JPH05189239A - Method and device for inputting natural language - Google Patents

Abstract

PURPOSE:To obtain method/device for inputting natural language, which can input a command and a natural language sentence without discriminating them by providing means analyzing sentence data and judging and processing a designated content. CONSTITUTION:An input part 101 inputs the command peculiar to the device which is inputted by a user. A judgement part 102 judges data inputted from the input part 101 and judges whether inputted data is the command peculiar to the device or the command by sentence data generated by the natural language sentence. A natural language processing part 103 analyzes the natural language sentence obtained from the judgement part 102 and judges the command of a back end system 109, which corresponds to the content which the sentence data means. Thus, the judgement part 102 outputs the command received from the natural language processing part 103 to a command execution part 108. It executes the command of the back end system 109 and outputs the result to a result display part 110.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a natural language input method and apparatus capable of interactively inputting commands using text data created in natural language.

[0002]

2. Description of the Related Art In recent years, natural language input devices have been researched and developed which are capable of inputting text data created in natural language and executing a back-end system command instructed by the text data. To briefly explain the configuration of such a device, a back-end system command (a non-natural language specific command) input by a user is received, the command is executed to the back-end system, and input by the user. The command of the natural language sentence is received, the natural language sentence is analyzed, the command of the back-end system that the natural language sentence means is determined, and the command is executed to the back-end system.

As described above, in the natural language input device capable of inputting two kinds of commands, the backend system is operated based on the command given by the device and the backend system is operated based on the command given by the natural language sentence. In the case of operating the end system, the input commands are processed by different processing systems.

Further, in an apparatus capable of inputting and processing a command created by a natural language sentence, the input natural language is analyzed by a sentence analysis unit, and as a result of the analysis, it is determined to be a command. The execution result is returned to the user according to the user level stored in advance.

[0005]

Since this stored user level is only inputted by the user at the start of the dialogue or changed depending on the contents of the dialogue, there is a long time interval before the device is used again. As a result of the vacancy, the user's own level becomes lower than that of the previous time and the level will not match.

Further, in the above-mentioned conventional example, the process of moving the backend system by the instruction of the specific command and the process of moving the backend system by the command input in the natural language are separately performed. There was a problem. Since each of these processes operates as a separate task, the user must switch between the functions of both natural language sentence command input and device-specific command input. .. Even if both of these command input screens are displayed at the same time by the multi-window system so that commands can be input, the user needs to perform an operation for selecting a window. Further, since different processing is performed for each window, a series of interactive processing by the user is interrupted.

The present invention has been made in view of the above-mentioned conventional example, and allows a user to input an instruction command in a natural language sentence and a command peculiar to the apparatus in a mixed manner.
An object of the present invention is to provide a natural language input method and device that allow a user to input a command and a natural language sentence without distinguishing between them.

It is another object of the present invention to provide a natural language input method and device capable of changing a user level according to a user's usage time or usage interval.

[0009]

In order to achieve the above object, the natural language input device of the present invention has the following configuration. That is, a natural language input device for inputting and executing sentence data created in natural language, and inputting both sentence data including instruction information created in natural language and command data specific to the device. Means, an analysis means for analyzing the text data when the text data is input from the input means, a determination means for determining the instruction content based on the text data analyzed by the analysis means, and a determination by the determination means And processing means for performing processing in accordance with the designated instruction content or the command input from the input means.

In order to achieve the above object, the natural language input method of the present invention comprises the following steps. That is, it is a natural language input method of inputting and executing sentence data created in natural language, and a step of inputting both sentence data including instruction information created in natural language and command data peculiar to the device. When the text data is input, a step of analyzing the text data, a step of determining the instruction content based on the analyzed text data, and a process according to the determined instruction content or an input specific command And the process.

In order to achieve the above object, the natural language input device of the present invention has the following configuration. That is, a natural language input device for inputting and executing sentence data created in natural language, the input device being operated by a user, for inputting sentence data including instruction information created in natural language, and the sentence. An analyzing means for analyzing data, a judging means for judging an instruction content based on the sentence data analyzed by the analyzing means, a processing means for performing processing according to the instruction content judged by the judging means, and an input by the input means According to the user level determined by the determining means, the determining means determining the user's level based on at least the user's experience information and the user level information at the time of previous use, and the result processed by the processing means. And display means for displaying and outputting.

In order to achieve the above object, the natural language input method of the present invention comprises the following steps. That is, a natural language input method for inputting and executing sentence data created in natural language, the process of inputting sentence data including instruction information created in natural language by a user, and the input sentence data The step of analyzing the text data to determine the instruction content based on the text data, the step of performing the processing in accordance with the determined instruction content, and the level of the user who input the text data, at least the user's experience information and the time of previous use. And a step of displaying and outputting the processed result in accordance with the determined user level.

[0013]

With the above structure, when both the sentence data including the instruction information created in natural language and the command data peculiar to the device are input, and the sentence data is input,
The sentence data is analyzed, and the instruction content by the analyzed sentence data is determined. Then, the operation is performed according to the determined instruction content or the input specific command.

According to another invention, sentence data including instruction information created by a user in natural language is input, the input sentence data is analyzed, and the instruction content by the sentence data is determined. Then, the level of the user who inputs the text data is determined based on at least the user's experience information and the user level information at the time of the previous use, and the result of processing according to the instruction content of the text data is determined according to the determined user level. Display and output.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing the schematic arrangement of a natural language input device according to the first embodiment of the present invention.

In FIG. 1, reference numeral 101 denotes an input unit for inputting a device-specific command or the like input by a user, as well as a command based on text data created in natural language. A determination unit 102 determines the data input from the input unit 101 and determines whether the input data is a command peculiar to this device or a command based on text data created by a natural language sentence. .. According to the determination result, the determination unit 102 determines the data (command) if the command is a peculiar command.
Is output to the command execution unit 108, and if it is a natural language sentence, the data is output to the natural language processing unit 103.

The natural language processing section 103 includes a determination section 10
The back-end system 109 that analyzes the natural language sentence obtained from 2 and corresponds to the meaning of the sentence data
Command is determined. This natural language processing unit 103
A sentence analysis unit 104 that analyzes the language meaning of the natural language sentence obtained from the determination unit 102, an instruction content determination unit 105 that determines the user's instruction content determined from the language meaning of the natural language sentence, and The command determination unit 10 that determines the command of the back-end system 109 corresponding to the instruction content.
6 and a database 107 that the command determination unit 106 refers to in order to determine the type of command. Then, the command of the back-end system 109 determined by the natural language processing unit 103 is the determination unit 102.
Then, the determination unit 102 outputs the command received from the natural language processing unit 103 to the command execution unit 108.

Reference numeral 108 is a command execution unit, which is a judgment unit 102.
The input command of the back-end system 109 is executed to the back-end system 109, and the result executed by the back-end system 109 is output to the result display unit 110. A result display unit 110 displays / outputs the result executed by the command execution unit 108.

FIG. 2 is a flow chart showing a process of inputting and executing a command in a natural language sentence and a specific command of the backend system 109 in a mixed manner in the natural language input device of the first embodiment. ..

First, in step S1, the user inputs data through the input unit 101. Here, both a specific command input and a command input by a natural language sentence can be input. Next, in step S2, step S
The data input in 1 is determined, and determination processing is performed to determine whether a specific command is input or a command in a natural language sentence. According to the result of this determination process, in step S3, when a specific command is input, step S3 is executed.
7, the information is output to the command execution unit 108.

On the other hand, if it is determined in step S3 that the input is a natural language sentence, the process proceeds to step S4 and the natural language sentence is output to the sentence analysis processing 104 for analysis. In this sentence analysis process, the meaning of this natural language sentence is analyzed, and the content is passed to the instruction content determination process in step S5. In the instruction content determination processing in step S5, it is determined from the analysis result obtained by the sentence analysis in step S4 what purpose or what function the instruction content determination unit 105 is instructing. Next, in step S6, the command determination unit 106 refers to the command database 107 that stores information related to the commands of the backend system 109 based on the determination result, and determines the corresponding command of the backend system 109. It Next, in step S7, the determined command is output to the command execution unit 108 via the determination unit 102. As a result, the command is executed for the backend system 109.

The command execution processing in step S7 is
The command determined by the command determination unit 106 or the specific command input by the user through the input unit 101 is executed by the back-end system 109 according to the determination in step S3. Then, the execution result of the command in the backend system 109 is received from the backend system 109. When the execution result is obtained in this way, the process proceeds to step S8, and display processing for displaying the result is performed. The result display process of step S8 displays the result obtained by the command execution process of step S7 on the result display unit 110. Next, in step S9, it is determined whether or not to terminate the interaction processing with the user, and when an instruction to terminate is input,
The dialogue processing is ended. If not, the process returns to step S1 to perform processing for receiving the next data from the user. By repeating these processes, the command input by the natural language sentence and the command input by the command specific to the back-end system 109 can be mixed and executed.

Next, the dialogue processing in the first embodiment will be described with reference to FIG. In FIG. 3, a UNIX command interpreter is used as the back-end system 109.

In FIG. 3, as indicated by 61, (Us
When the user's instruction “rwd” “pwd” is input from the input unit 101, this data is determined by the determination process of step S2 executed by the determination unit 102. In the example of FIG. 3, since the back-end command is a UNIX command and the natural language sentence is Japanese, in the determination processing of step S2, the received input is an ASCII character or a Japanese character (for example, JIS code). And EUC code)
Is determined. Here, "pwd" is ASCII
Since it is a character, this input is determined to be an instruction by a command, it is determined to be a command peculiar to the back-end system 109 by the determination of step S3, and (Usr1) is input to the command execution process of step S7. Is passed.

In step S7, this input is executed to the UNIX command interpreter which is the back end system 109. Then, the command is executed in step S7, the command execution unit 108 receives the execution result by the backend system 109, and outputs it to the result display processing S8. This results in step S8
3 and the result is "(Sy
s1) / usr / usr1 "is displayed. Thereafter, if the dialogue continues due to the determination in step S9, the process returns to the data input process by the user.

Next, as indicated by reference numeral 63 in FIG. 3, when the user inputs a (Usr2) instruction, "take list of files", the input (Usr2) received from the input process from the user is input. It is passed to the determination processing in step S2. In step S2, it is determined whether this input data is a natural language sentence or a command. In this case, since the input is Japanese text data, this input is determined to be an instruction in a natural language sentence. As a result, the process proceeds from step S3 to step S4, and the sentence analysis processing by the sentence analysis unit 104 (U
The input of sr2) is passed. In this sentence analysis process, semantic analysis of Japanese sentences is performed, and it is interpreted as follows: style: command predicate: take object: file list. This analysis result is step S5.
Is passed to the instruction content determination processing of. In the instruction content determination process in the instruction content determination unit 105, it is determined from the analyzed information that the instruction content of the natural language sentence is: request: (action: list, content: file). Then, the information is passed to the command determination process of step S6. In the command determination process of step S6, the command determination unit 106 refers to the database 107 to search for a command suitable for the back-end system 109 based on the meaning of the sentence and the instruction content. In the example of FIG. 3, since the back-end system 109 is a UNIX command interpreter, (operation: list, content: file) → (command: l
s) is determined. In this way, the “ls” command, which means to take a list of files in the UNIX command interpreter, is passed to the command execution unit 108 via the determination unit 102. In the command execution processing in step S7, this input is input to the backend system 1
09 for the UNIX command interpreter. The command execution unit 108 receives the executed result from the backend system 109 and displays the result on the result display unit 110. Thus, the result display unit 110
The result is "(Sys2) file1 file2 fil
It is displayed as "e3".

As described above, according to the present embodiment, as shown in FIG. 3, the user can input a question or command in a natural language sentence and a command peculiar to the target system in a mixed manner. It became possible.

In the above embodiment, the UNIX command interpreter is used as the back-end system 109, but the present invention is not limited to this.
It can be combined with various back-end systems. For example, a natural language input device for inputting a mixture of SQL, which is a question answering language (command) of a relational database, and a question sentence of a natural language sentence, can be realized.

In the above embodiment, Japanese is used as the natural language sentence, but the present invention is not limited to this.
It is possible to use various types of natural languages.

The commands specific to the back-end system are subject to a stricter constraint condition than the natural language sentence in the subsequent structure (syntax). Therefore, instead of the determination of the character type performed by the determination unit 102 in the above-described embodiment, it may be determined whether or not the user input satisfies the command syntax of the backend system.

In the above embodiment, in the case of an instruction by a natural language sentence, the command decided by the command decision processing shown in FIG.
Although it was output in 08, the determined command may be displayed to the user for confirmation of command input by the natural language sentence. By configuring the apparatus in this way, the dialogue as shown in FIG. 4 can be realized.

That is, when a command is input in Japanese as indicated by 41, the command is converted into a command peculiar to the backend system 109 and displayed for confirmation as indicated by 42. 43 shows the result of executing the command, which corresponds to 62 in FIG. Similarly, when a Japanese sentence such as 44 is input, an example of interpreting it and displaying it as a command is shown by 45.

This confirms the command of the back-end system 109 corresponding to the command in the natural language sentence input by the user and realizes the educational function for the user. This allows the user to memorize this confirmed command, and as shown by 46 in FIG. 4, in the next dialogue, instead of command input in Japanese as shown by 41, according to the backend system 109. You can direct the display of the directory by directly entering the command.

As described above, according to the first embodiment, the user input unit 101 and the determination unit 102 determine the type of the user's input and distribute the instruction to the natural language sentence and the instruction by the command. The effect of is now available. It has become possible for the user to enter questions or instructions in natural language and commands for the target system in a mixed manner. Since the instruction by the natural language sentence and the instruction by the command can be accepted as the same user input, it is possible to eliminate the unnecessary work of switching between the both processes. By displaying the command corresponding to the instruction by the natural language sentence as a confirmation of the analysis result of the natural language processing unit,
Command education for users can be realized.

FIG. 5 is a block diagram showing a schematic structure of a natural language input device according to the second embodiment of the present invention. The parts common to the respective parts of FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted.

In this embodiment, the input unit 101 inputs a command in the form of a natural language sentence, and the natural language sentence input from the input unit 101 is analyzed by the sentence analysis unit 104 to generate a command of the backend system 109. To decide. The command thus determined is executed by the command execution unit 108 to the backend system 109. 1
A response generation unit 21 receives the execution result from the command execution unit 108 and generates a response result in response to the value of the response level input from the response level determination unit 123. The result display unit 110 displays the response result thus generated. A response level determination unit 123 is a user level stored in the user level storage unit 124, a timer 122 for measuring the usage time and usage interval by the user, and a user level table 125 for determining the response level to the user. Based on the experience information of the user from the user use experience information holding unit 126, the level of the response content to the user is determined. Here, the user level holding unit 124 divides the user's proficiency level into stages and stores the degree of response content displayed on the result display unit 110 according to each stage.

FIG. 6 is a flow chart showing command input processing in the natural language input device of the second embodiment.

First, in step S11, a command in a natural language sentence input from the input unit 101 is input, and in step S12, the sentence analysis unit 104 analyzes the input sentence,
The command corresponding to the input sentence is determined. This command is output to the command execution unit 108, where it is executed to the backend system 109. When the execution result is obtained from the backend system 109, the process proceeds to step S14, and the result is output to the response generation unit 121.

Next, in step S15, the response generator 1
When the command execution result is output from the response level determination unit 123 to the response level determination unit 123,
2, the user's use time and use time interval, the value of the user level table 125 for determining the level of the response content to the user, and the information from the user's use experience information holding unit 126 are referred to. Determine the user level according to time and usage interval. Based on the user level thus determined and the execution result of the command in the command execution unit 108, the degree of the amount of information to be displayed to the user is determined. Next, in step S16, the response generation unit 121 creates a response according to the degree of information amount to be displayed, and outputs the response to the result display unit 110 for display. In step S17, it is determined whether or not the interactive process is to be ended. If the interactive process is to be ended, the process is ended. If not, the process returns to step S11.

FIG. 7 is a flow chart showing the response level determination processing in step S14 of FIG.

First, in step S21, the usage experience information of the user (the user who is currently operating) stored in the user usage experience information storage unit 126 is retrieved. Next, step S
In step 22, the current time and the time elapsed since the user started the operation previously are calculated based on the information from the timer 122. Next, in step S23, the level of the user who is currently operating is extracted based on the user level information stored in the user level holding unit 124. Then, the process proceeds to step S24, and the user use experience information holding unit 12
The user level table 125 is referred to based on the experience information stored in No. 6, the elapsed time from the timer 122, and the user level, and the user level is newly determined. The user level thus determined is the response generation unit 12
It is output to 1.

Next, with reference to FIGS. 8 to 10, the interactive input processing in the natural language input device of the present embodiment will be described.

FIG. 8 is a diagram showing an example of user level information stored in the user level holding unit 124.
Indicates the beginner's level. In the user level information, all the information is responded at the level 1, but a part of the information is displayed at the level 2 (the number of operations of 1 to 10 times). In the case of a user who has operated 10 times or more, the minimum amount of information is displayed.

FIG. 9 is a diagram showing an example of the experience information stored in the user experience information holding unit 126, in which the time when the person has used before and the user level at that time are associated with the name of the user. And are remembered.

FIG. 10 shows the user level table 12
5 is a diagram showing an example of the data structure of FIG. 5, in this table 125, the time elapsed since the user used the device previously, and the user level newly set corresponding to the user level at that time (previous) Is remembered. That is, according to this table, if the elapsed time is within 3 hours, the level of the user does not change, but if within 3 hours to 1 day, the level of the user is, for example, "2". It shows that the value drops to "1".

Based on these data, for example,
It is assumed that the user "Taro Yamada" is using the device four days ago and the user level of the user is "2" as shown in FIG. When this user (Taro Yamada) inputs a command in a natural language sentence from the input unit 101, sentence analysis and command execution are performed as described above, and the execution result is sent to the response generation unit 121. In this way, the response level determination unit 123 reads the experience information from the user use experience information holding unit 126, the measurement information from the timer 122, and the level information from the user level holding unit 124.

In this case, the previous user level is "2".
Then, since the elapsed time from the time of previous use is between one day and one week, referring to the user level table 125 shown in FIG. 10, it can be seen that the column indicated by 81 corresponds. As a result, the new user level of the user "Taro Yamada" is changed to "1". In this way, the response generator 1
Reference numeral 21 outputs all response information to the result display unit 110 and displays it in response to the user level “1”.

As described above, according to the second embodiment, it is possible to change the user level according to the usage time and usage interval of the user.

In the above embodiment, as the user level table 125, the time interval from the last use of the apparatus and the previous user level are shown in association with each other, but the present invention is not limited to this. Instead, various user levels can be set. For example, the user level can be changed according to the length of the usage time by associating the elapsed time during which a certain user continuously uses the device with the user level.

Further, as the data of the user level table 125, in addition to the time interval from the last use of the device and the previous user level, the elapsed time during which the user continuously uses the device and the user level thereof. By combining and, it is possible to more finely control the timing of returning from the level down of the user. For example, in the user level table 125 described above, the user level when the user starts using the device is determined, and based on the elapsed time when the user continuously uses the device, the original user is determined according to the user level. You can also return to the level.

As described above, according to the second embodiment, the proficiency level of the user is stored, and the usage time and usage time interval of the user, the usage experience information of the user, and the user level table are used. This allows the device to automatically adjust the level of the user according to the frequency of use and time interval of use by the user. The device can automatically set the level of the user according to the usage time of the user.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program for implementing the present invention to a system or an apparatus.

[0054]

As described above, according to the present invention, the user can input an instruction command in a natural language sentence and a command peculiar to the apparatus in a mixed manner so that the user can input the command and the natural language sentence. There is an effect that can be input without distinguishing.

According to another aspect of the invention, there is an effect that the user level can be automatically changed according to the usage time and usage interval by the user.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of a natural language input device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an input process in the natural language input device according to the first embodiment of the present invention.

FIG. 3 is a diagram showing a command input example in the natural language input device according to the first embodiment of the present invention.

FIG. 4 is a diagram showing an example of command input in a natural language input device according to another embodiment of the present invention.

FIG. 5 is a block diagram showing a schematic configuration of a natural language input device according to a second embodiment of the present invention.

FIG. 6 is a flowchart showing command input processing in the natural language input device according to the second embodiment of the present invention.

FIG. 7 is a flowchart showing a response level determination process in the natural language input device according to the second embodiment of the present invention.

FIG. 8 is a diagram showing a data configuration example of a user level holding unit in the second example of the present invention.

FIG. 9 is a diagram showing a data configuration example of a user use experience information holding section in the second example of the present invention.

FIG. 10 is a diagram showing a data configuration example of a user level table in the second exemplary embodiment of the present invention.

[Explanation of symbols]

 Reference numeral 101 input unit 102 determination unit 103 natural language processing unit 104 sentence analysis unit 105 instruction content determination unit 106 command determination unit 107 command database 108 command execution unit 109 back-end system 110 result display unit 121 response generation unit 122 timer 123 response level determination Part 124 User level holding part 125 User level table 126 User use experience information holding part

Claims (4)

[Claims] 1. A natural language input device for inputting and executing text data created in natural language, wherein both text data including instruction information created in natural language and command data specific to the device are received. Input means for inputting; text data input from the input means; analysis means for analyzing the text data; determination means for determining instruction content based on the text data analyzed by the analysis means; A natural language input device comprising: a processing unit that performs processing according to the instruction content determined by the unit or the command input from the input unit. 2. A natural language input method for inputting and executing text data created in natural language, wherein both text data including instruction information created in natural language and command data specific to an apparatus are input. , The step of analyzing the text data when the text data is input, the step of determining the instruction content by the analyzed text data, and the step of determining the instruction content or the input specific command according to the determined instruction content. A natural language input method, comprising: a step of performing processing. 3. A natural language input device for inputting and executing text data created in natural language, the input device being operated by a user and inputting text data including instruction information created in natural language. An analyzing means for analyzing the sentence data; a judging means for judging an instruction content by the sentence data analyzed by the analyzing means; a processing means for performing a process according to the instruction content judged by the judging means; Deciding means for deciding the level of the user input by means based on at least the experience information of the user and user level information at the time of the previous use, and the user decided by the deciding means for the result processed by the processing means A natural language input device comprising: display means for displaying / outputting according to a level. 4. A natural language input method for inputting and executing text data created in natural language, comprising the steps of inputting text data containing instruction information created by a user in natural language, and The step of analyzing the written text data to determine the instruction content by the text data, the step of performing processing according to the determined instruction content, and the level of the user who input the text data is at least the user's experience information and A natural language input method comprising: a step of determining based on user level information at the time of previous use, and a step of displaying / outputting a processed result according to the determined user level.