JPH0488460A - Information processor - Google Patents

Abstract

PURPOSE:To obtain a useful answer and advice to a user's question at all times by collecting and pigeonholing knowledge and knowhow decentralized in a user group. CONSTITUTION:This unit is equipped with interface parts 11 and 12 for the user (questioner), an interface part 13 for other users (answerer), a reasoning and retrieval part 14 which retrieves or generates an answer to a question from the user, a result evaluation part 15 which evaluates the obtained result, a knowledge acquisition part 16 for obtaining information from other users, and a knowledge base 17 stored with the knowledge of the system in an object area. When necessary information is not obtained or deficient, the user makes a request to collect another piece of information from other users, the system collects information from other users through a network, and the result is sent to the user who makes the question. At the same time, the information is stored in the knowledge base of the system and the updated knowledge base is used to answer. Consequently, the proper answer and advice are obtained at all time.

Description

[Detailed Description of the Invention] Industrial Field of Application This invention relates to a conventional information processing device that provides appropriate answers and advice in response to questions asked by users to the system, such as how to operate a computer. With the development of technical information processing technology, systems with a wide variety of functions have appeared, but at the same time, the operating procedures for activating these processing functions have also become more complex, and in order to make full use of them, it is difficult for users to In order to deal with this problem, development of help systems has become active. This is a system that displays the solution to the problem.The basic structure of such a help system C is shown in Figure 2.

Figure 2 21 shows an interface section that detects a user's help request and displays solutions and advice for the request. Although the necessary knowledge is stored in Knowledge Base 23, for example, if it is written in Knowledge Base 23 that ``To delete a file, use the command rm,'' a certain user would like to delete a file. If you don't know how to do this, please ask the help system.
On the other hand, with the advancement of network technology and the spread of e-mail, communication between users across different locations has become more popular. One of the most common behavior patterns in life is to first ask someone close to you when you don't understand something.However, the person who has the answer is not always close to you. In such cases, it is possible to contact people who are not present by using e-mail, or by using a network-based electronic bulletin board (BBS). Although it is possible to obtain advice from users, the problem that the invention aims to solve cannot be solved using methods based on conventional knowledge bases. If the user lacks even a little bit of the necessary knowledge, the user will not be able to get a sufficient answer. However, when the target field is as complex as operating systems, the amount of knowledge required becomes enormous, and in practice it is impossible to build a complete knowledge base. However, the following method is usually used: Questions that the system cannot answer are recorded and the knowledge required for them is periodically added to the knowledge base by the maintainers of the knowledge base.However, in this method Users who first ask questions often receive no information;
As a help system, it is often not actually useful. If there are many answers to a question, they will not be systematically organized and saved. Even if another user has the same question, it is not possible to search for previous information. It's not impossible, but it's not easy (I don't know how many times the same question has been asked.
It is not desirable for it to appear on an electronic bulletin board (~ The purpose of the present invention is to collect and organize knowledge and know-how distributed within a user group.
To provide an information processing device that can always provide useful answers and advice.A.Means for Solving the ProblemsIn order to solve the above problems, the information processing device of the present invention (such as
A means for inputting a user's help request, an inference/search means for automatically searching a main point or a knowledge base for a solution or advice thereto, and a means for confirming with the user whether the solution or advice was effective; If you are unable to give a sufficient answer, there is a means to ask other users for the proper answer, a means to convey the obtained answer to the first person who asked the question, and a method to memorize the contents and do the same in the future. According to the above-described structure, the present invention has a structure in which the inference/retrieval means uses the information when a question arises. When a user asks a question, the system first tries to provide an answer using the knowledge in the knowledge base.
It is better for users to find the answer themselves by searching for information in the knowledge base. Either way, if the information you are looking for is found, you are done. If not obtained, or
If the information obtained is not sufficient, the user can request the system to collect further information from other users, and the system may collect information from other users over the network. Also, the obtained result ζ is conveyed to the user who first asked the question, and at the same time, it is also saved in the system's knowledge base.If another user asks a similar question, CAL will respond with the updated knowledge base. You can use it to reply.

EXAMPLE Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing the configuration of a help system in an embodiment of the present invention. 13 is the number of interfaces for other users (answerers) 14 is the number of inferences for searching or generating answers to the user's questions
Search wide 15 is a result evaluation unit that evaluates the results obtained by 14, 16 is a knowledge acquisition unit for obtaining information from other users, and 7 is a knowledge base that stores system knowledge in the target area.Knowledge base Figure 6 shows part of the knowledge stored in . Figure 6: This is part of the system's knowledge regarding the operation of IT computers.The arrows in the diagram represent the hierarchical relationship between nodes.For example, "copying a file" and "deleting a file" are both "file operations".
"File operation" is also a "copy operation".
is also one of the ``operations on a computer.''a The nodes drawn with ellipses in the diagram represent concepts related to computer operations.
Although the nodes drawn as rectangles represent specific processing methods, knowledge such as ``Use the ζ command Cp to copy the file''! Figures 3 and 4 are flowcharts showing the flow of the process, even though it is represented by a link between the node ``Copy file'' and the node ``Use command Cp''. In the system, the first
3 @ Centered on Figure 4 & Go To explain its operation, first of all, if there is something that the user does not understand, the user will type in a natural language such as Japanese or English on the keyboard,
Questions are asked of the system by selecting help items from a menu using a mouse or the like.

The help request is detected by the system interface unit 11 (step 31 in FIG. 3).

As the processing contents of the interface unit 11 (for example,
When a user inputs "How do I copy a file?", it generates an internal expression like the one shown in Figure 3. Inference/Search Department 14th Friend Ability to find the solution to the user's question.
The knowledge necessary for this is stored in knowledge base 17 (
Figure 3 step 32).

For example, if the internal structure shown in Figure 3 is obtained from the user's question, the concepts such as "copy" and "file" that appear there, or the concept "copy file" in Figure 6.
tic in the knowledge base 17. The internal representation obtained by the interface unit 11 (in this example, Figure 3) is the concept node in Figure 4 (in this case, "copy file"). Even though it includes the knowledge to check whether the user intends to
Let's check whether the nodes ``what'' are connected by the links ``what'' and ``what'' are the links a. In this case, since that condition is satisfied, it is determined that what the user wants to know is how to ``copy a file.'' , the processing method linked from that node is displayed to the user by the interface unit 11 (step 34 in FIG. 3).

If the user's question is ``I copied a file from machine A to machine B (...''), the internal representation would be as shown in Figure 7.However, if the knowledge held by the system remains as shown in Figure 6. As before, the system is
Use command cp.

” would be the answer. The structure shown in FIG. 7 includes the structure shown in FIG. Push,
It is checked whether the results given to the user are sufficient (step 35 in Figure 3). For example, if you want to copy a file to another machine, use the command rcpJ to copy a file to another machine. If the user is dissatisfied with the advice given (Step 36 in Figure 3), the answer cannot be found by the inference/search unit 14. (Step 33 in Figure 3), if the user (questioner) wishes (Steps 37 and 38 in Figure 3), the knowledge acquisition unit 16 is activated, but as long as the content of the user's question remains unchanged.
The internal representation The node finally reached in the current knowledge base, that is, the location where an incorrect answer was given, etc., is recorded (step 39 in FIG. 3).

Through the knowledge acquisition unit CAL interface unit 13,
Request help from other users on the network (step 41 in FIG. 4). This can be done by sending an email to each user.

The user ζ, who is sent a message from the system, can be an unspecified number of people, even members of a specific pre-registered group. Among the users who received the message from the system, the person who knows the answer or who can give advice is the interface section 1.
3, the contents are conveyed to the system (step 42 in FIG. 4). In the previous example, even though the system sends an e-mail with the content shown in Figure 9, it is not necessary to inform the inquirer of the affiliation, name, etc. For example, as shown in Figure 10, person 1 fills in the "Answer" column and sends a reply to the system.If there are replies from multiple users and they contain the same answer, there will be duplicates. removes (
Figure 4 step 43). The newly acquired knowledge is added to the knowledge base 17 of the system by the knowledge acquisition unit 16 (step 44 in Figure 4), and is added to the system C
The answer is transmitted to the first questioner through the interface section 12 (step 45 in FIG. 4). In the previous example, the bottom of the "Copy file" node in Figure 6 is the new area.
A node called "Copying files between machines" is added (Figure 8). Using this updated knowledge base, from now on, when a user asks the question expressed in Figure 7, the answer can be ``Use the command rcp.'' How do I copy the system?'
If that does not work, it is also possible to infer that the user's goal is to ``copy files between machines.'' Steps in Figure 4 Steps 44 and 45 can be performed in the reverse order.The obtained information can be communicated not only to the questioner but also to all desired users.

Also, the knowledge base does not necessarily have to be one, and only the knowledge obtained from other users can be managed separately. Even if the knowledge base is incomplete, the user can be guaranteed to get some kind of response.
Although the following effects can be obtained, (1) there is no need to prepare a complete knowledge base from the beginning - (2) there is no need to have extra knowledge that may or may not be used; Keeping the size of the knowledge base compact (3) The knowledge base does not limit the scope to any particular area, so users can essentially ask all kinds of questions and can also ask all kinds of questions. By making it possible to add knowledge to the user exchange system, (4) there is no need to provide special personnel to update the knowledge base. It is common for there to be more than one answer to a given problem (each user is often unaware of alternative solutions), even though they can be collected and shared as a group. (6) Tell all users in the group who want to know about it, other than the person asking the question.

[Brief explanation of drawings]

FIG. 1 shows the configuration of an information processing device in an embodiment of the present invention. FIG. 2 shows the configuration of a help system based on a conventional knowledge base. FIG. 3 shows the processing of the inference/search section and result evaluation section in the embodiment. FIG. 4 is a flow doctor showing the processing of the knowledge acquisition unit in the embodiment. FIGS. 5 and 7 are flow doctors showing an example of internal expression of the user's question in the embodiment. Figure 8 shows part of the knowledge example stored in the knowledge base in the example; Figure 9 shows part of the knowledge in the knowledge base after being updated in the example. Figure 10 is a diagram showing an example of a reply from another user to a message from the system. Interface inquiry for other users (answerers) 24...Inference/search unit 25...24 that searches for or generates answers to user questions.
A result evaluation unit that evaluates the results obtained by 26...
・Knowledge acquisition unit for obtaining information from other users 27・
...Knowledge board that stores system knowledge in the target area 11...Interface question 12...Inference/search IQ 13...Knowledge board name of agent Patent attorney Shigetaka Awano Haka1 persontlEj Figure 1 2 Figure User Chio Base Haku's User nth! Kasuri 1 Knowledge 11hlIIllIl Figure Figure Figure End Figure File is copied from machine AtP to machine 8 Figure M. The file is also copied from machine A to machine 6. Supplementary note: 1 method of working that I enjoyed using the command cp No. 101! 1. How to process (if you are able to read and write) How to copy and write files between machines

Claims (1)

[Claims] A means for inputting a user's help request, an inference/search means for automatically generating or searching a solution or advice from a knowledge base, and a means for confirming with the user whether the solution or advice was effective; If a sufficient answer cannot be given, a means to ask another user for the original appropriate answer, a means to convey the answer obtained as a result to the first questioner, and a means to memorize the contents, An information processing device characterized in that when a similar question is asked thereafter, the inference/search means uses the information.