WO2024150703A1

WO2024150703A1 - Information processing system, information processing method, and method for generating learning model

Info

Publication number: WO2024150703A1
Application number: PCT/JP2023/047268
Authority: WO
Inventors: 秀明渡辺
Original assignee: ソニーグループ株式会社
Priority date: 2023-01-11
Filing date: 2023-12-28
Publication date: 2024-07-18

Abstract

An information processing system according to one embodiment of the present disclosure comprises a management index estimation unit that, on the basis of voice information of a user, estimates a management index value relating to the condition of a disease of the user.

Description

Information processing system, information processing method, and learning model generation method

This disclosure relates to an information processing system, an information processing method, and a method for generating a learning model.

For diseases such as asthma, which require continual visits to the hospital to control the condition, it is important to correctly understand the condition in the patient's daily life and take appropriate measures to prevent the condition from worsening. Currently, peak flow meters and asthma diaries are used by asthma patients to understand their condition on a daily basis. Typically, asthma patients measure their condition with a peak flow meter several times a day, recording each measurement in an asthma diary to understand their condition in their daily lives.

Peak flow meters are inexpensive, easy-to-use devices approved by medical insurance that can measure peak flow values. Peak flow values are the maximum instantaneous speed of airflow when breathing out with all one's might, and are a numerical value that makes it possible to objectively grasp the state of asthma. They are used by doctors as reference information for confirming treatment plans and diagnosis, and by patients as an indicator for daily management. Patients record the measured peak flow values, as well as the state of their daily life, such as the occurrence of attacks and medication status, in an asthma diary.

Special Publication No. 2007-530327 Special Publication No. 2022-500713

However, because the above-mentioned methods of managing disease in daily life place a burden on patients, patients are often unable to continue managing their condition, making it difficult for them to correctly know their condition on a daily basis. For example, when measuring peak flow values with a peak flow meter, patients must breathe out with all their might, which places a great physical burden on the patient. In addition, many patients are unable to continue measuring every day due to the effort required or forgetting to measure, and it is often not possible to continuously record the necessary data.

In this disclosure, we propose an information processing system, information processing method, and learning model generation method that can quantitatively grasp a disease condition while minimizing the burden on the user.

The information processing system according to one embodiment of the present disclosure includes a management index estimation unit that estimates a management index value related to a user's medical condition based on the user's voice information.

In one embodiment of the information processing method disclosed herein, a computer estimates a management index value related to a user's medical condition based on the user's voice information.

In one embodiment of the present disclosure, a method for generating a learning model involves a computer generating a management index estimation model for estimating a management index value related to a user's medical condition based on the user's voice information.

FIG. 1 is a diagram illustrating a configuration example of an information processing system according to an embodiment of the present disclosure. FIG. 1 is a diagram illustrating an example of a configuration of a portion of an information processing device according to an embodiment of the present disclosure. 11 is a flowchart illustrating an example of a management index estimation process according to an embodiment of the present disclosure. FIG. 11 is a diagram for explaining an example of a management index estimation process according to an embodiment of the present disclosure. 13 is a flowchart illustrating an example of a generation process of a management index estimation model according to an embodiment of the present disclosure. 11 is a diagram for explaining an example of a generation process of a management index estimation model according to an embodiment of the present disclosure. FIG. 11 is a flowchart illustrating an example of a response process according to an embodiment of the present disclosure. FIG. 11 is a diagram for explaining a first modified example of an information processing system according to an embodiment of the present disclosure. FIG. 11 is a diagram for explaining a second modified example of an information processing system according to an embodiment of the present disclosure. FIG. 11 is a diagram for explaining a third modified example of an information processing system according to an embodiment of the present disclosure. FIG. 11 is a diagram for explaining a fourth modified example of an information processing system according to an embodiment of the present disclosure. FIG. 2 is a diagram illustrating an example of a hardware configuration according to an embodiment of the present disclosure.

Below, an embodiment of the present disclosure will be described in detail with reference to the drawings. The embodiment includes examples and modified examples. Note that the embodiment does not limit the systems, devices, methods, etc. related to the present disclosure. Furthermore, in the following embodiments, essentially the same parts are designated by the same reference numerals, and duplicated explanations will be omitted.

The following one or more embodiments can be implemented independently. However, at least a portion of the following embodiments may be implemented in appropriate combination with at least a portion of the other embodiments. These embodiments may include novel features that are different from one another. Thus, each embodiment may contribute to solving a different purpose or problem, and may provide different effects.

The present disclosure will be described in the following order.
1. Embodiment 1-1. Example of the configuration of an information processing system 1-2. Example of the configuration of a part of an information processing device 1-3. Example of a management index estimation process 1-4. Example of a management index estimation model generation process 1-5. Example of a response process 1-6. Modified example of an information processing system 1-7. Effects 2. Other embodiments 3. Example of the hardware configuration 4. Supplementary notes

1. Embodiment
<1-1. Example of information processing system configuration>
A configuration example of an information processing system 1 according to the present embodiment will be described with reference to Fig. 1. Fig. 1 is a diagram showing a configuration example of an information processing system 1 according to the present embodiment. In the example of Fig. 1, the information processing system 1 functions as a voice dialogue system that supports disease treatment.

As shown in FIG. 1, the information processing system 1 includes a sound input unit 10, a biometric information detection unit 20, a sound output unit 30, a user terminal 40, and an information processing device 50. Various information is transmitted and received between the sound input unit 10, the biometric information detection unit 20, the sound output unit 30, the user terminal 40, and the information processing device 50. This transmission and reception is performed via wireless and/or wired communication networks, wiring, etc.

The sound input unit 10 detects sounds such as voice and inputs them to the information processing device 50. For example, the sound input unit 10 detects user speech and inputs it to the information processing device 50. For example, a microphone is used as the sound input unit 10.

User utterances are voices that a user speaks to the information processing system 1 to obtain a response. For example, a user may say, "What's the weather going to be like in Tokyo tomorrow?" or "What are your plans for today?" Voice information related to user utterances is an example of user voice information.

The biometric information detection unit 20 detects the user's biometric information and inputs it to the information processing device 50. For example, a wearable device is used as the biometric information detection unit 20. There are various types of wearable devices, such as wristband type, neckband type, and earphone type.

User biometric information is biometric information obtained from a user. This user biometric information is collected implicitly when the user gives permission to the information processing system 1 to collect the information in advance. For example, the user biometric information includes heart rate, sleep state, amount of exercise, pulse, blood pressure, blood flow, etc., collected by the biometric information detection unit 20 worn by the user.

The sound output unit 30 outputs sounds such as voice. For example, the sound output unit 30 outputs sounds based on response information or the like. For example, a speaker such as a smart speaker is used as the sound output unit 30.

The user terminal 40 is a terminal for a user. The user terminal 40 presents various information to the user through displays, sounds, etc. An example of the user terminal 40 is a smartphone.

The information processing device 50 has a management index estimation unit 51, a management index database 52, a voice recognition unit 53, a semantic analysis unit 54, a response generation database 55, a response generation unit 56, and a response control unit 57.

The control index estimation unit 51 estimates a control index value related to the user's medical condition by analyzing the user's speech and the user's biometric information, and outputs control index information related to the control index value. Note that if the control index estimation unit 51 is unable to obtain user biometric information, it may estimate the control index value by analyzing only the user's speech.

The management index value is an objective numerical value for understanding and managing the user's medical condition. Furthermore, the management index information is information that includes the management index value estimated by the management index estimation unit 51.

The management index database 52 is a database that records the management index information (user's management index value) output from the management index estimation unit 51.

Here, examples of management index values for diseases such as asthma include forced vital capacity (FVC), forced expiratory volume in one second (FEV1), rate of expiratory volume in one second (FEV1%), predicted forced expiratory volume in one second (%FEV1), and peak flow (PEF), all of which are measured using a spirometer. Any one or more of these may be used as the management index value.

The voice recognition unit 53 converts the voice spoken by the user into a spoken string. The spoken string is a character string spoken by the user.

The semantic analysis unit 54 analyzes the spoken string generated by the speech recognition unit 53 to generate the first response generation information required by the response generation unit 56 to generate response information.

The first response generation information is information that is formed by analyzing the intention of the user's utterance by the semantic analysis unit 54 so that the response generation unit 56 can generate response information. For example, the first response generation information for a user utterance of "What's the weather in Tokyo today?" is "Subject: weather, date and time: today, location: Tokyo", etc.

The response generation database 55 is a database that records the second response generation information required by the response generation unit 56 to generate response information. The response generation database 55 stores in advance the information required to generate response information.

The second response generation information is information that is formed so that the response generation unit 56 can generate response information together with the first response generation information. For example, the second response generation information includes environmental information about the user, such as weather forecasts and air pollution information, as well as the user's schedule information. Environmental information about the user includes, for example, environmental information about the user's living area, including the user's home, workplace, shopping destinations, etc., but may also include environmental information about the user's travel destinations, etc. The user's schedule information includes information about the user's plans, such as the content of the errand, date and time, and location.

The response generation unit 56 generates response information for the user from the first response generation information input from the semantic analysis unit 54, the management index information input from the management index database 52, and the second response generation information input from the response generation database 55.

The response information includes various types of information for responding to the user. The various types of information include, for example, information according to the content of the user's utterance, information about the user's current condition, and reference information for preventing the user's condition from worsening in the future. For example, if the user utterance is "What's the weather in Tokyo today?", the response information may be "It's sunny today. Your asthma has been getting worse since yesterday, so be careful."

The response control unit 57 performs a system response based on the response information input from the response generation unit 56, providing information tailored to the device used by the user.

For example, if the device used by the user is a sound input unit 10 such as a smart speaker, the response control unit 57 responds by voice, and if the device used by the user is a user terminal 40 such as a smart watch or smartphone, the response control unit 57 provides a notification by text.

Here, each of the functional units, such as the above-mentioned management index estimation unit 51, speech recognition unit 53, semantic analysis unit 54, response generation unit 56, and response control unit 57, may be configured with both hardware and software, or one of them. Their configuration is not particularly limited. For example, each of the above-mentioned functional units may be realized by a computer, such as a CPU (Central Processing Unit) or MPU (Micro Control Unit), executing a program pre-stored in ROM using a RAM or the like as a working area. Furthermore, each of the functional units may be realized with an integrated circuit, such as an ASIC (Application Specific Integrated Circuit) or FPGA (Field-Programmable Gate Array).

<1-2. Example of a Part of the Configuration of an Information Processing Device>
An example of the configuration of a portion of the information processing device 50 according to this embodiment will be described with reference to Fig. 2. Fig. 2 is a diagram showing an example of the configuration of a portion of the information processing device 50 according to this embodiment.

As shown in FIG. 2, the management index estimation unit 51 estimates the management index value based on the management index estimation model 58a. This management index estimation model 58a is generated by the model generation unit 58.

The model generation unit 58 generates a management index estimation model 58a by machine learning, for example, based on the speech voice database 58b and the management index measurement value database 58c. The management index estimation model 58a is, for example, a model that performs regression analysis on data collected in advance.

Such a model generation unit 58 may be provided in the information processing device 50, or may be provided in a device other than the information processing device 50.

<1-3. Example of management index estimation process>
An example of the management index estimation process according to the present embodiment will be described with reference to Fig. 3 and Fig. 4. Fig. 3 is a flowchart showing an example of the management index estimation process according to the present embodiment. Fig. 4 is a diagram for explaining an example of the management index estimation process according to the present embodiment.

As shown in FIG. 3, in step S11, the control index estimation unit 51 acquires the user's speech input from the sound input unit 10. In step S12, the control index estimation unit 51 calculates acoustic features from the acquired user's speech. In step S13, the control index estimation unit 51 estimates a control index value from the aforementioned acoustic features using the control index estimation model 58a. In step S14, the control index estimation unit 51 outputs the estimated control index value, i.e., the control index estimated value, to the control index database 52, and ends the process.

In this process, the management index estimation unit 51 estimates the management index value from the user's speech, i.e., the user's speech, in order to reduce the measurement load on the user and to implicitly grasp the user's condition while using the information processing system 1.

As shown in FIG. 4, specifically, the management index estimation unit 51 processes the user's speech through an acoustic feature calculation process to calculate the acoustic feature.

An acoustic feature is a numerical value (vector) that represents the characteristics of a sound. Examples of acoustic features include MFCC (Mel Frequency Cepstral Coefficients), zero cross, spectral centroid, spectral flatness, and spectral rolloff.

The control index estimation unit 51 estimates the control index value by performing a control index value estimation process using the above-mentioned acoustic feature amount and the control index estimation model 58a obtained by prior learning. This allows the control index value, i.e., the control index estimated value, to be obtained.

The control index estimation process is a process of calculating a control index estimate value by regression using, for example, control index estimation model 58a. Control index estimation model 58a is, for example, a model that performs regression analysis on data collected in advance using acoustic features as explanatory variables and control index values as objective variables. This control index estimation model 58a is generated in advance by machine learning or the like.

<1-4. Example of management index estimation model generation process>
An example of the management index estimation model generation process according to the present embodiment will be described with reference to Fig. 5 and Fig. 6. Fig. 5 is a flowchart showing an example of the management index estimation model generation process according to the present embodiment. Fig. 6 is a diagram for explaining an example of the management index estimation model generation process according to the present embodiment.

As shown in FIG. 5, in step S21, the model generation unit 58 acquires the patient's speech data from the speech database 58b, and acquires the management index measurement values corresponding to the speech data from the management index measurement value database 58c. In step S22, the model generation unit 58 calculates acoustic features from the acquired speech data. In step S23, the model generation unit 58 performs model learning to generate a management index estimation model 58a using the calculated acoustic features and management index measurement values. In step S24, the model generation unit 58 stores the generated management index estimation model 58a.

In such processing, the model generation unit 58 generates the management index estimation model 58a in advance. Note that the management index estimation model 58a may be caused to execute re-learning using a user utterance and an estimated management index value corresponding to the user utterance, thereby updating the management index estimation model 58a.

As shown in FIG. 6, specifically, the model generation unit 58 collects in advance, for learning purposes, pairs of speech data (speech information) from asthma patients with various symptoms and the control index measurement values corresponding to the speech data. The control index measurement values corresponding to the speech data are, for example, control index measurement values measured by spirometry when the corresponding speech data is acquired. Speech data for patients with different symptoms and control index measurement values for each of the speech data are prepared in advance.

The model generation unit 58 calculates acoustic features from the speech data, estimates a control index value from the acoustic features, and performs learning so as to minimize the error between the estimated control index value and the control index measurement value corresponding to that control index value, thereby generating a control index estimation model 58a.

(User utterance)
In this embodiment, the user's speech is analyzed, but if the system can collect the sounds, coughing and breathing sounds can also be analyzed. Therefore, coughing and breathing sounds are included in the speech. In addition, if the system can acquire user biometric information, vital signs such as heart rate and respiratory rate, and facial images can also be analyzed to estimate the management index value.

(Acoustic features)
In this embodiment, 12-dimensional MFCC (Mel Frequency Cepstral Coefficients) are calculated for input data at regular time frames, and the average of the calculated values is used as the acoustic feature. However, it is also possible to use other acoustic features, such as statistics such as the median, maximum value, minimum value, standard deviation, skewness, and kurtosis, as well as mel spectrogram, chroma vector, zero cross, spectral centroid, spectral flatness, and spectral rolloff.

(Control index value)
In this embodiment, the use of, for example, peak flow (PEF) is assumed as the control index value, but the use of forced vital capacity (FVC), forced expiratory volume in one second (FEV1), rate in one second (FEV1%=FEV1/FVC), and forced expiratory volume in one second relative to a predicted value (%FEV1) are also assumed. The priority order of the control index values is as follows: (1) rate in one second, (2) forced expiratory volume in one second relative to a predicted value, (3) peak flow, (4) forced vital capacity, (5) forced vital capacity relative to a predicted value, and (5) forced expiratory volume in one second. The smaller the number in parentheses, the higher the priority order. This priority order is, for example, the importance of understanding the condition.

Respiratory diseases include asthma (bronchial asthma), but other respiratory diseases include chronic obstructive pulmonary disease (COPD) and pulmonary fibrosis. The management index value for chronic obstructive pulmonary disease is, for example, forced expiratory volume in 1 second (FEV1%). The management index value for pulmonary fibrosis is, for example, forced vital capacity (FVC). There are also various diseases such as eosinophilic granulomatosis with polyangiitis (EGPA), sarcoidosis (lung), collagen disease of the lung, metastatic lung tumor, bronchiectasis, and foreign bodies in the airway. The management index value is set appropriately according to the type of disease.

The management index value (management index estimated value) is estimated by the management index estimation unit 51 and stored in the management index database 52 for each user. For example, the management index value is stored for each user utterance and managed for each user. Such management index information may be configured to be viewable by the user's doctor, family, etc. For example, the user's management index information may be appropriately transmitted to a terminal of the user's doctor, family, etc. in response to access from that terminal.

(Model learning)
In this embodiment, linear regression is used as the regression model, but it is also assumed that gradient boosting decision tree, support vector regression, model learning by deep learning, etc. are also used. In this embodiment, the system learns using data collected in advance and estimates the management index value based on the learned model, but it is also assumed that a function for personal optimization of the model by re-learning using the voice input when the user uses the service and for improving the accuracy of the entire model are introduced.

<1-5. An example of response processing>
An example of the response process according to this embodiment will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the response process according to this embodiment.

As shown in FIG. 7, in step S31, the response generation unit 56 acquires response generation information (first response generation information) from the semantic analysis unit 54. In step S32, the response generation unit 56 refers to the management index database 52 and determines whether the management index value is available.

If the response generation unit 56 determines in step S32 that the management index value is available (Yes in step S32), in step S33, it refers to the response generation database 55 and determines whether there is available information regarding exacerbation in the response generation database 55.

On the other hand, if the response generation unit 56 determines in step S32 that the management index value is not available (No in step S32), normal response information is generated in step S36, and the process ends.

In step S33, if the response generation unit 56 determines that there is available information regarding exacerbation in the response generation database 55 (Yes in step S33), in step S34, response information including information regarding the exacerbation (worsening of the condition) is generated, and the process ends.

On the other hand, if in step S33 the response generation unit 56 determines that there is no information regarding exacerbation available in the response generation database 55 (No in step S33), in step S35 it generates response information including information regarding the condition, and ends the process.

According to this processing, the response generation unit 56 generates response information for the user based on the first response generation information input from the semantic analysis unit 54, the management index information input from the management index database 52, and the second response generation information input from the response generation database 55.

(Normal response)
The response generating unit 56 generates normal response information when it is unable to acquire a management index value from the management index database 52. For example, when the user utterance is "What's the weather in Tokyo today?", the response generating unit 56 generates the response information "It's sunny today."

(Response regarding medical condition)
When the response generating unit 56 can acquire a management index value from the management index database 52, the response generating unit 56 generates response information including information on the medical condition in addition to normal response information. For example, when the user utterance is "What's the weather in Tokyo today?", the response generating unit 56 generates response information saying "It's sunny today. Your current asthma condition is good."

In addition, when the most recent continuous information can be acquired from the management index database 52, the response generation unit 56 may generate response information according to changes in the most recent continuous information. For example, when the user utterance is "What's the weather in Tokyo today?", the response generation unit 56 generates response information such as "It's sunny today. Your asthma condition has been worsening since yesterday, so be careful."

(Responses regarding exacerbations)
When the response generating unit 56 can obtain environmental information such as the user's schedule, pollen, and air pollution together with the management index value, the response generating unit 56 generates response information including information for suppressing exacerbation together with normal response information. For example, when the user utterance is "What's the weather in Tokyo today?", the response generating unit 56 generates response information such as "It's sunny today. Your current asthma condition is good, but there is a lot of pollen scattered, so be careful when going out." Also, when the user utterance is "What's the weather in Tokyo today?", the response generating unit 56 generates response information such as "It's sunny today. We're planning to have dinner from 19:00, but your asthma condition has been getting worse since yesterday, so be careful not to drink too much alcohol."

(Information provided by smart speakers and AI agents)
In the information processing system 1 based on providing information by voice, responses to the user are provided by voice, for example, using a sound output unit 30. Note that, since voice input to the information processing system 1 is also based on user speech, the management index value is estimated using voice input by the sound input unit 10.

(Information provided via smartphones and wearable devices)
In the information processing system 1 based on providing information by displaying a screen, responses to the user may be provided by notifying a message on the screen. For example, a user terminal 40 or the like is used. Note that the sound input unit 10 is used for inputting voice to the information processing system 1, and the user terminal 40 is used for providing information by displaying a screen, but for example, the user terminal 40 such as a smartphone may be used for inputting voice to the information processing system 1. In this case, the user terminal 40 functions as the sound input unit 10.

As described above, according to this embodiment, the condition of a user who uses the voice dialogue system, which is the information processing system 1, can be quantitatively understood while reducing the burden on the user by estimating the condition of the disease by performing analysis using the user's speech to the system. In addition, the system can generate information to prevent the condition from worsening based on the condition estimated from the user's speech, the user's predicted behavior information, and environmental information such as weather, and present it to the user in an easy-to-understand manner. Therefore, the user can measure the condition of the disease unconsciously and without physical burden through voice dialogue with the system, and can receive information required to prevent the condition from worsening from the system. Conventionally, even if an asthma diary was written, it was difficult for the patient to determine from the information what to do to prevent the condition from worsening, and the reality is that the patient has no choice but to maintain the current condition until the next visit to the hospital to hear the doctor's diagnosis.

<1-6. Modified examples of information processing system>
Modifications of the information processing system 1 according to the present embodiment will be described with reference to Fig. 8 to Fig. 11. Fig. 8 to Fig. 11 are diagrams showing configuration examples of the information processing systems 1A to 1D according to the modifications of the present embodiment.

As shown in FIG. 8, the information processing system 1A includes a wearable device 110, a smartphone 120, and a server 200. The wearable device 110 and the smartphone 120 are configured to be able to communicate with the server 200. Data detected by the wearable device 110 is sent directly to the server 200.

As shown in FIG. 9, information processing system 1B, like information processing system 1A described above, includes wearable device 110, smartphone 120, and server 200. Wearable device 110 is configured to be able to communicate with smartphone 120, and smartphone 120 is configured to be able to communicate with server 200. Data detected by wearable device 110 is first stored in smartphone 120 and then transmitted to server 200.

In Figs. 8 and 9, the wearable device 110 corresponds to the biometric information detection unit 20 described above, and the smartphone 120 corresponds to the sound input unit 10 and sound output unit 30 described above. In other words, the smartphone 120 includes the sound input unit 10 and sound output unit 30. The server 200 corresponds to the information processing device 50.

As shown in FIG. 10, the information processing system 1C includes a wearable device 110, a smartphone 120, a server 200, and a service provider server 300. The wearable device 110 and the smartphone 120 are configured to be able to communicate with the server 200. The server 200 is configured to be able to communicate with the service provider server 300. Data detected by the wearable device 110 is sent directly to the server 200.

As shown in FIG. 11, information processing system 1D, like information processing system 1C described above, comprises a wearable device 110, a smartphone 120, a server 200, and a service provider server 300. The wearable device 110 is configured to be able to communicate with the smartphone 120, and the smartphone 120 is configured to be able to communicate with the server 200. The server 200 is configured to be able to communicate with the service provider server 300. Data detected by the wearable device 110 is first stored in the smartphone 120 and then transmitted to the server 200.

10 and 11, the wearable device 110 corresponds to the biometric information detection unit 20 described above, and the smartphone 120 corresponds to the sound input unit 10 and sound output unit 30 described above. In other words, the smartphone 120 includes the sound input unit 10 and sound output unit 30. One or both of the server 200 and the service provider server 300 corresponds to the information processing device 50.

In the examples of Figures 10 and 11, the device owned by the service provider is shown to be a server, i.e., the service provider server 300. However, the device owned by the service provider does not necessarily have to be a server, and may be an information terminal such as a smartphone, a tablet terminal, a laptop computer, or a desktop computer.

<1-7. Effects>
As described above, according to this embodiment, the information processing system 1 includes a management index estimation unit 51 that estimates a management index value related to the user's medical condition based on the user's voice information (e.g., user utterance). This makes it possible to obtain the management index value from the user's voice information, making it possible to quantitatively grasp the user's medical condition while reducing the burden on the user.

The management index value may also be a management index value related to the condition of the user's disease when the user's disease is a respiratory disease. This makes it possible to quantitatively grasp the condition of the user's disease when the user's disease is a respiratory disease.

The control index estimation unit 51 may also estimate the control index value based on the user's voice information and the user's biometric information. This allows the control index value to be determined with high accuracy.

The control index estimation unit 51 may also calculate acoustic features from the user's voice information and estimate the control index value based on the calculated acoustic features. This allows the control index value to be determined with high accuracy.

The management index estimation unit 51 may also estimate the management index value using the management index estimation model 58a, which is a learning model. This allows the management index value to be determined with high accuracy.

The information processing system 1 may further include a model generation unit 58 that generates a management index estimation model 58a. This makes it possible to reliably obtain the management index estimation model 58a.

The model generation unit 58 may also generate the management index estimation model 58a using voice information for each patient with different symptoms related to the user's disease and the management index measurement value for each voice information. This makes it possible to obtain a highly accurate management index estimation model 58a.

The model generation unit 58 may also calculate acoustic features from the user's voice information, estimate a control index value from the calculated acoustic features, and perform learning so as to minimize the error between the estimated control index value and the control index measurement value corresponding to the control index value, thereby generating a control index estimation model 58a. This makes it possible to obtain a highly accurate control index estimation model 58a.

The information processing system 1 may further include a management index database 52 that stores the management index values. This allows the management index values to be managed.

The information processing system 1 may further include a response generation unit 56 that generates response information regarding the user's medical condition based on the management index value. This allows the user to know the medical condition.

The response information may also include one or both of information regarding the user's current condition and information regarding the worsening of the user's condition. This allows the user to know the current state of the condition or the worsening of the condition.

In addition, the information regarding the worsening of the user's condition may include information for preventing the worsening of the user's condition. This can help prevent the user's condition from worsening.

The response generation unit 56 may also generate response information based on the management index value and the first response generation information. This makes it possible to obtain appropriate response information.

The first response generation information may also include information for generating response information according to the intention of the user's speech regarding the user's voice information. This makes it possible to obtain appropriate response information.

The information processing system 1 may further include a semantic analysis unit 54 that analyzes the intention of the user's utterance and generates the first response generation information. This makes it possible to obtain appropriate first response generation information.

The response generation unit 56 may also generate response information based on the management index value, the first response generation information, and the second response generation information that is different from the first response generation information. This makes it possible to obtain appropriate response information.

The second response generation information may also include one or both of environmental information about the user and schedule information about the user. This makes it possible to obtain appropriate response information.

The information processing system 1 may further include a response generation database 55 that stores the second response generation information. This allows the second response generation information to be managed.

2. Other embodiments
Each configuration and each process according to the above-mentioned embodiment (example, modified example) may be implemented in various different forms other than the above-mentioned embodiment. For example, the configuration and process may be various forms without being limited to the above-mentioned example. In addition, for example, among the processes described in the above-mentioned embodiment, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by a known method. In addition, the configurations, processing procedures, specific names, or information including various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown in the figure.

Furthermore, each configuration and each process of the above-mentioned embodiment (example, modified example) does not necessarily have to be physically configured as illustrated. In other words, the specific form of distribution and integration of each device is not limited to that illustrated, and all or part of it can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

Furthermore, the configurations and processes of the above-described embodiments (examples, modified examples) may be combined as appropriate. For example, at least a part of an embodiment may be implemented in combination with at least a part of another embodiment as appropriate. Furthermore, the effects of the embodiments are merely examples and are not limiting, and other effects may also be provided.

3. Hardware configuration example
A specific hardware configuration example of the various information devices according to the above-mentioned embodiment (or modified example) will be described. The various information devices according to the embodiment (or modified example) may be realized by a computer 500 having a configuration as shown in Fig. 12, for example. Fig. 12 is a diagram showing a hardware configuration example according to this embodiment.

As shown in FIG. 12, computer 500 has a CPU 510, RAM 520, ROM (Read Only Memory) 530, HDD (Hard Disk Drive) 540, a communication interface 550, and an input/output interface 560. Each part of computer 500 is connected by a bus 570.

The CPU 510 operates based on the programs stored in the ROM 530 or the HDD 540 and controls each part. For example, the CPU 510 loads the programs stored in the ROM 530 or the HDD 540 into the RAM 520 and executes processes corresponding to the various programs.

The ROM 530 stores boot programs such as the BIOS (Basic Input Output System) that is executed by the CPU 510 when the computer 500 starts up, as well as programs that depend on the hardware of the computer 500.

HDD 540 is a recording medium readable by computer 500 that non-temporarily records programs executed by CPU 510 and data used by such programs. Specifically, HDD 540 is a recording medium that records the information processing program related to the present disclosure, which is an example of program data 541.

The communication interface 550 is an interface for connecting the computer 500 to an external network 580 (the Internet as an example). For example, the CPU 510 receives data from other devices and transmits data generated by the CPU 510 to other devices via the communication interface 550.

The input/output interface 560 is an interface for connecting the input/output device 590 and the computer 500. For example, the CPU 510 receives data from an input device such as a keyboard or a mouse via the input/output interface 560. The CPU 510 also transmits data to an output device such as a display, a speaker, or a printer via the input/output interface 560.

The input/output interface 560 may also function as a media interface that reads programs and the like recorded on a specific recording medium. Examples of media that can be used include optical recording media such as DVDs (Digital Versatile Discs) and PDs (Phase change rewritable Disks), magneto-optical recording media such as MOs (Magneto-Optical disks), tape media, magnetic recording media, and semiconductor memories.

Here, for example, when the computer 500 functions as the information processing device 50 according to this embodiment, the CPU 510 of the computer 500 executes the information processing program loaded on the RAM 520 to realize all or part of the functions of the management index estimation unit 51, the voice recognition unit 53, the semantic analysis unit 54, the response generation unit 56, the response control unit 57, etc. Also, the information processing program and data according to this embodiment are stored in the HDD 540. Note that the CPU 510 reads and executes the program data 541 from the HDD 540, but as another example, these programs may be obtained from other devices via the external network 580.

<4. Notes>
The present technology can also be configured as follows.
(1)
a control index estimation unit that estimates a control index value related to a medical condition of the user based on voice information of the user;
Information processing system.
(2)
The management index value is a management index value related to a medical condition when the disease of the user is a respiratory system disease.
The information processing system according to (1) above.
(3)
the control index estimation unit estimates the control index value based on the voice information and the biometric information of the user.
The information processing system according to (1) or (2).
(4)
the control index estimation unit calculates an acoustic feature from the speech information, and estimates the control index value based on the calculated acoustic feature.
The information processing system according to any one of (1) to (3).
(5)
the management index estimation unit estimates the management index value using a management index estimation model which is a learning model;
The information processing system according to (4) above.
(6)
A model generation unit that generates the management index estimation model,
The information processing system according to (5) above.
(7)
The model generation unit generates the management index estimation model using voice information for each patient having a different symptom related to the user's disease and a management index measurement value for each of the voice information.
The information processing system according to (6) above.
(8)
the model generation unit calculates acoustic features from the speech information, estimates a control index value from the calculated acoustic features, performs learning so as to minimize an error between the estimated control index value and the control index measurement value corresponding to the estimated control index value, and generates the control index estimation model.
The information processing system according to (7) above.
(9)
A management index database for storing the management index values is further provided.
The information processing system according to any one of (1) to (8).
(10)
A response generating unit that generates response information regarding the medical condition of the user based on the management index value.
The information processing system according to (9) above.
(11)
The response information includes one or both of information regarding the user's current medical condition and information regarding the worsening of the user's medical condition.
The information processing system according to (10) above.
(12)
The information regarding the worsening of the user's condition includes information for suppressing the worsening of the user's condition.
The information processing system according to (11) above.
(13)
The response generation unit generates the response information based on the management index value and first response generation information.
The information processing system according to any one of (10) to (12).
(14)
The first response generation information includes information for generating the response information according to an intention of a user's utterance regarding the voice information.
The information processing system according to (13) above.
(15)
The system further includes a semantic analysis unit that analyzes the intention of the user utterance to generate the first response generation information.
The information processing system according to (14) above.
(16)
the response generation unit generates the response information based on the management index value, the first response generation information, and second response generation information different from the first response generation information.
The information processing system according to any one of (13) to (15).
(17)
The second response generation information includes one or both of environmental information about the user and schedule information of the user.
The information processing system according to (16) above.
(18)
a response generation database for storing the second response generation information;
The information processing system according to (16) or (17).
(19)
The computer
estimating a management index value related to the user's medical condition based on the user's voice information;
Information processing methods.
(20)
The computer
generating a control index estimation model for estimating a control index value related to a medical condition of the user based on the voice information of the user;
How to generate a learning model.
(21)
An information processing method using the information processing system according to any one of (1) to (18).
(22)
A method for generating a learning model, which generates a learning model for an information processing system described in any one of (1) to (18).

REFERENCE SIGNS LIST 1 Information processing system 1A Information processing system 1B Information processing system 1C Information processing system 1D Information processing system 10 Sound input unit 20 Biometric information detection unit 30 Sound output unit 40 User terminal 50 Information processing device 51 Management index estimation unit 52 Management index database 53 Voice recognition unit 54 Semantic analysis unit 55 Response generation database 56 Response generation unit 57 Response control unit 58 Model generation unit 58a Management index estimation model 58b Speech voice database 58c Management index measurement value database

Claims

a control index estimation unit that estimates a control index value related to a medical condition of the user based on voice information of the user;
Information processing system.
The management index value is a management index value related to a medical condition when the disease of the user is a respiratory system disease.
The information processing system according to claim 1 .
the control index estimation unit estimates the control index value based on the voice information and the biometric information of the user.
The information processing system according to claim 1 .
the control index estimation unit calculates an acoustic feature from the speech information, and estimates the control index value based on the calculated acoustic feature.
The information processing system according to claim 1 .
the management index estimation unit estimates the management index value using a management index estimation model which is a learning model;
5. The information processing system according to claim 4.
A model generation unit that generates the management index estimation model,
6. The information processing system according to claim 5.
The model generation unit generates the management index estimation model using voice information for each patient having a different symptom related to the user's disease and a management index measurement value for each of the voice information.
7. The information processing system according to claim 6.
the model generation unit calculates acoustic features from the speech information, estimates a control index value from the calculated acoustic features, performs learning so as to minimize an error between the estimated control index value and the control index measurement value corresponding to the estimated control index value, and generates the control index estimation model.
The information processing system according to claim 7.
A management index database for storing the management index values is further provided.
The information processing system according to claim 1 .
A response generating unit that generates response information regarding the medical condition of the user based on the management index value.
10. The information processing system according to claim 9.
The response information includes one or both of information regarding the user's current medical condition and information regarding the worsening of the user's medical condition.
The information processing system according to claim 10.
The information regarding the worsening of the user's condition includes information for suppressing the worsening of the user's condition.
The information processing system according to claim 11.
The response generation unit generates the response information based on the management index value and first response generation information.
The information processing system according to claim 10.
The first response generation information includes information for generating the response information according to an intention of a user's utterance regarding the voice information.
The information processing system according to claim 13.
The system further includes a semantic analysis unit that analyzes the intention of the user utterance to generate the first response generation information.
15. The information processing system according to claim 14.
the response generation unit generates the response information based on the management index value, the first response generation information, and second response generation information different from the first response generation information.
The information processing system according to claim 13.
The second response generation information includes one or both of environmental information about the user and schedule information of the user.
17. The information processing system according to claim 16.
a response generation database for storing the second response generation information;
17. The information processing system according to claim 16.
The computer
estimating a management index value related to the user's medical condition based on the user's voice information;
Information processing methods.
The computer
generating a control index estimation model for estimating a control index value related to a medical condition of the user based on the voice information of the user;
How to generate a learning model.