a question, and when the question obtained by the semantic matching has a corresponding answer, the semantic matching device outputs the answer as a query result. The above steps are repeated until the corresponding answer and the feedback information obtained are indication information for the user to end the query.

The advantages of the method and system provided by the present invention are:

1) Reduce the reinvestment of the enterprise in the customer service center, and calmly cope with the large increase in the number of users;

2) Reduce the operating cost of the enterprise customer service center;

3) Reduce the unpleasant experience of customers - waiting time, accuracy of answers;

5) Analyze customers and self-update service capabilities;

6) Provide round-the-clock and holiday service at low cost

By adopting the technical solution of the invention, it can provide problem triage support for the traditional customer service center, so that the traditional service center can focus on complicated and high-value customers. Therefore, in the two channels, enterprises can provide users with more rapid, accurate and standard services. Under the condition of low cost, it provides effective support for the rapid development of enterprises. For enterprises or governments, the use of this leading technology also adds a new image (image of technology and innovation).

Description of drawings

Other characteristics, objects and advantages of the present invention will become more apparent by reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings.

In the drawings, the same or similar reference numerals represent the same or corresponding components.

Fig. 1 is intelligent customer service equipment-user terminal network system diagram according to the present invention;

Fig. 2 is a flow chart of a system method for realizing intelligent automatic customer service according to the intelligent customer service equipment of the present invention;

3 is a schematic diagram of a parallel data table adopted by the knowledge base in the intelligent customer service device according to the present invention;

FIG. 4 is a schematic diagram of a tree structure data representation adopted by the knowledge base in the intelligent customer service device according to the present invention;

Fig. 5 is a device block diagram of providing intelligent automatic customer service equipment through human-computer interaction technology according to a specific embodiment of the present invention;

FIG. 6 is a device block diagram of an intelligent automatic customer service device according to a variation example of a simplified solution of a specific embodiment of the present invention;

FIG. 7 is a schematic diagram of a specific architecture of a server system of an intelligent customer service device according to a specific embodiment of the present invention;

Fig. 8 is a schematic diagram of a specific architecture of a server system of an intelligent customer service device according to a preferred embodiment of the present invention.

Detailed ways

The present invention can be used to build an intelligent customer service device (such as the intelligent customer service device 1 shown in Figure 1), that is, the customer service robot system or customer service robot platform described in the present invention, so as to realize the first aspect of the present invention. The method provided for realizing intelligent automatic customer service.

Based on the intelligent query system-user terminal network system diagram shown in Figure 1, the specific process of the method provided by the first aspect of the present invention will be introduced in detail below, and Figure 2 is the realization of intelligent automatic customer service according to the intelligent customer service equipment of the present invention Flowchart of the system method.

The intelligent customer service device 1 shown in Fig. 1 of the present invention has a network interface, specifically, the network can be various networks that can realize data transmission between the user terminal 2 and the intelligent customer service device 1, including the Internet , mobile Internet, or local area network or other types of networks. In this implementation manner, the network interface may be used to exchange data with the Internet.

In step S210, the user can input information based on natural language into the user terminal 2, and transmit the data of the information input to the intelligent customer service device 1 through the Internet.

Then proceed to step S220, the intelligent customer service device 1 performs semantic analysis on the received information input based on natural language, for example, extracts the semantic keywords in the natural language, and rehearses and combines it into a form suitable for intelligent customer service equipment. 1 Recognize or classify the retrieved sentence fragments, and then the intelligent customer service device 1 will perform semantic matching on the natural language data that has undergone semantic analysis according to the established knowledge base, so as to obtain the semantics of the natural language in the knowledge base and the natural language Matching answers or query information matching the semantics of the natural language.

Then execute the judgment step S230, if after the above steps, the intelligent customer service device 1 has obtained an answer matching the semantics of the natural language input by the user in the knowledge base, the intelligent customer service device 1 will directly The answer is used as feedback information, step S260 is executed, and the information is output to the user terminal 2 through the network, and the user can receive the feedback information.

If after the above steps, the intelligent customer service device 1 cannot obtain an answer that matches the semantics of the natural language input by the user in the knowledge base, then the intelligent customer service device 1 continues to execute the judgment step S240, and judges that in the knowledge base Whether there is any query information that matches the semantics of the natural language input by the user. If there is query information matching the semantics of the natural language input by the user, go to step S241.

Step S241, the intelligent customer service device 1 extracts the query information from the knowledge base, and sends it to the user terminal 2 as feedback information via the Internet.

The user can further input information in natural language according to the inquiry information displayed by the user terminal 2, and the input natural language data will be sent by the user terminal 2 to the intelligent customer service device 1 through the Internet, and then return to step S210 and repeat The intelligent customer service device 1 performs semantic analysis on natural language and performs semantic matching and other steps. .

If after the above steps, the intelligent customer service device 1 cannot obtain an answer matching the semantics of the natural language input by the user in its knowledge base, nor can it obtain inquiry information matching the semantics of the natural language input by the user, the intelligent The customer service device 1 will continue to execute the predetermined default step S250, for example, the feedback information can be output according to the preset default information; or, the received natural language input content that cannot obtain the semantic matching result is sent by email To a pre-set email address, the pre-set email address is, for example, regularly managed by a dedicated human customer service staff; it is also possible to link the human-computer interaction interface to the pre-set address, specific methods include, through instant messaging tools MSN displays a link address to the customer and guides the customer to link to the pre-set address. The pre-set address can be a customer service phone number in charge of a manual customer service person, or a manual customer service in the form of MSN, QQ and other instant messaging tools. , the artificial customer service personnel supplementarily answer some of the questions for which the intelligent customer service device 1 cannot obtain answers matching the semantics of the received natural language.

Normally, when the corresponding answer and the feedback information received by the intelligent customer service device 1 from the user terminal 2 are indication information for the user to end the query, the process ends.

In the above process, the data stored in the knowledge base of the intelligent customer service device 1 can be improved and increased by automatic means or background management.

When the intelligent customer service device 1 receives natural language-based query information input by the user, the query information is usually different from person to person. Specifically, the user can use multiple expressions for the same semantic query information. The knowledge base in the intelligent customer service device 1 may not necessarily cover multiple expressions of query information based on natural language under the same semantics, but after the semantic analysis step of the intelligent customer service device 1, these expressions are different The query information of the query information can usually be corresponded with the corresponding semantic query information existing in the knowledge base, so that the answer associated with the query information or the associated query information can be further obtained in the knowledge base.

The intelligent customer service device 1 of the present invention can automatically store the query information based on natural language input by the user in a variety of expressions that are not available in the knowledge base, and after semantic analysis, it will be the same as that stored in the knowledge base. Semantic query information is stored associatively. This method is an additional method of storage. When adding query information in a new expression mode, it does not affect any content already stored in the knowledge base, which can enrich the content covered by the knowledge base. Various ways of expressing the query information. On the other hand, the stored content can also be used to check whether the intelligent customer service device 1 performs semantic analysis on the natural language-based inquiry information input by the user. For example, the above-mentioned stored content can be checked manually, if If it is found that the natural language-based query information automatically stored by the intelligent customer service device 1 through semantic analysis is not associated with query information with the same semantics in its knowledge base, then the intelligent customer service device 1 can automatically store the query information. The inquiry information is deleted, or it is manually associated with another existing inquiry information in the knowledge base with appropriate semantics for storage. In this way, incorrect semantic analysis of the same query information can be avoided next time.

In the process of information exchange between the intelligent customer service device 1 and the user terminal 2, in particular, there are artificial customer service personnel involved in answering the question that the intelligent customer service device 1 cannot obtain an answer that matches the semantics of the received natural language , the knowledge base of the intelligent customer service device 1 can be controlled to store relevant content through the corresponding background management interface in the intelligent customer service device 1, and the knowledge base can be updated, for example, it can store the received natural language input content , and the semantic matching result of the natural language input content, as well as the number of occurrences; and the received natural language input content for which the semantic matching answer and semantic matching query information cannot be obtained and the feedback information correspondingly replied by the human customer service personnel; In addition, it also includes storing some language vocabulary related to professional fields. Usually, one or more contents stored in the knowledge base can optionally update the knowledge base through replacement or addition.

Specifically, the knowledge base in the intelligent customer service device 1 may adopt a data table whose storage mode is parallel, and in this case, it is divided into two data tables, as shown in FIG. 3 , wherein the first data The table is used to store preset answers that match the semantics of specific questions based on natural language. For example, preset answer a1 is associated with question q1, preset answer a2 is associated with question q2, etc. These are related to user input based on natural language The relationship between the answers matching the questions in the database is parallel; the second data table is used to store the preset query information matching the semantics of a specific natural language, for example, the preset query information a1' is associated with the question q1' , the preset query information a2' is associated with the question q2', etc., and the relationship of these query information associated with the question input by the user based on natural language in the database is also juxtaposed.

In addition, the knowledge base can also use a data table in the form of a tree structure. In this case, the specific natural language in the data table is pre-classified and arranged step by step according to its semantics, as shown in FIG. 4, the tree data The table classifies and arranges the questions entered by users based on natural language level by level. For example, when the questions entered by users based on natural language are semantically analyzed and matched into Q1 questions, there will be several corresponding inquiries in this type of questions Information can continue to be asked or prompted to make the user answer further, such as Q11, Q12, Q13 and other questions to ask the user, so as to finally get the answer semantically related to the question entered by the user based on natural language. Of course, such a tree-structured data table can also directly obtain an answer when a natural language-based question input by a user has a directly associated answer.

The device block diagram of the intelligent automatic customer service device 1 provided by the human-computer interaction technology shown in FIG. 5 will be used below, and the device for realizing automatic customer service provided by the second aspect of the present invention will be described in detail in conjunction with an embodiment.

As shown in Figure 5, according to a specific embodiment of the present invention, an intelligent customer service device 1 that provides intelligent automatic customer service through human-computer interaction technology is mainly applicable to units that provide consultation, inquiry or similar services or units that are responsible for corresponding functional organization. The intelligent customer service device 1 mainly includes a human-computer interaction device 51 and a feedback device 52 , preferably, the intelligent customer service device 5 may also include a background management device 53 .

The human-computer interaction device 51 is mainly used for the received data from the user terminal 2, such as the text input by the user on the user terminal 2 using the above-mentioned Microsoft Pinyin input method based on windows, or the user uses the microphone on the user terminal 2 to The input voice, etc. is converted into information for computer recognition through human-computer interaction technology; specifically, it includes converting the natural language contained in the data into information suitable for identification or classification retrieval by the intelligent customer service device 1, for example, converting The user's voice input is converted into data that the computer can perform semantic analysis on. On the other hand, the human-computer interaction device 51 is also used to send the feedback information from the feedback device 52 to the user terminal 2 via the Internet.

The feedback device 52 generally includes a knowledge base storage device 521 and a semantic matching device 522, wherein the specific representation form of the knowledge base storage device 521 is a relational database in various formats such as DB2, Oracle, or MySQL, Microsoft SQL Server 2000 withSP3/SP4 etc. Answers matching the semantics of a specific natural language and questions matching the semantics of a specific natural language are stored therein. Semantic matching device 522 is used for carrying out semantic analysis to natural language and carries out semantic matching, and this semantic matching device 522 converts the above-mentioned various data through human-computer interaction device 51 (that is, the user inputs through user terminal 2 and sends to human-machine through Internet. The natural language data of interactive device 51) carry out semantic analysis, for example, extract wherein key word, then match it with the answer or the question that is stored in the knowledge base storage device 521 and matches the semantics of specific natural language, and The matching result is fed back to the human-computer interaction device.

Specifically, when outputting feedback information, the feedback device 52 uses the semantic matching device 522 to perform semantic matching on the natural language data converted by the human-computer interaction device 51 according to the database in the established knowledge base storage device 521, so as to obtain the The answer matched with the semantics of the natural language is fed back to the human-computer interaction device 51 . If the semantic matching device 522 does not obtain an answer that matches the semantics of the natural language, it further matches query information that matches the semantics of the natural language in the database of the knowledge base storage device 521. If the query information matches the semantics of the natural language, the query information is output to the human-computer interaction device 51, and the human-computer interaction device 51 sends the query information to the user terminal 2 through the Internet to ask the user a question; then the user The natural language used to answer the query information can be input through the user terminal 2, and sent to the human-computer interaction device 51 via the Internet, and the semantic matching device 522 then performs semantic matching on the feedback natural language data processed by the human-computer interaction device 51. ; Repeat the above steps until the user terminates the input of natural language on the user terminal 2.

In order to make the answer of the intelligent automatic customer service equipment 1 to the user tend to be perfect, the above-mentioned knowledge base storage device 521 can also store the natural language input content received by the human-computer interaction device 51 and the content processed by the human-computer interaction device 51 And the matching result of the semantic matching device 522 is used to update the knowledge base in the knowledge base storage device 521 , the semantic matching result includes an answer or query information matching the semantics of the natural language input by the user on the user terminal 2 .

Since the questions entered by the user on the user terminal 2 are various, for example, some of them have no direct relationship with the consulting service that the corresponding functional department in the unit applying the intelligent automatic customer service equipment 1 is responsible for, or although they are related to the functional department's Consulting services are relevant but the corresponding questions do not yet have corresponding or relevant answers in the knowledge base of the knowledge base storage device 521. In this case, this intelligent automatic customer service device 5 will not be able to obtain the semantics of the received natural language Matching answers, even the query information cannot be fed back to the user according to the semantics of the received natural language. According to a preferred solution of the present invention, when the above-mentioned feedback device 52 does not obtain answers and query information that match the semantics of the received natural language, it can execute a predetermined default instruction, and send a message to the user through the human-computer interaction device 41. Terminal 2 feeds back compensatory content. For example, the feedback device 52 can output feedback information according to preset default information; For example, the preset email address is regularly managed by a dedicated manual customer service personnel; the human-computer interaction interface can also be linked to the preset address, and the specific method includes sending a message to the customer through the instant messaging tool MSN Display a link address and guide the customer to link to the preset address. The preset address can be the customer service phone number in charge of the manual customer service personnel, or it can also be the manual customer service in the form of MSN, QQ and other instant communication tools. The customer service personnel supplementarily answer some of the questions for which the feedback device 52 cannot obtain answers matching the semantics of the received natural language.

In order to continuously improve and increase the knowledge base content stored in the knowledge base storage device 521 during the use of the intelligent automatic customer service equipment 1 of the present invention, the process of information exchange between the intelligent automatic customer service equipment 1 and the user terminal 2 , especially when there are human customer service personnel involved in answering questions that the feedback device 52 cannot obtain an answer that matches the semantics of the received natural language, the knowledge base storage device 521 can store relevant content for updating the knowledge base, such as its Can store the received natural language input content, the semantic matching result of the natural language input content, and the number of occurrences; The feedback information of the corresponding reply by the artificial customer service personnel; in addition, it also includes the storage of some language vocabulary related to the professional field. Usually, one or more contents stored in the knowledge base storage device 521 can optionally update the knowledge base in the knowledge base storage device 521 by overwriting or appending.

In order to realize the above preferred solution, the intelligent automatic customer service device 1 of the present invention may include a background management device 53 for improving and increasing the content of the knowledge base stored in the knowledge base storage device 521 .

The background management device 53 can help the user who uses the intelligent automatic customer service equipment 1 to manage and deploy the functions of the intelligent equipment. Specifically, the background management device 53 can be, for example, a computer terminal connected to the feedback device 52, which supports operating systems such as UNIX, Linux, and Windows, and managers can easily manage the intelligent system through the management background based on WEB. The functions of the automatic customer service equipment 1 include operations such as adjusting the expression form of the customer service system of the intelligent automatic customer service equipment 1 . Concretely comprise: 1), increase, delete and revise of customer service knowledge base; 2), the batch import of customer service knowledge base, for example the operating system that the present invention adopts can support the data import of multiple modes (Excel, Access, text file etc.) , and you can choose options such as overwriting/appending; 3), the processing strategy of the knowledge base, when there is no problem related to the user's question in the database, what strategy to use to deal with it, you can choose the default answer, Email to the manual customer service , Transfer to manual customer service (MSN robot support) and other methods; 4), manage (add and modify) the robot's welcome message; 5), manage (add and modify) the default answer of the robot; 6), modify the manual customer service 7), the log analysis function, you can check the detailed usage of the system, the specific functions are as follows: question content, answer content, IP address, question time, response time; for a certain IP, check the IP question list and answer content; through Log analysis, the system administrator can mark wrong answers, so that the robot can avoid wrong answers when encountering the same question again; provide the ranking of the number of questions asked; provide a filtering function, which can filter out correct answers, wrong answers, alternative answers, Unable to answer etc. At the same time, it provides question time screening, IP address screening, and question content screening; 8), analyzes the list of questions that the robot cannot answer, adds correct questions and answers to the knowledge base, and continuously enriches the knowledge base, thereby improving the intelligence of the robot; 9) , Increase professional vocabulary, enrich the robot's professional field knowledge; 10) Flexible management personnel authority settings.

The above-mentioned background management device 53 is not necessary. The function of this device is to further improve the intelligentization of the equipment provided by the present invention on the basis that the intelligent automatic customer service equipment 1 of the present invention can automatically reply to the natural language input by the customer. However, if the knowledge base in the intelligent automatic customer service equipment 1 is relatively complete, the background management device 53 may not be used. In other words, only the feedback device 52 and the human-computer interaction device 51 can complete the basic solution of the present invention, or the background management device 53 can be removed when the background management device 53 is used to complete the knowledge base. Therefore, as a variation of the embodiment of the present invention, the intelligent automatic customer service equipment may not include the background management device. As shown in FIG. 6 , the intelligent customer service device 5' mainly includes a human-computer interaction device 51 and a feedback device 52, and the feedback device 52 also includes a knowledge base storage device 521 and a semantic matching device 522. In the intelligent automatic customer service equipment 5' of this simplified embodiment, the functions that the above-mentioned devices can complete are identical to the embodiment of the previous equipment.

In addition, as another variation of the embodiment of the present invention, the knowledge base storage device 521 or the semantic matching device 522 described in the above embodiments may be independent of the feedback device 52, and its content or function may be controlled by the background management device. In actual application, those skilled in the art may disassemble or combine the various devices described in the embodiments of the present invention according to requirements and then apply them.

The following will introduce in detail the technical solution for forming a customer service robot according to the present invention by using specific equipment such as a server system according to the above-mentioned content. Before implementing the above-mentioned customer service robot solution according to the method and equipment provided by the present invention, the following preparatory work is usually required:

Preparation of the hardware environment

One or more personal computer terminal servers are used as database and application program servers; when more than one server is used, the database server and application program server can be set separately to improve performance and facilitate network architecture.

The above-mentioned servers can be servers of various brands and configurations, for example, servers of various models provided by manufacturers of brands such as Dell, HP, IBM, Lenovo, and Inspur can be used.

Preparation of the software environment

Corresponding to the hardware environment, the operating system on the server can also have multiple choices, for example, it can use Windows as the operating system, and also can use Unix, Linux, Linux, FreeBSD, Solaris, etc. as the operating system. Usually, the server that is used to realize technical solution of the present invention can install Microsoft Windows 2003 Enterprise operating system, and install other security and performance adjustment software such as firewalls. The database server also needs to install Microsoft SQL Server 2000 with SP3/SP4. In the implementation of customer service software, it is usually necessary to use the management password of the database.

In addition, the database involved in the technical solution of the present invention is a general concept, which may be a relational database in various formats, and is not limited to a database in a specific format. Specifically, when applying the technical solution of the present invention, a commercial database such as DB2, Oracle, etc. may be selected, or a free database such as MySQL may be selected.

Preparation for the network environment

Ensure that the servers provided according to the technical solution provided by the present invention are connected to each other through an intranet (when more than one server is used), for example, each server is formed into a local area network by a switch, and these servers are all assigned a fixed intranet IP address. Since Internet WEB services are to be provided, the application server must have an exit to the external network (an exit bandwidth is reserved in advance). Moreover, the customer service robot constructed according to the technical solution of the present invention needs to use the external network address and the mapped port number (for example, port 80) during implementation.

It should be understood by those skilled in the art that the various preparatory work described above are only used to exemplarily describe the specific implementation manners of the present invention, rather than to limit the present invention. During actual application, technicians may make appropriate adjustments or supplements to any or more of the hardware environment, software environment, and network environment according to actual needs. For example, in addition to the above-mentioned software environment, those skilled in the art can also install supporting software according to needs, or, on the basis of the technical solution of the present invention, install and configure the customer service system built through various servers, specifically, The installation supporting software may include, for example, JDK1.5, Resin-Pro, JDBC components, etc., and the installation customer service system includes an intelligent customer service engine, a natural language processing engine, a management console, and the like.

In addition, the described hardware environment is not limited to the above-mentioned various hardware environments used for example. For example, when a customer service robot constructed according to the technical solution of the present invention is implemented in a certain unit, the various application servers involved, The database server and the like can be integrated in various types of original servers of the unit, and are not limited to the various servers listed in the present invention. Or, the various application program servers and database servers used to construct the customer service robot according to the present invention are composed of separately purchased servers, but these servers can also be used in conjunction with the original server or server system of the unit, for example, the The original unit has an Internet WEB server, which has an outlet to connect to the internet, and this server is used to display the website of the unit to the outside world. In this case, the technical solution according to the technical solution of the present invention is usually used by technicians according to the actual situation of the network. And the Uniform Resource Locator (URL) of the WEB customer service robot of construction is integrated in the existing website of this unit by the mode such as JavaScript script.

Fig. 7 shows a schematic diagram of a system architecture of a customer service robot in a specific embodiment of the present invention. The usual application environment of the customer service robot may be a unit that provides consultation, inquiry or similar services, and it is assumed that it is used in a specific department in a certain unit that communicates with the public.

When constructing the customer service robot system, the server and other specific equipment used may correspond to the devices described above. For example, the human-computer interaction device 71 shown in FIG. 1 or FIG. It is a server responsible for exchanging data with users, such as a WEB server; the semantic matching device 522 shown in Fig. 5 or Fig. 6 can usually be a special server with an intelligent customer service engine; and the knowledge base storage device can usually be A knowledge base server for storage. Of course, in order to make reasonable use of hardware devices such as servers, all the devices shown in FIG. 5 or 6 above can be integrated into a server. On the other hand, more than one server can also be used to constitute a device described above. For example, the above-mentioned feedback device 52 can be composed of a knowledge base server and an intelligent engine server. Therefore, the various servers described below are only examples of a specific application solution of the present invention, and various adjustments can be made according to the above prompts in practical applications.

As shown in Figure 7, the server system adopted in the customer service robot platform architecture according to the present invention includes: a knowledge base server 71, an intelligent customer service engine server 72, and an Internet WEB server 73, and these servers constitute an intelligent automatic customer service device 1. Among them, three servers perform the functions of knowledge base (database) storage, intelligent customer service and data exchange with the Internet respectively. Moreover, as shown in FIG. 7 , these servers are interconnected through cables or an intranet, and can communicate with each other, and any two servers can exchange data with each other.

The user can log in to the website of the above-mentioned government department via the Internet through the user terminal 2, that is, connect to the Internet WEB server 73 shown in Figure 7 through the Internet, and interact with the WEB or various instant messaging tools, such as MSN, QQ, etc. Exchange information with the server system composed of the knowledge base server 71, the intelligent customer service engine server 72, and the Internet WEB server 73.

Specifically, when the user exchanges information with the intelligent automatic customer service device 1 built by the server system in the WEB interactive manner, the user terminal 2 only needs to have a web browser corresponding to the Internet WEB communication protocol, such as Microsoft's IE, this browser is usually integrated in a common operating system such as Windows, without installing or modifying any software, it can easily communicate with the intelligent automatic customer service device 1 . The user terminal 2 can support various existing operating systems, such as Windows, Unix, Linux, and so on.

For users who are accustomed to using instant messaging tools, the intelligent automatic customer service device 1 constructed according to the server system of the present invention can also support the use of instant messaging tools, such as common instant messaging tools such as MSN and QQ. In this case, the customer service robot usually appears in the form of a chat partner on the other end of the instant messaging tool, which is a more relaxed and casual communication environment, allowing users to partially eliminate the need to face important departments anxiety and tension.

The above two ways of information exchange with the user terminal 2 can be used alone or in combination, and can be selected according to specific conditions in actual application, and no matter which way can be used to complete the technical solution of the present invention.

When applying the intelligentized automatic customer service equipment 1 built by the server system of the present invention, if the user needs to exchange information with the specific department that is provided with the customer service robot system in the above-mentioned certain unit, such as consulting and the specific department business-related Information can be input in natural language through the user terminal 2. Assuming that the user terminal 2 conducts information exchange through the instant messaging tool MSN and the intelligent automatic customer service device 1, the above-mentioned information input in natural language will pass through the instant messaging tool MSN via The Internet sends to the Internet WEB server 73 in the server system, and the intelligent automatic customer service device 1 receives the question input in natural language through the Internet WEB server 73 .

Specifically, the above-mentioned way for the customer to input questions based on natural language can be through various input methods to the user terminal 2, such as using the Microsoft Pinyin input method on the keyboard for text input, or through a microphone to the user terminal 2 Voice input, in the case of voice input, will involve the process of recognizing the input voice to judge its semantics, because it belongs to mature prior art, it will not be described in detail in the present invention.

Then, the intelligent customer service engine server 72 in the intelligent automatic customer service device 1 converts the received natural language into information for computer recognition through human-computer interaction technology; then outputs feedback information according to the information for computer recognition , and send the feedback information to the user terminal 2 via the Internet via the Internet WEB server 73 via the instant messaging tool MSN.

Specifically, when outputting feedback information, the intelligent customer service engine server 72 performs semantic analysis on the input natural language and performs semantic matching on it according to the knowledge base in the established knowledge base server 71, so as to obtain the information preset in the knowledge base. Neutralize the answer that matches the semantics of the input natural language and output it. If the intelligent customer service engine server 72 does not obtain an answer matched with the semantics of the imported natural language, it will judge whether there is an inquiry message matching the semantics of the imported natural language, if there is a corresponding Then output the query information and ask the user a question; then the intelligent customer service engine server 72 receives the user's natural language feedback input according to the query information again; then the feedback input natural language is processed. Semantic analysis and semantic matching; repeat the above steps until the user stops inputting natural language on the user terminal 2, or until the user obtains an answer matching the natural language input by the user from the user terminal.

In the process described above, the knowledge base in the knowledge base server 71 can use parallel data tables, in this case, it is divided into two data tables, as shown in Figure 3, wherein the first data table It is used to store preset answers that match the semantics of specific questions based on natural language. For example, preset answer a1 is associated with question q1, preset answer a2 is associated with question q2, etc. These are related to user input based on natural language The relationship between the answers matching the questions in the database is parallel; the second data table is used to store preset inquiry information matching the semantics of a specific natural language, for example, the preset inquiry information a1' is associated with the question q1', The preset query information a2' is associated with the question q2', etc., and the relationship between these query information associated with the question input by the user based on natural language in the database is also parallel.

In addition, the knowledge base can also use a data table in the form of a tree structure. In this case, the specific natural language in the data table is pre-classified and arranged step by step according to its semantics, as shown in FIG. 4, the tree data The table classifies and arranges the questions entered by users based on natural language level by level. For example, when the questions entered by users based on natural language are semantically analyzed and matched into Q1 questions, there will be several corresponding inquiries in this type of questions Information can continue to be asked or prompted to make the user answer further, such as Q11, Q12, Q13 and other questions to ask the user, so as to finally get the answer semantically related to the question entered by the user based on natural language. Of course, such a tree-structured data table can also directly obtain an answer when a natural language-based question input by a user has a directly associated answer.

In order to make the answer of the intelligent automatic customer service equipment 1 to the user more perfect, the knowledge base server 71 can also store the received natural language input content and its semantic matching results to update the knowledge base contained in it, such as the above-mentioned A data table, the semantic matching result includes answers or query information that match the semantics of the natural language input by the user on the user terminal 2 .

Since the questions input by the user on the user terminal 2 are various, for example, some of them are not directly related to the consulting service of the government department, or although they are related to the consulting service of the government department, the corresponding questions are in the knowledge of the knowledge base server 71. There is no corresponding or related answer in the library. In this case, the answer matching the semantics of the received natural language cannot be obtained, and even the query information cannot be fed back to the user according to the semantics of the received natural language. According to the preferred solution of the present invention, when the intelligent customer service engine server 72 does not obtain answers and inquiry information that match the semantics of the received natural language, it can execute predetermined default instructions and feed back compensatory information to the user terminal 2. Content, for example, the intelligent customer service engine server 72 can output feedback information according to preset default information; or, the received natural language input content that cannot obtain semantic matching results is sent to a preset email address by email, The preset email address is, for example, regularly managed by a dedicated artificial customer service personnel; the human-computer interaction interface can also be linked to the preset address, and the specific method includes displaying a link address to the customer through the instant messaging tool MSN, Guide customers to link to the pre-set address, the pre-set address can be the customer service number in charge of the manual customer service staff, or it can also be the manual customer service in the form of MSN, QQ and other instant messaging tools, and the manual customer service staff supplementarily Answer some questions for which the intelligent customer service engine server 72 cannot obtain an answer matching the semantics of the received natural language.

In order to continuously improve and increase the content of the knowledge base stored in the knowledge base server 71 during the use of the intelligent automatic customer service device 1 of the present invention, in the process of information exchange between the intelligent automatic customer service device 1 and the user terminal, Especially when there are artificial customer service personnel involved in answering the question that the intelligent customer service engine server 72 cannot obtain an answer matching the semantics of the received natural language, the knowledge base server 71 can store relevant content for updating the knowledge base, for example, it can Store the received natural language input content, as well as the semantic analysis and semantic matching results of the natural language input content, as well as the number of occurrences; and the received natural language input content and It is the corresponding feedback information replied by human customer service personnel; in addition, it also includes the storage of some language vocabulary related to professional fields. Usually, one or more contents stored in the knowledge base server 71 can optionally update the knowledge base in the knowledge base server 71 by overwriting or appending.

Therefore, according to a preferred solution of an embodiment of the present invention, the server system used to construct the customer service robot platform actually includes two systems: a customer service robot management system and a customer service robot intelligent engine system.

As shown in FIG. 8 , the server system according to the above preferred solution includes: a knowledge base server 71 , an intelligent customer service engine server 72 , and an Internet WEB server 73 . Three servers perform the functions of knowledge base storage, intelligent customer service and data exchange with the Internet respectively. These servers are mainly used to form the intelligent engine system of the customer service robot. They are interconnected through cables or intranets, and can communicate with each other. Any two servers can exchange data with each other. In addition, on the basis of the customer service robot intelligent engine system formed by the above servers, the complete server system also includes an intelligent customer service management background 75, which is mainly used to form a customer service robot management system.

The customer service robot management system can help users manage and deploy intelligent automatic customer service equipment 1'. It supports UNIX, Linux, Windows and other operating systems. Managers can easily manage the content of the customer service system through the WEB-based management background, such as adjusting the form of customer service system and other operations. Concretely comprise: 1), increase, delete and revise of customer service knowledge base; 2), the batch import of customer service knowledge base, for example the operating system that the present invention adopts can support the data import of multiple modes (Excel, Access, text file etc.) , and you can choose options such as overwriting/appending; 3), the processing strategy of the knowledge base, when there is no problem related to the user's question in the database, what strategy to use to deal with it, you can choose the default answer, Email to the manual customer service , Transfer to manual customer service (MSN robot support) and other methods; 4), manage (add and modify) the robot's welcome message; 5), manage (add and modify) the default answer of the robot; 6), modify the manual customer service 7), the log analysis function, you can check the detailed usage of the system, the specific functions are as follows: question content, answer content, IP address, question time, response time; for a certain IP, check the IP question list and answer content; through Log analysis, the system administrator can mark wrong answers, so that the robot can avoid wrong answers when encountering the same question again; provide the ranking of the number of questions asked; provide a filtering function, which can filter out correct answers, wrong answers, alternative answers, Unable to answer etc. At the same time, it provides question time screening, IP address screening, and question content screening; 8), analyzes the list of questions that the robot cannot answer, adds correct questions and answers to the knowledge base, and continuously enriches the knowledge base, thereby improving the intelligence of the robot; 9) , Increase professional vocabulary, enrich the robot's professional field knowledge; 10) Flexible management personnel authority settings.

When using the intelligent automatic customer service equipment 1 or 7' according to the present invention to exchange information with the user, the process of the method for realizing intelligent automatic customer service can also refer to FIG. The schematic diagram of the system architecture of the customer service robot platform and Figure 2 describe the specific steps in the process as follows:

In step S210, the user inputs information to the customer service robot platform constructed by the server system shown in FIG. The data such as the sound that the user inputs with microphone sends to the Internet WEB server 73 in the described customer service robot platform by Internet.

Step S220, the Internet WEB server 73 sends the received data to the intelligent customer service engine server 72 through a cable or local area network, and the intelligent customer service engine server 72 carries out semantic analysis to the input natural language and communicates with the pre-stored knowledge base server 71. The knowledge base performs semantic matching, for example, extracting semantic keywords in the natural language, rehearsing and combining them into sentence fragments suitable for recognition or classification retrieval by the intelligent customer service engine. Then perform semantic matching on the knowledge base to obtain an answer matching the semantics of the natural language.

Then execute the judgment step S230, if after the above steps, the intelligent customer service engine server 72 has obtained the answer matched with the semantics of the natural language input by the user in the knowledge base server 71, then directly execute step S260, the intelligent customer service engine server 72 The answer that matches the semantics of the natural language input by the user is used as the feedback information, and it is sent to the Internet WEB server 73, and the Internet WEB server 73 passes through the Internet in an Internet WEB interactive mode or in the form of an instant communication tool. The feedback information is sent to the user terminal 2 to complete the information exchange process from receiving the user input information to the user obtaining the feedback information.

If after the above steps, the intelligent customer service engine server 72 cannot obtain an answer that matches the semantics of the natural language input by the user in the knowledge base of the knowledge base server 71, then continue to perform the judgment step S240, and judge that in the knowledge base server 71 Whether there is query information matching the semantics of the natural language input by the user in the knowledge base of 71 . If there is an inquiry information matching the semantics of the natural language input by the user, then step S241 is performed, and the intelligent customer service engine server 72 extracts the inquiry information from the knowledge base server 71, and feeds it back via the Internet through the Internet WEB server 73 to user terminal 2. Wherein, the inquiry information is used to ask questions to the user, for example, when the natural language input by the user is "I want to know the weather conditions." Since the weather conditions vary according to the region and time, the intelligent customer service engine server 72 is based on the knowledge An answer that matches the semantics of the natural language input by the user cannot be obtained from the knowledge base of the library server 71 . Usually, in the knowledge base of the knowledge base server 71, there will be query information related to the semantics of the question input by the user "what area and time do you want to know about the weather?" In this case, the customer service engine server 72 can The query information is obtained from the knowledge base, and then the customer service engine server 72 feeds it back to the user terminal 2 via the Internet WEB server on the Internet.

The user can further input information in natural language according to the inquiry information displayed by the user terminal 2, for example, "I want to know the weather conditions in Shanghai today." The input natural language data will be sent to the customer service robot by the user terminal through the Internet The Internet WEB server 73 in the platform repeats the steps starting from step S210.

If after the above-mentioned steps, the intelligent customer service engine server 72 cannot obtain the inquiry information matching the semantics of the natural language input by the user in the knowledge base of the knowledge base server 71, the intelligent customer service engine server 72 will continue to execute the predetermined default. Step S250 is omitted, for example, the intelligent customer service engine server 72 can output feedback information according to preset default information; or, the received natural language input content that cannot obtain semantic matching results is sent to the preset email address by email. Address, the preset mail address is, for example, regularly managed by a dedicated manual customer service personnel; the human-computer interaction interface can also be linked to the preset address, and the specific method includes displaying a link to the customer through the instant messaging tool MSN Address, guide customers to link to the pre-set address, the pre-set address can be the customer service number in charge of the manual customer service staff, or it can also be the manual customer service in the form of MSN, QQ and other instant messaging tools, supplemented by the manual customer service staff Responsively answer some questions for which the intelligent customer service engine server 72 cannot obtain answers matching the semantics of the received natural language.

Usually, when the user stops inputting natural language or obtains an answer matching the natural language input by the user, the process ends.

In conjunction with the specific description of the above embodiments, the analysis of the advantages of the method provided by the present invention and the equipment for realizing the method is as follows:

1) Reduce the reinvestment of the enterprise in the customer service center, and calmly cope with the large increase in the number of users;

According to the survey of customer service centers of banks, telecommunications, governments and ordinary enterprises, most of the questions received by customer centers are repetitive questions or simple consultation questions, such as product functions, characteristics, handling procedures, etc. If we can divert this part of the problem to other channels, such as the enterprise's website, it will be able to meet the needs of this part of users. Enterprises usually try to do this, but according to the feedback of end users, when they visit the enterprise website, they usually feel that the structure of the website is too complicated, and it takes too much time to find the content they need.

The customer service robot or the intelligent automatic customer service equipment of the present invention can understand the customer's needs through guided questions and answers, and give the user a satisfactory answer in a relatively short period of time. And users only need to chat with customer service representatives to achieve their goals. And due to the characteristics of the system architecture itself, the present invention can realize the service for a large number of users at the same time.

2) Reduce the operating cost of the enterprise customer service center;

According to Forrester Research, 30% of customers will get answers through Self Service on the Internet. According to Garner Group's research, the cost of customer service replies is $5.50/one, while the cost of using online customer service is only $0.24/one; companies can save more communication costs by transferring some inquiries to lower-cost online customer service and personnel costs.

3) Reduce customer's unpleasant experience - waiting time, accuracy of answers

Due to the transfer of some consulting issues to online customer service, enterprises have more and more resources (lines, personnel) to invest in complex issues. By adopting the technical solution of the present invention, for an enterprise, the time to queue up is shorter every time it dials the customer service center, and it can even be connected immediately.

For users who use online customer service, the system can give relevant answers immediately according to their questions, and because their answers are based on enterprise standard information, there will be no errors.

5) Analyze customers and self-update service capabilities;

The system can completely record all user problems, and companies or governments can analyze customer behavior characteristics based on these records, such as which problems users are most concerned about, what products/services specific users are interested in, etc. With traditional customers, the information is stored in the voice form or belongs to the private information of the customer service personnel, and it is difficult to reprocess the information (search and analysis).

The system can provide more accurate answers through the analysis of frequently asked questions. For problems that cannot be satisfied temporarily, enterprises can conduct analysis to provide information for similar problems in the future.

6) Provide round-the-clock and holiday service at low cost

During holidays and off-duty hours, users are willing to spend longer to solve product or service problems. In the traditional customer service center mode, enterprises or governments have to hire more customer service representatives on the one hand, and on the other hand, the average call time has changed. In the long run, the corresponding cost will increase sharply. Once deployed, Inspiron's intelligent customer service robot can provide full-time services with almost no cost increase.

As a brand-new customer service tool, the customer service robot is mainly used to solve problems that cannot be solved by manual customer service. It has low investment cost and good application effect. It can reply users' questions online in real time 24 hours a day. Units that apply this technology have significantly improved their ability to serve the public. The customer service robot can help the departments at all levels of the unit conveniently transition from the manual customer service mode to the computer automatic artificial intelligence robot customer service mode. Through the website, information can be transmitted completely, combined with pictures, text and even video to give users the most complete reply, allowing users to solve problems in pleasant and relaxed communication with customer service robots.

In short, enterprises or governments can provide problem triage support for traditional customer service centers by deploying intelligent customer service robots or intelligent automatic customer service equipment, so that traditional service centers can focus on complex and high-value customers. Therefore, in the two channels, enterprises can provide users with more rapid, accurate and standard services. Under the condition of low cost, it provides effective support for the rapid development of enterprises. For enterprises or governments, the use of this leading technology also adds a new image (image of technology and innovation).

The embodiments of the present invention have been described above, but the present invention is not limited to specific systems, devices and specific protocols, and those skilled in the art can make various changes or modifications within the scope of the appended claims.

Claims (14)

1. a method that realizes automatic customer service by human-computer interaction technology is characterized in that, may further comprise the steps:

A receives the Query Information based on natural language of user's input;

B1 to the Query Information as user's input information carry out semantic analysis with will be wherein the key word permutation and combination become to be suitable for to identify or the statement fragment of systematic searching, and carry out semantic matches with the content in the knowledge base that prestores, the problem that acquisition and this Query Information are complementary, described knowledge base comprises a plurality of problems with certain topological structure and the answer that is associated thereof;

The problem that b2 obtains when described semantic matches has corresponding answer, and this answer is exported as Query Result.

2. method according to claim 1 is characterized in that, described step a comprises:

-receive the user by the Query Information based on natural language of literal or phonetic entry.

3. method according to claim 1 is characterized in that, also comprises the following steps:

The problem that b3 obtains when described semantic matches does not have corresponding answer and only has the inquiry message that is associated, then with this inquiry message output;

B4 receives the feedback information based on natural language that the user inputs for described inquiry message;

Described feedback information is come repetition above-mentioned steps b1 to b4 as described user's input information, indicate the indication information that finishes inquiry for the user until obtain answer and the described feedback information of described correspondence.

4. according to claim 2 or 3 described methods, it is characterized in that, also comprise the following steps:

The described user of-storage input based on the Query Information of natural language and/or feedback information and semantic matches result thereof to upgrade described knowledge base, described semantic matches result comprises problem and corresponding answer or the inquiry message of described semantic matches.

5. according to claim 2 or 3 described methods, it is characterized in that, when the answer that is complementary when the semanteme that does not have to obtain with the natural language that receives and inquiry message, carry out any in the following steps or appoint a plurality of:

-according to predefined default information output feedback information;

-the natural language input content that can't obtain the semantic matches result that receives is sent to predefined addresses of items of mail by Email;

-described human-computer interaction interface is linked to predefined address.

6. method according to claim 5 is characterized in that, stores following each or appoints multinomial content, is used for refreshing one's knowledge the storehouse:

Query Information and/or feedback information and the semantic matches result thereof based on natural language of-the described user input that receives, with and occurrence number;

Can't obtain natural language input content and the corresponding feedback information thereof of the inquiry message of the answer of semantic matches and semantic matches in-the described knowledge base that receives;

-relate to the language vocabulary of professional domain.

7. method according to claim 6 is characterized in that, the one or more content of storing is upgraded described knowledge base with replacement or the mode of appending.

8. an equipment of realizing automatic customer service by human-computer interaction technology is characterized in that, comprising:

Human-computer interaction device is used for receiving the Query Information based on natural language that the user inputs;

Feedback assembly, it comprises the semantic matches device, be used for the Query Information as user's input information carried out that semantic analysis becomes to be suitable for to identify with key word permutation and combination wherein or the statement fragment of systematic searching, and carry out semantic matches with the content in the knowledge base that prestores, the problem that acquisition and this Query Information are complementary, described knowledge base comprises a plurality of problems with certain topological structure and the answer that is associated thereof;

The problem that obtains when described semantic matches has corresponding answer, and this semantic matches device is exported this answer as Query Result.

9. equipment according to claim 8 is characterized in that, described human-computer interaction device is used for receiving the user by the Query Information based on natural language of literal or phonetic entry.

10. equipment according to claim 9 is characterized in that, described semantic matches device also is used for:

-when the problem of described semantic matches acquisition does not have corresponding answer and only has the inquiry message that is associated, then this inquiry message is exported;

The feedback information based on natural language that-reception user inputs for described inquiry message.

11. according to claim 9 or 10 described equipment, it is characterized in that, described feedback assembly also comprises:

The knowledge base memory storage, be used for storing described user's input based on the Query Information of natural language and/or feedback information and semantic matches result thereof to upgrade described knowledge base, described semantic matches result comprises problem and corresponding answer or the inquiry message of described semantic matches.

12. according to claim 9 or 10 described equipment, it is characterized in that, when the answer that is complementary when the semanteme that does not have to obtain with the natural language that receives and inquiry message, described feedback assembly is carried out any in the following steps or is appointed a plurality of:

-according to predefined default information output feedback information;

-the natural language input content that can't obtain the semantic matches result that receives is sent to predefined addresses of items of mail by Email;

-described human-computer interaction interface is linked to predefined address.

13. equipment according to claim 11 is characterized in that, described knowledge base memory storage is used for storing following each or appoints multinomial content, with the storehouse of refreshing one's knowledge:

Query Information and/or feedback information and the semantic matches result thereof based on natural language of-the described user input that receives, with and occurrence number;

Can't obtain natural language input content and the corresponding feedback information thereof of the inquiry message of the answer of semantic matches and semantic matches in-the described knowledge base that receives;

-relate to the language vocabulary of professional domain.

14. equipment according to claim 13 is characterized in that, the one or more content that described knowledge base memory storage is stored is upgraded described knowledge base with covering or the mode of appending.

Images