Information processing system, information processing device, information processing method, and program

Abstract

To perform anonymization processing on text data based on voice data. [Solution] The information processing system includes a word extraction unit that extracts releasable words from text data consisting of speech data for identified speakers, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, are registered, and an output unit that outputs the releasable words corresponding to each speaker.

Description

[Technical Field]

[0001] The present invention relates to an information processing system, an information processing device, an information processing method, and a program. [Background technology]

[0002] Demand for text analysis has been increasing in recent years. For example, analyzing the content of calls at contact centers can provide insights that can lead to improvements in customer needs and response procedures. However, when outsourcing text analysis to an external organization, it is necessary to remove personal information about customers from the text.

[0003] For example, Patent Document 1 discloses a method for removing personal information from an audio stream, the method comprising: (A) receiving an audio stream in a data processing system, the audio stream carrying (i) a first audio waveform generated by a first party during a call and (ii) a second audio waveform generated by a second party during the call; (B) determining whether a first segment of the first audio waveform contains personal information, the step (B) being performed based on (i) the content of a second segment preceding the first segment of the first audio waveform, or (ii) the content of a predetermined segment of the second audio waveform; and (C) tagging the first segment when it is determined that the first segment contains personal information. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Laid-Open No. 2010-079235 Summary of the Invention [Problem to be solved by the invention]

[0005] However, the technology described in Patent Document 1 has the problem that it is not possible to extract non-personal information from data obtained by converting audio streams (voice) into text, and it is not possible to perform anonymization processing on text data based on voice data converted from voice.

[0006] One aspect of the present invention has been made in consideration of the above points, and has an object to provide an information processing system, an information processing device, an information processing method, and a program that can perform anonymization processing on text data based on voice data. [Means for solving the problem]

[0007] The present invention has been made to solve the above-mentioned problems, and one aspect of the present invention is an information processing system that includes a word extraction unit that extracts releasable words from text data consisting of speech data in which speakers are identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output unit that outputs the releasable words associated with each speaker.

[0008] Another aspect of the present invention is an information processing device that includes a word extraction unit that extracts releasable words from text data consisting of speech data in which a speaker is identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output unit that outputs the releasable words associated with each speaker.

[0009] Another aspect of the present invention is an information processing method executed by a computer of an information processing device, which includes a word extraction process for extracting releasable words from text data consisting of speech data in which a speaker is identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output process for outputting the releasable words corresponding to each speaker.

[0010] Another aspect of the present invention is a program that causes a computer of an information processing device to execute a word extraction step of extracting releasable words from text data consisting of speech data in which speakers are identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output step of outputting the releasable words associated with each speaker. [Effects of the Invention]

[0011] According to the present invention, it is possible to perform anonymization processing on text data based on voice data. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a system configuration diagram showing an example of the configuration of an information processing system according to a first embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram showing an example of a hardware configuration of the information processing device according to the present embodiment. [Figure 3] FIG. 2 is a block diagram illustrating an example of a functional configuration of the information processing device according to the present embodiment. [Figure 4] 10 is a diagram showing an example of categories of a whitelist stored in a storage unit of the information processing device according to the present embodiment. FIG. [Figure 5] 10 is a diagram showing an example of subcategories of a whitelist stored in a storage unit of the information processing device according to the present embodiment. FIG. [Figure 6] FIG. 2 is a diagram illustrating an example of a whitelist stored in a storage unit of the information processing device according to the present embodiment. [Figure 7] FIG. 2 is a block diagram showing an example of a functional configuration of an operation terminal device according to the present embodiment. [Figure 8] 10 is a flowchart illustrating an example of a whitelist generation process in the information processing device according to the present embodiment. [Figure 9] 10 is a flowchart showing a specific example of a whitelist generation process in the information processing device according to the present embodiment. [Figure 10] 10 is a flowchart illustrating another example of the whitelist generation process in the information processing device according to the embodiment. [Figure 11] 10 is a flowchart illustrating an example of an anonymization process in the information processing device according to the present embodiment. [Figure 12] 10 is a flowchart showing a specific example of an anonymization process in the information processing device according to the present embodiment. [Figure 13] 10 is a diagram showing an example of text data based on voice data input in the information processing device according to the present embodiment. FIG. [Figure 14] 10 is a diagram showing an example of processed data obtained by anonymizing text data based on voice data in the information processing device according to the present embodiment. FIG. [Figure 15] FIG. 2 is a diagram showing an example of an output image output by the information processing device according to the embodiment. [Figure 16] FIG. 10 is a diagram showing another example of text data based on voice data input in the information processing device according to the embodiment. [Figure 17] FIG. 10 is a diagram showing another example of processed data obtained by anonymizing text data based on voice data in the information processing device according to the present embodiment. [Figure 18] FIG. 10 is a block diagram showing an example of a functional configuration of an information processing device according to a second embodiment of the present invention. [Figure 19] 10 is a flowchart illustrating an example of an anonymization process in the information processing device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0013] [First embodiment] Hereinafter, each embodiment of the present invention will be described with reference to the drawings. <Configuration of information processing system SYS> First, the configuration of the information processing system SYS will be described.

[0014] FIG. 1 is a system configuration diagram showing an example of the configuration of an information processing system SYS according to the first embodiment of the present invention. The information processing system SYS includes an information processing device 100 and an operation terminal device 200. The information processing device 100 and the operation terminal device 200 are connected to each other so as to be able to communicate with each other via a network NW. The network NW may be a home network, a private network, the Internet, an intranet, or the like. The information processing device 100 and the operation terminal device 200 may be directly connected by wire or wirelessly, may be configured as a single device including the information processing device 100 and the operation terminal device 200, or may be configured such that the functions of the information processing device 100 and the operation terminal device 200 are distributed among three or more devices.

[0015] The information processing system SYS will now be described in more detail.

[0016] The information processing system SYS is a system that acquires voice data, extracts disclosable words that do not contain personal information from text data based on the voice data based on a whitelist, and outputs the extracted words. The voice data is voice data such as calls between customers and operators at call centers, conversations in meetings, etc. The text data based on the voice data is data that has been converted from voice data into text to identify the speaker and the content of the speech using voice recognition. The whitelist is a list consisting of multiple disclosable words that do not contain personal information.

[0017] Here, the term "word" includes both a single word and a phrase consisting of multiple words, but may also be either a single word or a phrase consisting of multiple words.

[0018] With this configuration, the information processing system SYS extracts words from text data based on voice data based on a whitelist containing publicly available words that do not contain personal information, thereby enabling anonymization processing to be performed by removing personal information (information that can identify an individual) from text data based on voice data. Furthermore, the words extracted by the information processing system SYS are publicly available words that do not contain personal information, eliminating the need for data analysts to check whether personal information is included in the anonymized data, thereby reducing operational costs. Furthermore, because the information processing system SYS extracts only words registered on a whitelist from text data based on voice data, it is possible to reduce the possibility of personal information being included due to mistranslation or other factors when converting voice data into text.

[0019] Here, the anonymization process according to this embodiment does not simply delete personal information from text data based on voice data, nor is it a general method of removing personal information by identifying, extracting, deleting, masking, or otherwise removing personal information, but rather extracts words from text data based on a whitelist from which personal information has been excluded.

[0020] The operation terminal device 200 is a terminal device used by an operator. In this embodiment, an example will be described in which the operation terminal device 200 inputs voice data before anonymization processing and outputs data after anonymization processing.

[0021] Next, the hardware configuration of the information processing device 100 will be described.

[0022] <Hardware configuration> FIG. 2 is a block diagram showing an example of the hardware configuration of the information processing device 100 according to this embodiment. The information processing device 100 includes a CPU 101, a storage medium interface unit 102, a storage medium 103, an input device 104, an output device 105, a ROM 106 (Read Only Memory), a RAM 107 (Random Access Memory), an auxiliary storage unit 108, and a network interface unit 109. The CPU 101, the storage medium interface unit 102, the input device 104, the output device 105, the ROM 106, the RAM 107, the auxiliary storage unit 108, and the network interface unit 109 are connected to each other via a bus.

[0023] The CPU 101 referred to here refers to a processor in general, and includes not only a device called a CPU in the narrow sense, but also, for example, a GPU, a DSP, etc. Furthermore, the CPU 101 referred to here is not limited to being realized by a single processor, but may be realized by combining multiple processors of the same or different types.

[0024] <cpu101> The CPU 101 controls the information processing device 100 by reading and executing programs stored in the auxiliary storage unit 108, the ROM 106, and the RAM 107, and by reading various data stored in the auxiliary storage unit 108, the ROM 106, and the RAM 107 and writing the various data to the auxiliary storage unit 108 and the RAM 107. The CPU 101 also reads various data stored in the storage medium 103 via the storage medium interface unit 102 and writes the various data to the storage medium 103.

[0025] <Storage medium 103> The storage medium 103 is a portable storage medium such as a magneto-optical disk, a flexible disk, or a flash memory, and stores various data.

[0026] <Storage medium interface unit 102> The storage medium interface unit 102 is an interface for reading and writing data from and to the storage medium 103 .

[0027] <Input device 104> The input device 104 is an input device such as a mouse, a keyboard, a touch panel, a microphone, a volume control button, a power button, a setting button, and an infrared receiver.

[0028] <Output Device 105> The output device 105 is an output device such as a display unit and a speaker.

[0029] <ROM106、RAM107> The ROM 106 and RAM 107 store programs for operating the various functional units of the information processing device 100 and various data.

[0030] <Auxiliary storage section 108> The auxiliary storage unit 108 is a hard disk drive, a flash memory, or the like, and stores programs for operating each functional unit of the information processing device 100 and various data.

[0031] <Network Interface Unit 109> The network interface unit 109 has a communication interface and is connected to the network NW via wireless communication.

[0032] For example, the CPU 101 of the information processing device 100 shown in Fig. 2 corresponds to the control unit 15 in the functional configuration shown in Fig. 3. The ROM 106, RAM 107, auxiliary storage unit 108, or any combination thereof of the information processing device 100 shown in Fig. 2 corresponds to the storage unit 12 in the functional configuration shown in Fig. 3. The input device 104 and the output device 105 of the information processing device 100 shown in Fig. 2 correspond to the input unit 13 and the output unit 14 in the functional configuration shown in Fig. 3.

[0033] Although illustration and description of the hardware configuration of the operation terminal device 200 are omitted, the operation terminal device 200 has the same hardware configuration as the information processing device 100 shown in FIG.

[0034] Next, the functional configuration of the information processing device 100 will be described.

[0035] <Functional configuration of information processing device 100> FIG. 3 is a block diagram showing an example of the functional configuration of the information processing device 100 according to this embodiment. The information processing device 100 includes a communication unit 11, a storage unit 12, an input unit 13, an output unit 14, and a control unit 15. The communication unit 11, the storage unit 12, the input unit 13, the output unit 14, and the control unit 15 are connected to each other via a bus.

[0036] <Communications Department 11> The communication unit 11 has a function of communicating with the operation terminal device 200. The communication unit 11 outputs various information received from the operation terminal device 200 to the control unit 15. In addition, the communication unit 11 transmits information input from the control unit 15 to the operation terminal device 200.

[0037] <Storage section 12> The storage unit 12 is configured by a storage medium, such as a hard disk drive (HDD), flash memory, electrically erasable programmable read-only memory (EEPROM), random access read / write memory (RAM), read-only memory (ROM), or any combination of these storage media. The storage unit 12 can be, for example, a nonvolatile memory.

[0038] The storage unit 12 stores a whitelist 121, a learning model 122, and processed data 123. The whitelist 121 is a list of multiple words extracted from text data generated by speech recognition of pre-training speech data, and is composed of multiple words (also called white words) that can be made public and do not contain personal information. The categories of the whitelist 121 are set to correspond to the categories in the speech data, such as bank, telephone, credit card, customer service, etc.

[0039] The categories of the whitelist 121 may be provided with subcategories such as customer satisfaction, procedures, and operator response, which correspond to purpose-specific categories. Furthermore, the whitelist 121 may include multiple whitelists according to categories.

[0040] Furthermore, the whitelist 121 may be provided with subcategories of words extracted from a plurality of whitelists corresponding to categories, and a whitelist may exist for each category and subcategory. Each category of the whitelist 121 is a category that specifies the industry or business content of the data to be anonymized, and may be composed of words such as industry jargon specific to the category or general words. In addition, subcategories of the whitelist 121 may be distinguished as those that extract necessary words depending on the purpose of analysis. Here, words according to the purpose of the subcategory include, for example, key words used in analyzing customer satisfaction, such as "helpful" or "well understood," as well as common words that form the basis of speech.

[0041] The learning model 122 includes morphological analysis based on a library, estimation of the part of speech of each word, removal of particles such as "te-ni-o-ha", scores (quantified) based on the part of speech of words, scores (quantified) based on frequently occurring words, scores (quantified) of the importance of words using TF-IDF (Term Frequency - Inverse Document Frequency), etc., and in the case of quantified learning data, it is a machine learning model that is strong in table data such as Lightgbm, and in the case of text learning data, it is a language processing AI (Artificial Intelligence) that performs language processing using a language model such as BERT.

[0042] In addition to particles, the morphological analysis may also perform language processing such as deleting various parts of speech, such as prefixes, auxiliary verbs, adjectives, and nouns, depending on the purpose.

[0043] The processed data 123 is data obtained after anonymization processing is performed on text data based on voice data. Specifically, the processed data 123 is data consisting of one or more words extracted from text data based on voice data based on the whitelist 121, or data consisting of one or more words extracted for each speaker based on the whitelist 121.

[0044] <Input section 13> The input unit 13 is an input device such as a mouse, keyboard, or microphone connected to the information processing device 100. The input unit 13 accepts an operation input input from outside. The input unit 13 outputs an operation signal corresponding to the operation input to the control unit 15.

[0045] <Output section 14> The output unit 14 is an output device such as a display device, etc. The output unit 14 outputs the processed data 123 output from the control unit 15 to the output device or another device such as the operation terminal device 200.

[0046] <Control unit 15> The control unit 15 has a function of controlling the information processing device 100. The control unit 15 reads out various data, applications, programs, etc. stored in the storage unit 12 and controls the information processing device 100.

[0047] The processing of the control unit 15 will be described in more detail. The control unit 15 includes a text data acquisition unit 151 , a learning unit 152 , and a word extraction unit 153 .

[0048] <Text data acquisition unit 151> The text data acquiring unit 151 acquires text data based on the voice data from the operation terminal device 200 via the network NW and the communication unit 11. The text data acquiring unit 151 outputs the text data based on the voice data to the word extracting unit 153.

[0049] Instead of acquiring text data based on voice data, the text data acquisition unit 151 may acquire voice data from the operation terminal device 200 and generate text data by voice recognition of the voice data. In this case, the text data may be text data in which speakers and the contents of speech of each speaker are converted into text through speaker recognition. Here, speaker recognition may be performed by recognizing the voices of multiple speakers using existing voice separation technology or the like.

[0050] <Word Extraction Unit 153> When text data based on the voice data is input from the text data acquisition unit 151, the word extraction unit 153 reads the learning model 122 from the storage unit 12, and performs morphological analysis for each utterance of the text data using the learning model 122 to extract words and phrases. Of the extracted words and phrases, the word extraction unit 153 stores the words and phrases included in the whitelist 121 as processed data 123 in the storage unit 12.

[0051] <Study Section 152> The learning unit 152 acquires text data based on speech data for pre-learning. The learning unit 152 performs morphological analysis on each utterance of the text data to extract words and phrases. The learning unit 152 registers and stores highly important words, such as words that appear a predetermined number of times, among the extracted words and phrases in a category whitelist 121 or a subcategory whitelist 121 corresponding to the category and subcategory of the speech data.

[0052] In addition, instead of words with high importance, the learning unit 152 may quantify the extracted words and phrases using a score (quantified) based on the word's part of speech, a score (quantified) based on frequently occurring words, or a score (quantified) of word importance using TF-IDF (Term Frequency - Inverse Document Frequency), etc., and register words with a score above a predetermined value in the whitelist 121.

[0053] The whitelist 121 may be presented so that it can be edited by the operator, and the edited whitelist may be stored as the whitelist 121.

[0054] Next, various data stored in the storage unit 12 of the information processing device 100 will be described.

[0055] FIG. 4 is a diagram showing an example of categories of the whitelist 121 stored in the storage unit 12 of the information processing device 100 according to this embodiment. As shown in the figure, the categories of the whitelist 121 are associated with a "number," which is identification information for identifying the whitelist, and a "category" that represents the category. For example, the whitelist 121 with a "number" of "1" and a "category" of "bank" has registered multiple white words related to banks, such as "bank," "telephone," "customer support," "billing," "withdrawal," "deposit," and "amount."

[0056] FIG. 5 is a diagram showing an example of subcategories of the whitelist 121 stored in the storage unit 12 of the information processing device 100 according to this embodiment. As shown in the figure, the subcategories of the whitelist 121 correspond to a "number" that is identification information for identifying the whitelist, a "category" that represents the category, and a "purpose-based category" that represents the subcategory. For example, the whitelist 121 with the "number" of "1," the "category" of "bank," and the "purpose-based category" of "customer satisfaction" has registered multiple white words related to banks and customer satisfaction, such as "bank," "telephone," "customer support," "honest," "strange," "teach," "please," and "date of birth."

[0057] Next, the whitelist 121 will be described.

[0058] FIG. 6 is a diagram showing an example of the whitelist 121 stored in the storage unit 12 of the information processing device 100 according to this embodiment. As shown in the figure, the whitelist 121 associates a "number," which is identification information for identifying a white word, a "white word" representing a white word, a "category" representing a category, and a "purpose category" representing a subcategory. For example, the whitelist 121 associates and registers the following: "number" = "1," "white word" = "bank," "category" = "bank," and "purpose category" = "customer satisfaction." Also, the whitelist 121 associates and registers the following: "number" = "2," "white word" = "telephone," "category" = "bank," and "purpose category" = "customer satisfaction."

[0059] In the illustrated example, white words, categories, and subcategories are associated with each other in one whitelist 121, but instead of using one whitelist 121, a whitelist 121 may be provided for each category or subcategory, and numbers may be associated with white words and registered in the whitelist 121. In other words, the whitelist 121 may be multiple whitelists in which white words are registered according to category and subcategory.

[0060] Next, the configuration of the operation terminal device 200 will be described.

[0061] <Functional Configuration of Operation Terminal Device 200> FIG. 7 is a block diagram showing an example of the functional configuration of the operation terminal device 200 according to this embodiment. The operation terminal device 200 includes a communication unit 21, a storage unit 22, an input unit 23, an output unit 24, and a control unit 25. The communication unit 21, the storage unit 22, the input unit 23, the output unit 24, and the control unit 25 are connected to each other via a bus.

[0062] <Communications Department 21> The communication unit 21 has a function of communicating with the information processing device 100. The communication unit 21 outputs various information received from the information processing device 100 to the control unit 25. In addition, the communication unit 21 transmits information input from the control unit 25 to the information processing device 100.

[0063] <Storage section 22> The storage unit 22 is configured by a storage medium, for example, a HDD, a flash memory, an EEPROM, a RAM, a ROM, or any combination of these storage media. The storage unit 22 can be, for example, a nonvolatile memory.

[0064] <Input section 23> The input unit 23 is an input device such as a mouse, keyboard, or microphone connected to the operation terminal device 200. The input unit 23 accepts an operation input input from outside. The input unit 23 outputs an operation signal corresponding to the operation input to the control unit 25.

[0065] <Output section 24> The output unit 24 is an output device such as a display device, etc. The output unit 24 outputs the processed data 123 received from the information processing device 100 and output from the control unit 25 to the output device.

[0066] <Control unit 25> The control unit 25 has a function of controlling the operation terminal device 200. The control unit 25 reads out various data, applications, programs, etc. stored in the storage unit 22 and controls the operation terminal device 200. The control unit 25 includes an operation reception unit 251 .

[0067] <Operation Reception Unit 251> The operation receiving unit 251 receives an operation by an operator. For example, the operation receiving unit 251 receives a selection of a category or subcategory of a whitelist by the operator. The operation receiving unit 251 transmits an operation signal corresponding to the received operation to the information processing device 100 via the communication unit 21.

[0068] Next, the flow of the whitelist generation process according to this embodiment will be described.

[0069] <Whitelist generation process> FIG. 8 is a flowchart showing an example of a whitelist generation process in the information processing device 100 according to this embodiment. In step S101, the information processing device 100 acquires text data based on speech data for pre-learning. The information processing device 100 performs morphological analysis on each utterance in the text data to extract words and phrases. Next, the information processing device 100 executes the process of step S102.

[0070] In step S101, the information processing device 100 may acquire voice data and generate text data from the voice data for pre-learning by voice recognition.

[0071] In step S102, the information processing device 100 creates a list of words with high importance from among the extracted words and phrases. Then, the information processing device 100 executes the process of step S103.

[0072] In step S103, the information processing device 100 registers the listed words with high importance as a whitelist 121, and stores the whitelist 121 in the storage unit 12. Then, the whitelist generation process in FIG. 8 ends.

[0073] The whitelist generation process will now be described in more detail.

[0074] <Example of whitelist generation process> FIG. 9 is a flowchart showing a specific example of a whitelist generation process in the information processing device 100 according to this embodiment. In step S201, the information processing device 100 acquires text data based on speech data for pre-learning from the operation terminal device 200. Next, the information processing device 100 executes the process of step S202.

[0075] In step S201, the information processing device 100 may acquire speech data for pre-learning from the operation terminal device 200, and generate speech text data based on the speech data.

[0076] In step S202, the information processing device 100 acquires selection information indicating a selection by the operator for a purpose-specific category that is a category or subcategory of the whitelist 121 from the operation terminal device 200. Next, the information processing device 100 executes the process of step S203.

[0077] In step S203, the information processing device 100 performs morphological analysis for each utterance in the text data and extracts words and phrases using the learning model 122. Next, the information processing device 100 executes the process of step S204.

[0078] In step S204, the information processing device 100 uses the learning model 122 to determine whether the extracted words or phrases contain important words. If the extracted words or phrases contain words with high importance, for example, words that appear a predetermined number of times or more, the information processing device 100 determines that the extracted words or phrases contain important words (YES in step S204) and executes the process of step S205. On the other hand, if the extracted words or phrases do not contain words with high importance, for example, words that appear a predetermined number of times or more, the information processing device 100 determines that the extracted words or phrases do not contain important words (NO in step S204) and executes the process of step S206.

[0079] In step S205, the information processing device 100 associates the extracted important words with the category and / or subcategory indicated by the selection information and registers them in the whitelist. Next, the information processing device 100 executes the process of step S206.

[0080] In step S206, the information processing device 100 checks whether the text is the last text in the text data. If it is the last text in the text data (YES in step S206), the information processing device 100 executes the process of step S207. On the other hand, if it is not the last text in the text data (NO in step S206), the information processing device 100 executes the process of step S203.

[0081] In step S207, the information processing device 100 stores the whitelist in which the key words are registered as the whitelist 121 in the storage unit 12. Then, the whitelist generation process in FIG. 9 ends.

[0082] The whitelist may be presented to the operator in an editable manner, and the edited whitelist may be stored in the storage unit 12 as the whitelist 121. This will be explained with reference to FIG.

[0083] <Another example of whitelist generation process> FIG. 10 is a flowchart showing another example of the whitelist generation process in the information processing device 100 according to this embodiment. In step S301, the information processing device 100 acquires text data based on speech data for pre-learning from the operation terminal device 200. Next, the information processing device 100 executes the process of step S302.

[0084] In step S301, the information processing device 100 may acquire speech data for pre-learning from the operation terminal device 200, and generate speech text data based on the speech data.

[0085] In step S302, the information processing device 100 performs morphological analysis on each utterance of the text data to extract words and phrases. Next, the information processing device 100 executes the process of step S303.

[0086] In step S303, the information processing device 100 determines whether the extracted words or phrases contain important words. If the extracted words or phrases contain words with high importance, for example, words that appear a predetermined number of times or more, the information processing device 100 determines that the extracted words or phrases contain important words (YES in step S303) and executes the process of step S304. On the other hand, if the extracted words or phrases do not contain words with high importance, for example, words that appear a predetermined number of times or more, the information processing device 100 determines that the extracted words or phrases do not contain important words (NO in step S303) and executes the process of step S305.

[0087] In step S304, the information processing device 100 associates the extracted important words with the category and / or subcategory corresponding to the text data based on the audio data and registers them in the whitelist. Next, the information processing device 100 executes the process of step S305.

[0088] In step S305, the information processing device 100 checks whether the text is the last text in the text data. If it is the last text in the text data (step S305: YES), the information processing device 100 executes the process of step S306. On the other hand, if it is not the last text in the text data (step S305: NO), the information processing device 100 executes the process of step S302.

[0089] In step S306, the information processing device 100 stores the whitelist in which the key words are registered as a whitelist candidate in the storage unit 12. Next, the information processing device 100 executes the process of step S307.

[0090] In step S307, the information processing device 100 reads out the whitelist candidate from the storage unit 12, transmits it to the operation terminal device 200, and presents each word in the whitelist candidate to the operator in an editable manner, and accepts editing. Next, the information processing device 100 executes the process of step S308.

[0091] In step S308, when the information processing device 100 receives the edited whitelist candidate from the operation terminal device 200, it stores the whitelist candidate in the storage unit 12 as the whitelist 121. Then, the whitelist generation process in FIG. 10 ends.

[0092] Next, the flow of anonymization processing in the information processing device 100 will be described.

[0093] <Anonymization processing example> FIG. 11 is a flowchart showing an example of an anonymization process in the information processing device 100 according to this embodiment. In step S401, the information processing device 100 acquires text data based on the voice data to be anonymized from the operation terminal device 200. The information processing device 100 performs morphological analysis on each utterance of the text data to extract and list words and phrases. Next, the information processing device 100 executes the process of step S402.

[0094] In step S401, the information processing device 100 may acquire voice data from the operation terminal device 200 and generate text data from the voice data to be anonymized by voice recognition.

[0095] In step S402, the information processing device 100 reads the whitelist 121 from the storage unit 12, and extracts, from the text data input in S401, words and phrases that match the words and phrases registered in the whitelist 121. Next, the information processing device 100 executes the process of step S403.

[0096] In step S403, the information processing device 100 stores the words and phrases extracted in step S402 and registered in the whitelist 121 as anonymized processed data 123 in the storage unit 12, and outputs the processed data 123 to the operation terminal device 200. Then, the anonymization process in FIG. 11 ends.

[0097] The anonymization process will be described in more detail.

[0098] <Examples of anonymization processing> FIG. 12 is a flowchart showing a specific example of anonymization processing in the information processing device 100 according to this embodiment. In step S501, the information processing device 100 acquires text data based on the voice data to be anonymized from the operation terminal device 200. Next, the information processing device 100 executes the process of step S502.

[0099] In step S501, the information processing device 100 may acquire the voice data to be anonymized from the operation terminal device 200, and generate text data from the voice data to be anonymized by voice recognition.

[0100] In step S502, the information processing device 100 acquires selection information indicating the operator's selection of a category or subcategory in the whitelist 121 from the operation terminal device 200. Next, the information processing device 100 executes the process of step S503.

[0101] In step S503, the information processing device 100 performs morphological analysis on each utterance in the text data to extract words and phrases. Then, the information processing device 100 executes the process of step S504.

[0102] In step S504, the information processing device 100 applies the whitelist 121 to the words and phrases extracted from the text data for each utterance in accordance with the selection information. Next, the information processing device 100 executes the process of step S505.

[0103] In step S505, the information processing device 100 determines whether the words and phrases extracted from the text data for each utterance include words and phrases in the whitelist 121. If it is determined that the words and phrases extracted from the text data for each utterance include words and phrases in the whitelist 121 (step S505: YES), the information processing device 100 executes the processing of step S506. On the other hand, if it is determined that the words and phrases extracted from the text data for each utterance do not include words and phrases in the whitelist 121 (step S505: NO), the information processing device 100 executes the processing of step S507.

[0104] In step S506, the information processing device 100 extracts words and phrases registered in the whitelist 121 from the words and phrases extracted for each utterance from the text data. Next, the information processing device 100 executes the process of step S507.

[0105] In step S507, the information processing device 100 stores the words and phrases registered in the whitelist 121, among the words and phrases extracted from the text data, in the storage unit 12 as anonymized processed data 123, and outputs the processed data 123 to the operation terminal device 200. Then, the anonymization process in FIG. 12 ends.

[0106] Next, the input data will be described.

[0107] <Input data> FIG. 13 is a diagram showing an example of text data based on voice data input to the information processing device 100 according to this embodiment. The text data based on the voice data is generated from the voice data by voice recognition, and is text data for each speaker and utterance by speaker recognition. The illustrated example is an example of text data. Note that the text data based on voice data in this example is text data based on voice data of a call.

[0108] The text data is text data in which a call identification number for identifying a call, the start time of an utterance in the call, the end time of the utterance, the speaker of the utterance, and the content of the utterance are associated in chronological order. Because the text data is based on pre-anonymized audio data, it contains personal information such as "XXX Bank," name "YYY," and date of birth "June 5, 1987."

[0109] Next, the output data will be described.

[0110] <Processing data> FIG. 14 is a diagram showing an example of processed data 123 obtained by anonymizing text data based on voice data in the information processing device 100 according to this embodiment. The information processing device 100 extracts words and phrases from the example text data shown in Figure 13, and applies a whitelist to the extracted words and phrases to extract words and phrases that do not contain personal information, thereby performing anonymization processing on the input text data. 14 is output after anonymization processing based on the whitelist 121, and is data in which a call identification number for identifying an utterance, the start time of the utterance in the call, the end time of the utterance, the speaker of the utterance, and the utterance content are associated in chronological order. The utterance content in the processed data 123 is words extracted from the input data "Thank you for calling. This is XXX Bank customer support." The words "telephone," "bank," and "customer support," which are registered in the whitelist 121, are listed separated by "|." Furthermore, the utterance content in the processed data 123 is extracted as the utterance content of the output data using words and phrases extracted based on the whitelist 121, which does not include personal information, and therefore the personal information contained in the input data has been removed.

[0111] Instead of separating each word with "|", punctuation marks or spaces may be used.

[0112] Next, the output image will be described.

[0113] <Output image> FIG. 15 is a diagram showing an example of an output image output by the information processing device 100 according to this embodiment. The example shown in FIG. 15 is an example in which text data "yyy.csv" is input as input data, the category "xxxxx" of the whitelist 121 is selected as the category, the purpose-specific category "xxxxx" of the whitelist 121 is selected as the purpose-specific category (subcategory), and "utterance content" is set as the column to be anonymized in the input data. In this case, the output image is provided with a display area for each setting, and each setting is visibly displayed in each display area. The output image also has a display area for a preview, and a preview of output data obtained by performing anonymization processing based on the whitelist 121 on a portion of the input data is displayed. The output image also has an "execute" button as an operator, and selecting and operating the "execute" button executes anonymization processing on all of the input data.

[0114] After the "Execute" button is selected, the start time and end time of the call for each call identification information, the speaker that identifies the speaker, and the white words are output as an output image as shown in FIG.

[0115] <Another example of input data> FIG. 16 is a diagram showing another example of text data based on voice data to be anonymized in the information processing device 100 according to this embodiment. The example shown in the figure is text data based on voice data before it is anonymized, so the text data includes personal information such as the name "Yamada Taro," the contract number "123456," the contract details "ABC Insurance," and the insurance premium "10,000 yen."

[0116] <Another example of output data> FIG. 17 is a diagram showing another example of processed data 123 obtained by anonymizing text data based on voice data in the information processing device 100 according to this embodiment. The processed data 123 shown in Fig. 17 is a pattern with more extracted words than the example shown in Fig. 14. This is output data that has been anonymized based on the whitelist 121 in which the number of extracted white words has been increased by limiting the parts of speech to be deleted compared to the whitelist 121 applied in Fig. 14 and extracting words with high importance in purpose-specific categories.

[0117] For example, if the category "insurance" and the purpose category "operator response" are selected, important insurance-related words such as "current," "content," "insurance," "monthly," "insurance," and "fee" are registered in the white list 121 as white words corresponding to "insurance" and "operator response," and further, polite expressions according to the purpose category such as "thank you," "please give," "regards," "receive," "receive," "convey," and "please" are registered. The output data with such whitelist 121 applied is "Thank you, Mr. Yamada. Please wait a moment while we confirm the details of your contract. Sorry to have kept you waiting. Your current contract is with ABC Insurance, and the monthly premium is 10,000 yen." However, personal information is deleted, and further, words that indicate the operator's response, such as "Thank you," "Contract," "Contents," "Confirm," "We will do it," "A little," "Wait," "Please," "Wait," "Let us," "We will do it," "Current," "Contents," "Insurance," "Monthly," "Insurance," and "Fee," are output, separated by "|," for example.

[0118] As such, the information processing system SYS of this embodiment includes a word extraction unit 153 that extracts releasable words from text data consisting of speech data in which speakers are identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output unit 14 that outputs the releasable words corresponding to each speaker.

[0119] In this way, the information processing system SYS extracts disclosable words from text data based on voice data based on a whitelist containing disclosable words that do not contain personal information, thereby enabling anonymization processing of text data based on voice data. Furthermore, because the words extracted by the information processing system SYS are disclosable words that do not contain personal information, data analysts do not need to confirm whether the anonymized data contains personal information, thereby reducing operational costs. Furthermore, because the information processing system SYS extracts only words registered on the whitelist from text data based on voice data, it is possible to reduce the possibility of personal information being included due to the effects of mistranslation, etc., when converting voice data to text.

[0120] Furthermore, when deleting personal information, some personal information remains, but by extracting words other than personal information based on a whitelist that excludes personal information, it is possible to anonymize the text data without leaving any personal information behind. In addition, by applying a whitelist of categories or subcategories, it is possible to extract words that can be made public for a specific purpose, preventing important words from being missed.

[0121] In the above explanation, an example of an operator selecting a whitelist category or subcategory is described, but it is also possible to extract a predetermined number of important words from the words and phrases extracted for each utterance in the text data, and select a whitelist category or subcategory based on that predetermined number of important words.

[0122] The anonymized output data (processed data 123) can also be used, for example, to generate minutes of a meeting, analyze various data such as inquiry trends during phone calls, etc. In this case, since the data (words and phrases) used for analysis and the data (words and phrases) used for generating minutes of a meeting do not contain personal information, it is possible to generate minutes of a meeting and analyze various data without personal information.

[0123] [Second embodiment] In the second embodiment, an example of extracting words that can be made public using artificial intelligence will be described. In addition, in the second embodiment, differences from the first embodiment will be mainly described.

[0124] The configuration of the information processing system SYS according to the second embodiment is the same as that of the information processing system SYS according to the first embodiment, as shown in Fig. 1 described above. The hardware configuration of the information processing device 100A according to this embodiment is the same as that of the information processing device 100 according to the first embodiment, as shown in Fig. 2 described above. The functional configuration of the operation terminal device 200 according to this embodiment is the same as that of the operation terminal device 200 according to the first embodiment, as shown in Fig. 7 described above.

[0125] <Configuration of information processing device 100A> FIG. 18 is a block diagram showing an example of the functional configuration of an information processing device 100A according to the second embodiment of the present invention.

[0126] Information processing device 100A includes a communication unit 11, a storage unit 12A, an input unit 13, an output unit 14, and a control unit 15A. Control unit 15A includes a text data acquisition unit 151, a learning unit 152, and a word extraction unit 153A. Communication unit 11, storage unit 12A, input unit 13, output unit 14, and control unit 15A are connected to each other via a bus.

[0127] Here, in the information processing device 100A according to the second embodiment shown in FIG. 18, the same configurations as those of the information processing device 100 according to the first embodiment shown in FIG. 3 are given the same symbols, and the same explanations are used, and the description is omitted.

[0128] The information processing device 100A according to this embodiment differs from the information processing device 100 according to the first embodiment in the configurations of a storage unit 12A and a word extraction unit 153A.

[0129] <Storage section 12A> The storage unit 12A stores a whitelist 121, processed data 123, and text data 124A.

[0130] The text data 124A is text data based on the voice data acquired by the text data acquisition unit 151.

[0131] <Word Extraction Unit 153A> The word extraction unit 153A includes an artificial intelligence input unit 1531A, an artificial intelligence unit 1532A, and an artificial intelligence output unit 1533A.

[0132] <Artificial Intelligence Input Unit 1531A> The artificial intelligence input unit 1531A reads out the whitelist 121 and the text data 124A stored in the storage unit 12A, and inputs the whitelist 121 and the text data 124A to the artificial intelligence unit 1532A.

[0133] <Artificial Intelligence Division 1532A> The artificial intelligence unit 1532A extracts, for each speaker, words that can be made public and do not contain personal information, based on the text data 124A input from the artificial intelligence input unit 1531A and the whitelist 121, and generates processed data. The artificial intelligence unit 1532A stores the generated processed data as processed data 123 in the storage unit 12A.

[0134] The artificial intelligence unit 1532A is a so-called generative AI. The artificial intelligence unit 1532A may be configured using algorithms such as BERT (Bidirectional Encoder Representations from Transformers) and GPT (Generative Pretrained Transformer), for example. The artificial intelligence unit 1532A may be configured by combining a plurality of algorithms including these algorithms.

[0135] <Artificial Intelligence Output Unit 1533A> The artificial intelligence output unit 1533A reads out the processed data 123 from the storage unit 12A, and outputs the processed data 123 to the output unit 14.

[0136] Next, the flow of the anonymization process according to this embodiment will be described.

[0137] <Anonymization process flow> FIG. 19 is a flowchart showing an example of an anonymization process in the information processing device 100A according to this embodiment. In step S601, the information processing device 100A inputs, to the artificial intelligence unit 1532A, the selection information specified by the operator acquired from the operation terminal device 200, the text data 124A based on the voice data read from the storage unit 12A, and the whitelist 121. Thereafter, the information processing device 100A executes the process of step S602.

[0138] In step S602, the information processing device 100A generates processed data by extracting, for each speaker, from the text data 124A, words that can be made public and do not contain personal information, based on the text data 124A, the selection information, and the whitelist 121. Thereafter, the information processing device 100A executes the process of step S603.

[0139] In step S603, the information processing device 100A stores processed data including the disclosable words extracted for each speaker and associated with the speaker, in the storage unit 12A as the processed data 123. Thereafter, the information processing device 100A executes the process of step S604.

[0140] In step S604, the information processing device 100A causes the output unit 14 to output the processed data 123. After that, the information processing device 100A ends the anonymization process in FIG.

[0141] As such, the information processing system SYS of this embodiment includes a word extraction unit 153A that extracts releasable words from text data consisting of speech data in which speakers are identified, corresponding to each speaker, based on a whitelist in which white words, which are words that can be made public, and an output unit 14 that outputs the releasable words corresponding to each speaker. In addition, in the above information processing system, the word extraction unit 153A further includes an artificial intelligence input unit 1531A that inputs text data and a whitelist to the artificial intelligence unit 1532A, and an artificial intelligence output unit 1533A that outputs publicly available words associated with the speaker generated by the artificial intelligence unit 1532A in response to the text data and the whitelist.

[0142] In this way, the information processing system SYS extracts disclosable words from text data based on voice data based on a whitelist containing disclosable words that do not contain personal information, thereby enabling anonymization processing of text data based on voice data. Furthermore, because the words extracted by the information processing system SYS are disclosable words that do not contain personal information, data analysts do not need to confirm whether the anonymized data contains personal information, thereby reducing operational costs. Furthermore, because the information processing system SYS extracts only words registered on the whitelist from text data based on voice data, it is possible to reduce the possibility of personal information being included due to the effects of mistranslation, etc., when converting voice data to text.

[0143] Furthermore, when deleting personal information, some personal information remains, but by extracting words other than personal information based on a whitelist that excludes personal information, it is possible to anonymize the text data without leaving any personal information behind. In addition, by applying a whitelist of categories or subcategories, it is possible to extract words that can be made public for a specific purpose, preventing important words from being missed.

[0144] Furthermore, in this embodiment, by using the artificial intelligence unit 1532A, the operator can obtain the processed data 123 more easily than in the first embodiment.

[0145] In addition, the artificial intelligence input unit 1531A may be configured to input the text data 124A and the whitelist 121 to the artificial intelligence unit 1532A, and the artificial intelligence unit 1532A may extract white words associated with the speaker based on the input text data 124A and the whitelist 121, and the artificial intelligence output unit 1533A may be configured to output the publicly available words associated with the speaker generated by the artificial intelligence unit 1532A.

[0146] Furthermore, the information processing device 100A may configure a prompt equivalent to the whitelist 121, instead of the whitelist 121, as input data for the artificial intelligence unit 1532A. Here, if the selection information specified by the operator is also to be used as input data for the artificial intelligence unit 1532A, the selection information may also be input by a prompt. In this embodiment, the whitelist 121 may be generated using the learning unit 152 as in the first embodiment, or may be configured to use a previously generated whitelist.

[0147] Each embodiment of the present invention has been described in detail above with reference to the drawings, but the specific configuration is not limited to that described above, and various design changes can be made within the scope that does not deviate from the gist of the present invention.

[0148] For example, in each of the above-described embodiments, an example configured with respective devices such as information processing devices 100 and 100A has been described, but one aspect of the present invention may also be realized by a device that combines some or all of these devices, or a device that rearranges some of these devices.

[0149] The program running on the information processing device 100, 100A according to one aspect of the present invention may be a program that controls one or more processors, such as a central processing unit (CPU), to implement the functions described in the above-described embodiments and modifications of the present invention (a program that causes a computer to function). The term "computer" as used herein also includes quantum computers. Information handled by each of these devices may be temporarily stored in random access memory (RAM) during processing, and then stored in various storage devices such as flash memory and hard disk drives (HDDs), where it may be read, modified, or written by the CPU or the like as needed.

[0150] Note that part or all of the information processing devices 100 and 100A in the above-described embodiments and modifications may be implemented by a computer having one or more processors. In this case, a program for implementing the control functions may be recorded on a computer-readable recording medium, and the program may be read into and executed by a computer system.

[0151] The term "computer system" used here refers to a computer system built into the information processing device 100 or 100A, and includes hardware such as an OS and peripheral devices. The term "computer-readable recording medium" refers to portable media such as flexible disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into the computer system.

[0152] Furthermore, the term "computer-readable recording medium" may include a medium that dynamically stores a program for a short period of time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line, or a medium that stores a program for a fixed period of time, such as volatile memory within a computer system that serves as a server or client in such a case. The program may also be one that realizes part of the above-mentioned functions, or one that can realize the above-mentioned functions in combination with a program already stored in the computer system.

[0153] Furthermore, part or all of the information processing device 100 in each of the above-described embodiments and modifications may be realized as an LSI, which is typically an integrated circuit, or as a chipset. Furthermore, each functional block of the information processing device 100 in each of the above-described embodiments and modifications may be individually formed into a chip, or part or all of the functional blocks may be integrated into a chip. Furthermore, the integrated circuit method is not limited to LSI, and may be realized using a dedicated circuit and / or a general-purpose processor. Furthermore, if an integrated circuit technology that replaces LSI emerges due to advances in semiconductor technology, it is also possible to use an integrated circuit based on that technology.

[0154] While the embodiments and modifications have been described above in detail with reference to the drawings as one aspect of the present invention, the specific configuration is not limited to the embodiments and modifications, and design changes within the scope of the present invention are also included. Furthermore, various modifications of one aspect of the present invention are possible within the scope of the claims, and embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. Furthermore, configurations in which elements described in the above embodiments and modifications are substituted with elements that achieve the same effect are also included. [Explanation of symbols]

[0155] 100, 100A Information processing equipment 101 CPU 102 Storage medium interface unit 103 Storage medium 104 Input Device 105 Output Device 106 ROM 107 RAM 108 Auxiliary storage 109 Network Interface Unit 11 Communications Department 12, 12A storage section 121 Whitelist 122 Learning Model 123 Processing Data 124A Text Data 13 Input section 14 Output section 15, 15A control section 151 Text data acquisition unit 152 Learning Department 153, 153A Word Extraction Unit 1531A Artificial Intelligence Input Unit 1532A Artificial Intelligence Department 1533A Artificial Intelligence Output Unit 200 Operation terminal device 21 Communications Department 22 Memory section 23 Input section 24 Output section 25 Control Unit 251 Operation reception section NW Network SYS Information Processing System

Claims

1. a word extraction unit that extracts, from text data consisting of utterance data for which a speaker is identified, words that can be made public, in association with each speaker, based on a whitelist in which white words that can be made public are registered; an output unit that outputs the disclosable words associated with each speaker; An information processing system comprising:

2. the word extraction unit extracts the disclosable words that match the white words corresponding to a category according to selection information designated by an operator; The white list has the white words registered for each category. The information processing system according to claim 1 .

3. the text data is text data in which the speaker and the speaker's speech time, which is made up of the speaker's speech start time and the speaker's speech end time, are identified; When extracting the releasable words from the text data, the word extraction unit extracts the releasable words in association with the speaker and the speech time. The information processing system according to claim 1 .

4. the word extraction unit extracts the disclosable words in the order in which they are spoken in the text data, in association with the speaker and the speaking time; The information processing system according to claim 3 .

5. the selection information specifies the category; a learning unit that outputs white words that appear in the text data a predetermined number of times or more for each category and registers the white words in a white list; Further comprising: The information processing system according to claim 2 .

6. The whitelist is a list generated from white words extracted from the text data using any one of a score based on the part of speech of the word, the frequency of occurrence of the word, the importance of the word, TF-IDF, and BERT. The information processing system according to claim 1 .

7. The word extraction unit an artificial intelligence input unit that inputs the text data and the whitelist into an artificial intelligence unit; an artificial intelligence output unit that outputs disclosable words associated with the speaker generated by the artificial intelligence unit in response to the text data and the whitelist; Further provided with The information processing system according to claim 1 .

8. a word extraction unit that extracts, based on a whitelist in which white words that are words that can be made public are registered, words that can be made public from text data consisting of utterance data in which speakers are identified, in association with the speakers; an output unit that outputs the disclosable words associated with each speaker; An information processing device comprising:

9. An information processing method executed by a computer of an information processing device, a word extraction process for extracting, from text data consisting of utterance data for which a speaker is identified, words that can be made public, in association with each speaker, based on a whitelist in which white words that can be made public are registered; an output step of outputting the disclosable words associated with each speaker; An information processing method comprising:

10. The computer of the information processing device a word extraction step of extracting, from text data consisting of utterance data for which a speaker is identified, words that can be made public, in association with each speaker, based on a whitelist in which white words that can be made public are registered; an output step of outputting the disclosable words associated with each speaker; A program to execute.

Images