JP5202163B2 - Apparatus and method for supporting language learning - Google Patents

Description

  The present invention relates to an apparatus and method for supporting language learning by a user.

  As methods for obtaining language learning, (A) self-study using materials such as books, (B) learning by direct instruction of a language teacher through face-to-face or communication are known. However, the method (A) has failed to learn knowledge of expression and syntax of an appropriate difficulty level in an appropriate order for each individual. In addition, the method (B) has a problem that the cost, time and labor for receiving the instruction are very large. In order to solve these problems, (C) e-learning for learning interactively using a computer has been developed and used.

  Further, Patent Document 1 proposes a natural language learning support device that can provide explanations related to syntax structure by outputting explanations about dictionary information used for syntactic analysis of an input sentence, and has good search efficiency. ing.

JP 2006-330628 A

  However, since the method of Patent Document 1 is a method of learning a natural language by commentary on an analysis result of an input sentence input from a keyboard or a web page, it is suitable for learning about a dialogue between a user and another person. Absent. In e-Learning, interactive learning is possible, but only learning by simulated dialogue can be performed. In other words, the conventional method has a problem that it cannot efficiently acquire “language ability in an actual environment” in which actual communication is performed. In other words, there is a problem that the user cannot learn the live expression that the user actually encountered on the spot efficiently.

  The present invention has been made in view of the above, and an object of the present invention is to provide an apparatus and method that can efficiently learn expressions encountered in actual communication.

  In order to solve the above-described problems and achieve the object, the present invention provides a knowledge storage unit for storing processing language knowledge used in natural language analysis processing, and a learning language used for language learning among the processing language knowledge A comment storage unit that stores knowledge and commentary information that represents commentary for the user regarding the learning language knowledge, and an input character string that represents the utterance content of the user and another person different from the user An input unit that inputs a natural language analysis process on the input character string based on the processing language knowledge, and an analysis that identifies the learning language knowledge among the processing language knowledge used in the natural language analysis process The learning unit specified when the input character string is an utterance of the other person, and a determination unit that determines whether the input character string is an utterance of the user or the other person. Language knowledge An acquiring unit for acquiring the description information associated from the description storage unit, characterized by comprising an output unit that outputs the acquired the explanatory information to the user, the.

  The present invention is also a method capable of executing the above apparatus.

  According to the present invention, there is an effect that an expression encountered in actual communication can be efficiently learned.

  Exemplary embodiments of an apparatus and a method according to the present invention will be described below in detail with reference to the accompanying drawings.

  The language learning support device according to the present embodiment is the language knowledge used by the user of the language learning support device and the utterance of the other person who is the conversation partner and used in the utterance of the other person. The user is provided with explanations about linguistic knowledge that is not used in their own utterances.

  In the following, an example will be described in which the present invention is applied to a language learning support device in which a speaker (user) whose native language is Japanese learns English conversation, but the applicable device is not limited to this, and the user and others It can be applied to any device that supports the learning of the language expressed in a dialogue with a person.

  For example, you may apply to the apparatus which learns the conversation between languages other than Japanese and English. In addition, devices that allow Japanese speakers to learn correct Japanese, devices that support the learning of industry terms and honorifics in specific industries, devices that support the learning of Japanese by young people, etc. The present invention may be applied to a device that supports learning related to a certain language. Moreover, you may apply to the apparatus which supports learning of foreign languages other than Japanese. Further, the number of languages that can be learned is not limited to one, and a plurality of languages may be learned.

  FIG. 1 is a block diagram showing a configuration of a language learning support apparatus 100 according to the present embodiment. As shown in FIG. 1, the language learning support apparatus 100 includes a language knowledge storage unit 121, a comment storage unit 122, a history storage unit 123, an association storage unit 124, a situation storage unit 125, and an input unit 101. , An analysis unit 102, a speaker determination unit 103, a storage unit 104, a comment acquisition unit 105, and an output unit 106.

  The language knowledge storage unit 121 stores processing language knowledge representing language knowledge used in natural language analysis processing such as morphological analysis, syntactic structure analysis, and semantic relationship analysis. Processing language knowledge includes words, idioms, idiomatic expressions, phrase expressions, word form change rules, morphological analysis rules, syntactic structure analysis rules, semantic relation analysis rules, and the like. In the following, a case will be described as an example where processing language knowledge related to an English idiom is recorded in the language knowledge storage unit 121.

  FIG. 2 is a diagram illustrating an example of a data structure of processing language knowledge stored in the language knowledge storage unit 121. As shown in FIG. 2, each entry in the language knowledge storage unit 121 is recorded with a pre-prepared processing language knowledge with a language knowledge ID for identifying the processing language knowledge. For example, the entry of the language knowledge ID = A1 in the figure represents processing language knowledge that the English expression “take part in” is an English idiom (hereinafter referred to as “English idiom expression”). Similarly, the entry of the language knowledge ID = A4 in the drawing represents the processing language knowledge that the English expression “take up for” is an English idiom expression.

  Returning to FIG. 1, the comment storage unit 122 stores comment information prepared in advance regarding learning language knowledge such as the meaning and usage of each learning language knowledge. The learning language knowledge means processing language knowledge used for language learning among the processing language knowledge recorded in the language knowledge storage unit 121. Further, in the present embodiment, the Japanese meaning of the English idiom expression, which is the language knowledge shown in FIG. 2, the English example, and the Japanese translation thereof are stored in the explanation storage unit 122 as the explanation information of the learning language knowledge. Is done.

  FIG. 3 is a diagram illustrating an example of a data structure of the comment information stored in the comment storage unit 122. As shown in FIG. 3, each entry of the comment storage unit 122 stores a comment ID for identifying comment information, a language knowledge ID, a Japanese meaning, and an English example in association with each other. In the Japanese meaning column, the Japanese meaning for the English idiom expression, which is the corresponding learning language knowledge, is recorded. In the English example column, an English example using English idiom expression corresponding to the learning language knowledge and a Japanese translation of the English example are recorded. In the figure, an example of recording a Japanese translation in parentheses is shown.

  Returning to FIG. 1, the history storage unit 123 stores, for each learning language knowledge, history information indicating that the user has spoken an expression using the learning language knowledge. Specifically, the history storage unit 123 stores history information in which a usage history ID for identifying history information, a language knowledge ID, and a user ID are associated with each other.

  FIG. 4 is a diagram illustrating an example of a data structure of history information stored in the history storage unit 123. As shown in FIG. 4, in each entry of the history storage unit 123, the language knowledge ID of the learning language knowledge used is recorded in the language knowledge ID column, and the learning language is recorded in the user ID column. The user ID of the user who has input a linguistic expression including knowledge is recorded.

  In the figure, the user ID is represented by “Un (n is a natural number)” or “UX”. “Un (n is a natural number)” represents a user ID of a known user registered in advance. “UX” represents a user ID representing all unregistered users who are not known users. In addition, the symbol “−” indicates that the use of the corresponding learning language knowledge has been detected, as in the case where any user including the unknown user has never used the learning language knowledge. It means not.

  For example, the entry of the usage history ID = C3 in the figure represents that the learning language knowledge with the language knowledge ID = A3 is used by two users with the user ID = U2 and the user ID = U3. ing. Similarly, the entry with the usage history ID = C2 indicates that the learning language knowledge with the language knowledge ID = A2 is used only by the user with the user ID = UX, that is, the unregistered user. Furthermore, since the user ID of the entry with the usage history ID = C4 is the symbol “−”, the language knowledge for learning with the language knowledge ID = A4 is used by any user including an unknown user. Not shown.

  Returning to FIG. 1, the relevancy storage unit 124 stores relevance information representing the relevance between learning language knowledge. FIG. 5 is a diagram illustrating an example of the data structure of the relevance information stored in the relevance storage unit 124.

  As shown in FIG. 5, in each entry of the relationship storage unit 124, the relationship ID for identifying the relationship information, the language knowledge ID 1, and other learning language knowledge related to the learning language knowledge of the language knowledge ID 1. Relevance information including linguistic knowledge ID 2 for identifying (relevant knowledge) and the type of relevance is recorded. In the drawing, “basic / application”, “similarity”, “part / whole”, and “general / exception” are shown as examples of the types of relevance.

  “Basic / application” means that the learning language knowledge of the language knowledge ID2 is knowledge representing an application of the learning language knowledge of the language knowledge ID1, which is basic knowledge. For example, the entry with the relevance ID = D2 in the figure is the knowledge that the learning language knowledge with the language knowledge ID1 = A1 becomes the “basic” of the learning language knowledge with the language knowledge ID2 = A3. This indicates that the learning language knowledge with ID = A3 is knowledge that is an “application” of the knowledge with language knowledge ID1 = A1.

  “Similar” means that the learning language knowledge of the language knowledge ID 1 is similar to the learning language knowledge of the language knowledge ID 2. For example, it is assumed that the learning language knowledge of the language knowledge ID 1 is knowledge that the plural form of the English noun “index” changes irregularly to “indices”. Further, it is assumed that the learning language knowledge of the language knowledge ID 2 is knowledge that the plural form of the English noun “vetex” changes irregularly to become “vetices”. Since these are knowledge representing similar irregular changes, they can be registered in the relationship storage unit 124 as similar knowledge.

  “Partial / whole” means that the learning language knowledge of the language knowledge ID1 is included in the learning language knowledge of the language knowledge ID2. For example, in the Japanese ancient sentence, the verbs that utilize the ra-line modification are limited to four words: Yes, Iri, Hori, and Ikari. In addition, there are only two verbs that make use of the Na line transformation, “Die” and “Do n’t leave”. Therefore, the learning language knowledge indicating that the use of the limited number of applicable verbs is the La line modification utilization and the Na line modification utilization is set as language knowledge ID2, and the learning language knowledge or the Na line modification utilization regarding the La line modification utilization is used. By using the language knowledge ID for learning as language knowledge ID 1, it is possible to specify relevance information having a “partial / whole” relevance.

  “General / exception” means that the learning language knowledge of the language knowledge ID2 is knowledge that is an exception to the learning language knowledge of the language knowledge ID1. For example, it is assumed that the language knowledge for learning of the language knowledge ID 1 is knowledge that an article (definite article “the” or indefinite article “a”) always precedes English nouns and noun phrases. Further, it is assumed that the learning language knowledge of the language knowledge ID 2 is knowledge that no article is attached to English nouns having uniqueness such as “earth”, “king”, “sun”, and the like. The learning language knowledge of the language knowledge ID 2 is knowledge representing an exception related to the article to be given to the noun, and therefore, between the learning language knowledge of the language ID 1 representing general knowledge about the article to be given to the noun and the noun phrase. It can be registered in the relationship storage unit 124 as knowledge having the relationship of “general / exception”.

  Returning to FIG. 1, the situation storage unit 125 stores situation information representing an utterance situation such as the date and time when a language expression including each learning language knowledge is input. FIG. 6 is a diagram illustrating an example of the data structure of the situation information stored in the situation storage unit 125.

  As shown in FIG. 6, each entry of the situation storage unit 125 includes a situation ID for identifying situation information, a language knowledge ID, an input date, an input time, a user ID of a speaker, and a use of a conversation partner. The situation information associated with the person ID is stored.

  The language knowledge ID of the learning language knowledge included in the input language expression is recorded in the language knowledge ID. Date information at the time of inputting the language expression is recorded as the input date. In addition, the time when the language expression is input is recorded as the input time. In the speaker and the conversation partner, the user ID of the speaker who inputs the language expression and the user ID of the conversation partner are recorded based on a series of outputs from the speaker determination unit 103 (described later). .

  The situation information is configured to further store information such as a sound source direction, position information obtained by GPS (Global Positioning System), information obtained from an RFID (Radio Frequency Identification) tag installed in a use environment, and the like. Also good. The sound source direction can be obtained using any conventional technique such as a method for estimating the direction of the sound source by using the difference in time to reach the multiple microphones from the sound source or the difference in sound pressure intensity. Can do.

  In addition, the user ID of the conversation partner may be specified by recognizing the conversation partner from the contents of previous and subsequent utterances. Further, the content of the previous and next utterances may be recorded as situation information. Further, the content of the utterance may be analyzed, and the topic and the instruction expression may be further recorded as the situation information.

  The language knowledge storage unit 121, the explanation storage unit 122, the history storage unit 123, the relationship storage unit 124, and the situation storage unit 125 are an HDD (Hard Disk Drive), an optical disk, a memory card, a RAM (Random Access Memory), and the like. It can be constituted by any commonly used storage means.

  Returning to FIG. 1, the input unit 101 executes various data input processing for the language learning support device 100. For example, the input unit 101 captures a voice uttered by a user and another person, recognizes the voice, and outputs a recognition result as language input information (input character string) that is text information. For example, the input unit 101 first captures an audio signal uttered by a user into a system using a microphone or the like, converts the signal into an electric signal, and then uses an A / D (analog / digital) conversion device or the like for each minute time unit. Sampled into digital data such as a time series of waveform amplitude.

  Then, the input unit 101 applies, for example, a method such as FFT (Fast Fourier Transform) analysis to the digital data, for example, analyzes time change of the frequency and extracts feature data of the uttered voice signal. To do. Subsequently, the input unit 101 generates a recognition candidate for the utterance content by comparing and collating a standard pattern such as a phoneme or a sequence thereof prepared in advance as a dictionary in association with the recognition result and the extracted feature data. To do. As a method for comparison and collation, any conventionally used method such as HMM (Hidden Markov Model), DP (Dynamic Programming), and NN (Neural Network) can be applied. Further, the input unit 101 estimates and selects the most likely candidate using a statistical language model represented by, for example, n-gram, for the generated recognition candidate in order to improve recognition accuracy. Recognize the utterance content.

  The analysis unit 102 analyzes the input character string output from the input unit 101 with reference to the processing language knowledge stored in the language knowledge storage unit 121. Then, the analysis unit 102 specifies learning language knowledge included in the input character string. For example, the analysis unit 102 specifies the predetermined processing language knowledge as the learning language knowledge among the processing language knowledge. In the present embodiment, the analysis unit 102 performs language learning among processing language knowledge such as words, idioms, idiomatic expressions, phrase expressions, word form change rules, morphological analysis rules, syntactic structure analysis rules, and semantic relation analysis rules. Language knowledge for processing related to useful idiom is specified as learning language knowledge.

  Note that the processing language knowledge specified as learning language knowledge is not limited to idioms, such as words, idioms, idiomatic expressions, phrase expressions, word form change rules, morphological analysis rules, syntactic structure analysis rules, and semantic relation analysis rules. Of these, any one or more processing language knowledge can be specified as learning language knowledge.

  Further, the method for identifying learning language knowledge is not limited to the above. For example, the analysis unit 102 may be configured to specify the processing language knowledge that matches the language knowledge ID stored in the comment storage unit 122 as the learning language knowledge. Further, the analysis unit 102 may be configured to specify all of the processing language knowledge used at the time of analysis as learning language knowledge. Even in such a case, if only the explanation information of the processing language knowledge useful for language learning is registered in the explanation storage unit 122, the explanation of the processing language knowledge not useful for language learning is presented to the user. It will never be done.

  The analysis unit 102 is a morphological analysis process that extracts words and morphemes contained in the input character string, a syntactic analysis process that extracts syntactic structures between words and morphemes extracted based on grammatical rules and statistical information, And semantic analysis processing for extracting a semantic relationship such as dependency between extracted words and morphemes and a dependency structure. Hereinafter, a case will be described as an example where the analysis unit 102 extracts a word included in the input character string and analyzes whether or not the English idiom expression recorded in the language knowledge storage unit 121 is included.

  The speaker determination unit 103 identifies the speaker of the voice captured by the input unit 101. In the speaker determination process by the speaker determination unit 103, a method of determining a speaker by comparing feature information for identifying a specific user with feature information of input voice has been used conventionally. Any speaker determination technique can be applied. The speaker determination unit 103 identifies each of at least one specific user who uses the language learning support apparatus 100 and other users. The feature information for identifying a specific user is stored in advance in a storage unit (not shown) in the language learning support device 100 by a prior learning process or the like.

  The saving unit 104 saves history information in the history storage unit 123 and saves status information in the status storage unit 125. For example, when it is determined that the speaker of the input character string is a specific user, the storage unit 104 determines the language knowledge ID of the learning language knowledge referred to when the input character string is analyzed, and the determined use History information associated with the user ID of the user is stored in the history storage unit 123. When history information of the language knowledge ID is already recorded in the history storage unit 123, the storage unit 104 adds the user ID of the determined user to the user ID column of the recorded history information.

  The comment acquisition unit 105 determines that the speaker of the input character string is someone other than the user of the language learning support device 100, and the current user of the language learning support device 100 refers to it when analyzing the input character string. When the history information indicating that the learned learning language knowledge is used is not recorded in the history storage unit 123, the explanation information corresponding to the referenced learning language knowledge is acquired from the explanation storage unit 122.

  The output unit 106 outputs the acquired explanation information together with the input character string and the like to the user. The output unit 106 presents the commentary information to the user by at least one of character information and voice information. For example, the output unit 106 presents commentary information or the like as character information on a display device capable of displaying character information. The output unit 106 outputs a synthesized voice obtained by synthesizing commentary information, which is character information, using a voice synthesis technique that has been conventionally used, through a speaker or the like (not shown). Further, the output unit 106 can generate commentary information to be output as character information by any method conventionally used, such as a method of generating the description information by applying it to a typical perforated template.

  Next, a learning support process performed by the language learning support apparatus 100 according to this embodiment configured as described above will be described with reference to FIG. FIG. 7 is a flowchart showing the overall flow of the learning support processing in the present embodiment.

  First, the input unit 101 inputs the user ID of the current user (hereinafter referred to as user UC) of the language learning support apparatus 100 (step S701). For example, the input unit 101 inputs a user ID specified or input by a user who is currently using the language learning support device 100 from an operation unit (not shown) such as a keyboard or a numeric keypad. When a plurality of users use, for example, the user can specify his / her own user ID from a plurality of user IDs registered in advance and stored in a storage unit or the like.

  Note that the method for inputting the user ID of the current user is not limited to this. For example, when the number of users of the language learning support device 100 is limited to one, the user ID registered in advance and stored in the storage unit or the like may be acquired.

  Next, the input unit 101 determines whether or not sound has been input (step S702). If not input (step S702: NO), the input unit 101 stands by until input. When speech (hereinafter referred to as speech V) is input (step S702: YES), the input unit 101 recognizes the input speech V and generates a recognition result R representing the input character string (step S702). S703).

  Next, the analysis unit 102 refers to various processing language knowledge stored in the language knowledge storage unit 121, and executes natural language analysis processing on the recognition result R (step S704). For example, the analysis unit 102 executes a morphological analysis process for the recognition result R, a syntax analysis process for the morpheme analysis result, a semantic analysis process, and the like.

  Then, the analysis unit 102 determines whether or not learning language knowledge (hereinafter referred to as learning language knowledge K) referred to when the recognition result R is analyzed exists (step S705). For example, the analysis unit 102 is the processing language knowledge referred to among the processing language knowledge in the language knowledge storage unit 121 during the analysis process, and the language of the processing language knowledge determined in advance as the learning language knowledge A knowledge ID is temporarily stored in a storage unit or the like, and the learning language knowledge referred to is specified by referring to the storage unit. If the referenced learning language knowledge K does not exist (step S705: NO), the next voice is input and the process is repeated (step S702).

  If the referenced learning language knowledge K exists (step S705: YES), the speaker determination unit 103 determines the speaker of the voice V (hereinafter referred to as speaker U) (step S706). Next, the storage unit 104 includes situation information (hereinafter, situation information S) including the learning language knowledge K referred to, the input date and time when the corresponding voice V is input, and the user ID of the determined speaker U. Is recorded in the status storage unit 125 (step S707).

  Next, the storage unit 104 determines whether the speaker U and the user UC match (step S708). Note that the storage unit 104, for example, the user ID associated with the user feature information used when the speaker determination unit 103 determines the speaker, and the user ID of the user UC input in step S701. To determine whether the speaker U and the user UC match.

  When the speaker U and the user UC match (step S708: YES), the storage unit 104 stores all the language knowledge IDs 1 associated with the language knowledge ID 2 that matches the learning language knowledge K with the association storage unit. The history information indicating that the speaker U has used each learning language knowledge of the acquired language knowledge ID1 is stored in the history storage unit 123 (step S709).

  For example, when the learning language knowledge K is the learning language knowledge of the language knowledge ID = A2 in FIG. 2, the language knowledge ID = A2 is the learning language knowledge ID2 from the relationship storage unit 124 as shown in FIG. The relevance information of the relevance ID = D1 that matches can be searched. And the language knowledge for learning of language knowledge ID = A1 set to the language knowledge ID1 column of this relevance information can be acquired. Therefore, the storage unit 104 stores history information including the user ID of the speaker U in the history storage unit 123 as history information of learning language knowledge with the language knowledge ID = A1.

  After the history information is stored in the history storage unit 123, the next voice is input and the process is repeated without executing the comment information acquisition process and the output process (step S702). If the speaker U matches the user UC, the input speech means that the user UC has spoken, so there is no need to present learning language knowledge obtained from the utterance content to the user UC. Because.

  If the speaker U and the user UC do not match (step S708: NO), the storage unit 104 further determines whether or not the speaker U has a usage history of the learning language knowledge K (step S710). ). Specifically, the storage unit 104 has history information corresponding to the language knowledge ID of the learning language knowledge K in the history storage unit 123, and the user ID of the history information includes the speaker U's ID. It is determined whether or not a user ID is included.

  When the use history of the learning language knowledge K does not exist in the speaker U (step S710: NO), the storage unit 104 stores history information indicating that the speaker U has used the learning language knowledge K as the history storage unit 123. (Step S711). For example, when the history information of the learning language knowledge K is already recorded in the history storage unit 123, the storage unit 104 adds the user ID of the speaker U to the user ID column of the history information. For example, when the history information of the learning language knowledge K is not recorded in the history storage unit 123, the storage unit 104 associates the language knowledge ID of the learning language knowledge K with the user ID of the speaker U. The appended history information is newly generated and stored in the history storage unit 123.

  When the use history of the learning language knowledge K exists in the speaker U (step S710: YES), or after the history information is stored in step S711, the comment acquisition unit 105 sends the learning language knowledge K to the user UC. It is determined whether there is a usage history (step S712). Specifically, the comment acquisition unit 105 has history information corresponding to the language knowledge ID of the learning language knowledge K in the history storage unit 123, and the user UC is included in the user ID of the history information. It is determined whether or not the user ID is included.

  If the user UC has a usage history of the learning language knowledge K (step S712: YES), the user UC does not need to output a commentary on the learning language knowledge K, so the next voice is input. The process is repeated (step S702).

  If the user UC does not have a usage history of the learning language knowledge K (step S712: NO), the comment acquisition unit 105 explains comment information (hereinafter referred to as comment E) corresponding to the learning language knowledge K. Obtained from the storage unit 122 (step S713). Specifically, the comment acquisition unit 105 acquires comment information including the meaning of Japanese corresponding to the language knowledge ID of the learning language knowledge K, the English example, and the Japanese translation of the English example.

  Next, the comment acquisition unit 105 determines whether or not the comment E corresponding to the learning language knowledge K has been acquired (step S714). If not acquired (step S714: NO), the next voice is input and the process is repeated (step S702).

  If it can be obtained (step S714: YES), the output unit 106 presents the situation information S, the recognition result R, and the comment E to the user (step S715). Note that the output unit 106 does not need to output all these pieces of information, and it is sufficient to present at least the explanation E to the user.

  Next, a specific example of the learning support process will be described. FIG. 8 is a diagram illustrating a specific example of learning support processing by the language learning support device 100 according to the present embodiment.

  First, it is assumed that the user of the current language learning support device 100 is a speaker whose native language is Japanese and user ID = U3. Further, since the conversation partner is a speaker whose native language is English of a travel agency in the United States and has not been registered in the language learning support apparatus 100, it is assumed that the user ID = UX. In the following description, “SYSTEM” represents the language learning support apparatus 100. Furthermore, in the following, the speaker with user ID = U3 is referred to as speaker U3, and the speaker with user ID = UX is referred to as speaker UX.

  In the following example, data such as those shown in FIGS. 2, 3, 4, and 5 are stored in the language knowledge storage unit 121, the comment storage unit 122, the history storage unit 123, and the relationship storage unit 124, respectively. It is assumed that

  First, it is assumed that the speaker U3 greets “Hello” to the speaker UX of the conversation partner at time T1. At this time, the language learning support device 100 detects the voice V1 (step S702) and recognizes the voice V1 (step S703). Further, the language knowledge storage unit 121 is referenced to search for learning language knowledge included in the utterance (steps S704 and S705). However, since there is no learning language knowledge corresponding to “Hello” in the language knowledge storage unit 121 of FIG. 2 (step S705: NO), the process returns to step S702.

  Assume that the speaker UX utters “May I help you?” At time T2. In this case as well, the same processing is executed for the corresponding speech V2, but the language knowledge storage unit 121 in FIG. 2 does not have the learning language knowledge included in this utterance, and the process returns to step S702.

  Similar processing is executed for the voices V3 and V4 corresponding to the time points T3 and T4, respectively. Also in this case, since the corresponding learning language knowledge does not exist, the process returns to step S702.

  At time T5, the speaker U3 inputs the voice V5 (step S702), and it is assumed that the recognition result = “Could you TAKE some place FOR INSTANCE?” Is obtained by voice recognition (step S703). For convenience of explanation, some words in the recognition result are shown in capital letters.

  Language analysis is performed on the recognition result (step S704). In this example, it is detected that the English idiom expression corresponding to the language knowledge ID = A3 in the language knowledge storage unit 121 of FIG. 2 is used in the recognition result (step S705: YES).

  Therefore, the speaker determination process is executed, and it is detected that the voice V5 is uttered by the speaker U3 (step S706). Subsequently, the situation information indicating that the detected language knowledge ID = A3 is used by the speaker U3 at the time corresponding to the time point T5 is stored in the situation storage unit 125 (step S707).

  In this case, since the determined speaker U3 matches the current user of the language learning support device 100 (step S708: YES), the language knowledge ID used this time is referred to with reference to the relationship storage unit 124. = Learning language knowledge that is the basis of A3 is searched. In this example, it is found from the entry D2 in FIG. 5 that the language knowledge ID = A1 has a “basic / application” relationship with the learning language knowledge with the language knowledge ID = A3. Therefore, “U3” is added to the user ID column of the entry C1 corresponding to the language knowledge ID = A1 in the history storage unit 123.

  With the above processing, when a speech related to the language knowledge ID = A1 is input thereafter, the speaker U3 who has already used the language knowledge ID = A3, which is applied knowledge of the language knowledge ID = A1, is used. Can prevent commentary of language knowledge ID = A1, which is basic knowledge, from being presented.

  In this case, the comment is not output and the process returns to step S702. Thereby, it is possible to prevent the explanation from being presented for the learning language knowledge used by the user.

  Thereafter, similar processing is executed for the voices V6 and V7 corresponding to the time points T6 and T7, respectively. In this case, since the corresponding learning language knowledge does not exist, the process returns to step S702.

  For the voice V8 input at the time T8, it is detected that the recognition result includes the English idiom expression “TAKE PART IN” corresponding to the language knowledge ID = A1 (step S705: YES). However, in the process for the voice V5 at the time point T5 described above, it is detected that the speaker U3 himself has used the learning language knowledge of the language knowledge ID = A3 that is more advanced (in the application relationship) than the language knowledge ID = A1. Yes. Then, “U3” is added to the user ID column of the entry C1 corresponding to the language knowledge ID = A1 of the history storage unit 123. Therefore, by referring to the history information, it can be determined that the explanation about the language knowledge ID = A1 is not necessary for the speaker U3 (step S712: YES). Therefore, commentary information about language knowledge ID = A1 is not presented.

  The same processing is executed for the voice V9 input at time T9, but since there is no corresponding learning language knowledge, the explanation information is not presented and the process returns to step S702.

  For the speech V10 of the speaker UX input at time T10, the English idiom expression corresponding to the language knowledge ID = A4 is included in the recognition result “I'll TAKE UP FOR you to get the best hotel to see it.” It is detected that “take up for” is included (step S705). Further, it is detected that the speaker of the voice V10 is an unknown speaker UX (step S706). Further, the situation information indicating that the detected language knowledge ID = A4 is used from the speaker UX to the speaker U3 at the time corresponding to the time point T10 is stored in the situation storage unit 125 (step S100). S707).

  Then, it is determined whether there is a use history of learning language knowledge with language knowledge ID = A4 in the detected speaker UX (step S710). As shown in FIG. 4, since there is no speaker using the language knowledge ID = A4 (step S710: NO), “UX” is entered in the user ID column of the entry corresponding to the language knowledge ID = C4 in the history storage unit 123. Is added (step S711).

  The speaker U3 who is the current user has no record of learning language knowledge with the language knowledge ID = A4 (step S712: NO), so the comment acquisition unit 105 sends the corresponding comment information from the comment storage unit 122. Obtain (step S713). Specifically, an entry with the comment ID = B4 corresponding to the language knowledge ID = A4 is detected from the comment storage unit 122 as shown in FIG.

  Since the corresponding commentary information has been acquired (step S714), at time T11, the output unit 106 presents a commentary that includes the meaning of the English idiom expression in Japanese, an English example, and its Japanese translation to the current user. Is presented to the speaker U3 (step S715).

  As shown in FIG. 8, the output unit 106 applies the acquired Japanese meaning, English example, and Japanese translation to a predetermined template, so that “TAKE UP FOR "The English idiom means" to take on. " For example, “She TOOK UP FOR her friend.” Can mean “She took a friend”. The comment E1 such as “is output.

  Through the above process, when speech containing learning language knowledge that the user has never used before is input, commentary information including the meaning and examples of the learning language knowledge is appropriately presented to the user It becomes possible to do.

  The same processing is executed for the voice V11 input at time T12, but since there is no corresponding learning language knowledge, the explanation information is not presented and the process returns to step S702.

  As described above, in the language learning support device according to the present embodiment, the linguistic expression generated by the user and the linguistic expression generated by a person other than the user are distinguished and constantly monitored, and the analysis result is recorded. Then, it detects that a person other than the user has issued an expression including learning language knowledge that has never been used by the user, and presents the user with a commentary on the learning language knowledge.

  This makes it possible to efficiently learn expressions encountered in actual communication. In addition, since the usage history of learning language knowledge of the user is recorded and no explanation is output for the learning language knowledge used in the past, useless education about the expressions and learning language knowledge already used by the user Can be avoided to reduce the efficiency of learning.

  Heretofore, examples have been presented in which commentary information is presented immediately when an expression containing learning language knowledge that the user does not know is encountered. However, since information related to the utterance status is recorded in the status storage unit 125, it is possible to learn later based on the information recorded in the status storage unit 125. In this case, the content of the actual dialogue may be further recorded or recorded and stored in a storage unit or the like so that it can be referred to during subsequent learning.

  In addition, the case where voice information representing an utterance has been input has been described as an example, but it is input in the form of text information such as mail, chat, and documents, and in the form of image information input by a scanner or camera. The same processing can be realized for the language information. Similarly, the present invention can be applied to a case where handwritten characters are recognized and text information is input.

  Next, the hardware configuration of the language learning support apparatus according to the present embodiment will be described with reference to FIG. FIG. 9 is an explanatory diagram showing a hardware configuration of the language learning support device according to the present embodiment.

  The language learning support apparatus according to the present embodiment is a communication I that communicates with a control device such as a CPU (Central Processing Unit) 51 and a storage device such as a ROM (Read Only Memory) 52 and a RAM 53 by connecting to a network. / F54 and a bus 61 for connecting each part.

  The language learning support program executed by the language learning support device according to the present embodiment is provided by being incorporated in advance in the ROM 52 or the like.

  The language learning support program executed by the language learning support device according to the present embodiment is a file in an installable format or an executable format, and is a CD-ROM (Compact Disk Read Only Memory), a flexible disk (FD), a CD. The recording medium may be recorded and provided on a computer-readable recording medium such as -R (Compact Disk Recordable) or DVD (Digital Versatile Disk).

  Further, the language learning support program executed by the language learning support device according to the present embodiment is stored on a computer connected to a network such as the Internet and is provided by being downloaded via the network. Also good. Moreover, the language learning support program executed by the language learning support device according to the present embodiment may be provided or distributed via a network such as the Internet.

  The language learning support program executed by the language learning support device according to the present embodiment includes a module configuration including the above-described units (input unit, analysis unit, speaker determination unit, storage unit, comment acquisition unit, output unit). As actual hardware, the CPU 51 reads out the language learning support program from the ROM 52 and executes it, so that the above-described units are loaded onto the main storage device, and the respective units are generated on the main storage device. ing.

  As described above, the apparatus and the method according to the present invention are suitable for an apparatus and a method for learning a linguistic expression input by interaction with others.

It is a block diagram which shows the structure of the language learning assistance apparatus concerning this Embodiment. It is a figure which shows an example of the data structure of the language knowledge for a process memorize | stored in the language knowledge memory | storage part. It is a figure which shows an example of the data structure of the comment information memorize | stored in the comment memory | storage part. It is a figure which shows an example of the data structure of the log | history information memorize | stored in the log | history memory | storage part. It is a figure which shows an example of the data structure of the relevance information memorize | stored in the relevance memory | storage part. It is a figure which shows an example of the data structure of the situation information memorize | stored in the situation memory | storage part. It is a flowchart which shows the whole flow of the learning assistance process in this Embodiment. It is a figure which shows the specific example of the learning assistance process by the language learning assistance apparatus of this Embodiment. It is explanatory drawing which shows the hardware constitutions of the language learning assistance apparatus concerning this Embodiment.

Explanation of symbols

51 CPU
52 ROM
53 RAM
54 Communication I / F
61 Bus 100 Language learning support device 101 Input unit 102 Analysis unit 103 Speaker determination unit 104 Storage unit 105 Explanation acquisition unit 106 Output unit 121 Language knowledge storage unit 122 Explanation storage unit 123 History storage unit 124 Relevance storage unit 125 Status storage unit

Claims (9)

A knowledge storage unit for storing processing language knowledge used in natural language analysis processing;
A commentary storage unit for storing learning language knowledge used for language learning among the processing language knowledge and commentary information representing commentary for users regarding the learning language knowledge in association with each other;
An input unit for inputting an input character string representing the utterance content of either the user or another person different from the user;
An analysis unit that performs natural language analysis processing on the input character string based on the processing language knowledge, and identifies the learning language knowledge among the processing language knowledge used in the natural language analysis processing;
A determination unit that determines whether the input character string is the utterance of the user or the other;
When the input character string is an utterance of the other person, an acquisition unit that acquires the commentary information associated with the specified learning language knowledge from the comment storage unit;
An output unit for outputting the acquired commentary information to the user;
Equipped with a,
For each learning language knowledge, the user inputs the input character string using the learning language knowledge.
A history storage unit capable of storing history information indicating that uttered,
When the input character string is the user's utterance,
A storage unit that stores the history information in the history storage unit,
The acquisition unit stores the history information of the specified learning language knowledge in the history storage unit.
If it is not memorized, the specified language knowledge for learning is determined.
Acquiring the associated commentary information from the comment storage unit;
Language learning support device characterized by The history storage unit stores the history information in which the learning language knowledge is associated with identification information for identifying the user who has uttered the input character string using the learning language knowledge,
The storage unit stores the history information in which the learning language knowledge specified and the identification information of the user are associated with each other when the input character string is an utterance of the user. Save to
The acquisition unit determines whether or not the history information in which the identified learning language knowledge is associated with the identification information of the user is stored in the history storage unit, and is not stored Obtaining the commentary information associated with the specified learning language knowledge from the comment storage unit,
The language learning support device according to claim 1 . A relevancy storage unit that stores the learning language knowledge and related knowledge representing other learning language knowledge related to the learning language knowledge in association with each other;
The storage unit further includes, when the input character string is an utterance of the user, the learning language knowledge associated with the related knowledge that matches the learning language knowledge specified, as the relevance. Acquiring from the storage unit and storing the acquired history information of the acquired learning language knowledge in the history storage unit;
The language learning support device according to claim 1 . The relevancy storage unit stores the learning language knowledge, other learning language knowledge applying the learning language knowledge, other learning language knowledge similar to the learning language knowledge, and the learning language knowledge. Storing other learning language knowledge included in part and the related knowledge that is at least one of the other learning language knowledge that is an exception to the learning language knowledge,
The language learning support device according to claim 3 . It further includes a situation storage unit that stores the input character string and information representing the situation in which the input character string is uttered, including the date and time when the input character string was uttered, and the situation information including the location. ,
The output unit further outputs the acquired commentary information and the status information to the user;
The language learning support device according to claim 1. The input unit inputs the input character string in a first language representing the speech of the other person,
The knowledge storage unit stores the processing language knowledge used in a natural language analysis process for the input character string in the first language;
The comment storage unit stores the learning language knowledge and the comment information expressed in a second language in association with each other,
When the input character string is an utterance of the other person, the acquisition unit associates the comment information expressed in the second language with the specified language knowledge for learning from the comment storage unit. Acquired,
The output unit outputs the commentary information expressed in the second language to the user;
The language learning support device according to claim 1. The input unit is configured to input the input character string obtained by recognizing a voice representing an utterance of either the user or the other person;
The language learning support device according to claim 1. The input unit inputs the input character string obtained by image recognition of an image including the input character string;
The language learning support device according to claim 1. A language learning support method executed by a language learning support device,
The language learning support device includes:
A knowledge storage unit for storing processing language knowledge used in natural language analysis processing;
A learning storage unit that stores learning language knowledge used for language learning among the processing language knowledge and commentary information representing commentary for users regarding the learning language knowledge in association with each other;
For each learning language knowledge, the user inputs the input character string using the learning language knowledge.
A history storage unit capable of storing history information indicating that uttered,
When the input character string is the user's utterance,
A storage unit that stores the history information in the history storage unit,
An input step in which the input unit inputs an input character string representing the utterance content of either the user or another person different from the user;
An analysis step, wherein the analysis unit performs a natural language analysis process on the input character string based on the processing language knowledge, and identifies the learning language knowledge among the processing language knowledge used in the natural language analysis process;
A determination step for determining whether the input character string is the utterance of the user or the other;
When the acquisition unit acquires the commentary information associated with the specified learning language knowledge from the comment storage unit when the input character string is the utterance of the other person,
An output unit that outputs the acquired commentary information to the user;
Equipped with a,
The acquisition unit stores the history information of the specified learning language knowledge in the history storage unit.
If it is not memorized, the specified language knowledge for learning is determined.
Obtaining the associated commentary information from the comment storage unit;
A language learning support method characterized by further comprising :

Images