JP7655255B2 - Calculation result providing device - Google Patents

Description

This disclosure relates to a calculation result providing device.

A calculation result providing device is known that uses a model to provide calculation results (e.g., the amount of power consumption at the user's home) related to the user and the user's environment (e.g., the home address).

On the other hand, a trained model provision system is known that selects one or more trained models from multiple trained models stored in a database in advance in response to a usage request obtained from a user device and provides the selected trained model to the user device (see, for example, Patent Document 1).

International Publication No. 2018/142766

However, in the above-mentioned calculation result providing device, when the user's environment changes (for example, the user moves), the above-mentioned model is not necessarily suitable for the user's new environment. For this reason, it is difficult to accurately provide calculation results related to the user's new environment using the above-mentioned model.

Therefore, following the example of Patent Document 1, if a model suitable for the user's new environment is selected from multiple models stored in a database and the selected model is used, it may be possible to provide accurate calculation results related to the user and the new environment. However, a model suitable for the user's new environment is not necessarily stored in the database beforehand. For this reason, even if Patent Document 1 is taken into consideration, it is difficult to provide accurate calculation results related to the user's new environment.

According to the present disclosure, the following is provided:
[Configuration 1]
A calculation result providing device for providing a calculation result using a model suitable for a user and an environment of the user, comprising:
an output unit configured to provide, as a calculation result, an output of a corrected old model obtained by correcting an old model suitable for the user and the user's old environment when the user's environment is changed from an old environment to a new environment, until a new model suitable for the user and the new environment is completed, and to provide an output of the new model as a calculation result when the new model is completed;
An operation result providing device comprising:

When the user's environment changes, calculation results relating to the user and the new environment can be provided more accurately.

1 is a schematic diagram of an operation result providing device according to an embodiment of the present disclosure. FIG. 1 is a schematic diagram for explaining an embodiment according to the present disclosure. 10 is a flowchart illustrating a calculation result providing routine according to an embodiment of the present disclosure. FIG. 2 is a functional block diagram of a processor of the computation result providing device according to the embodiment of the present disclosure.

Figure 1 shows a schematic diagram of a computation result providing device 10 according to an embodiment of the present disclosure. Referring to Figure 1, the computation result providing device 10 comprises one or more processors 11, one or more memories 12, a storage device 13, and an input/output interface (IF) 14, which are communicatively connected to each other via a bidirectional bus. The computation result providing device 10 according to an embodiment of the present disclosure is also composed of a user terminal or a server.

The memory 12 in the embodiment of the present disclosure includes a volatile or non-volatile memory. Various programs and the like are stored in the memory 12, and these programs are executed by the processor 11. The storage device 13 in the embodiment of the present disclosure stores created models and the like.

The input/output IF 14 of the embodiment of the present disclosure is communicatively connected to a communication device 15, an input/output device 16, and one or more sensors 17. The communication device 15 of the embodiment of the present disclosure is communicatively connected to a communication network such as the Internet. The input/output device 16 of the embodiment of the present disclosure includes, for example, a keyboard, a mouse, a media reader/writer, a display, and the like. The sensor 17 of the embodiment of the present disclosure acquires one or more data related to the user's environment. In one example, the sensor 17 is installed in the user's environment. The sensor 17 detects, for example, one or more of data related to the weather in the user's environment (temperature, precipitation, humidity, etc.), traffic volume, power consumption, and the like.

The calculation result providing device 10 of an embodiment of the present disclosure is configured to provide calculation results related to a user and the user's environment using a model appropriate for the user and the user's environment. In an embodiment of the present disclosure, the user's environment includes a location (e.g., a city, town, region, etc.) associated with the user. Examples include the user's home address, the user's work address, etc.

Cities or towns in embodiments of the present disclosure include smart cities or connected cities that utilize big data, etc.

In an embodiment of the present disclosure, a model suited to a user and the user's environment outputs a calculation result relating to the user and the user's environment in response to a certain input. In addition, in an embodiment of the present disclosure, the calculation result relating to the user and the user's environment includes a quantity or numerical value obtained in a combination of the user and the user's environment. Examples include electricity consumption in the user's home, the user's time of leaving for work, and control parameters for autonomous driving suited to the user's preferences in the user's environment.

The models of the embodiments of the present disclosure are created by, for example, one or more of AI or artificial intelligence, neural networks, machine learning or deep learning, support vector machines (SVM), k-nearest neighbors (k-NN), k-means, association analysis, logistic regression, linear regression, Monte Carlo methods, random forests, boosting, least squares, maximum likelihood estimation, Bayesian estimation, Gaussian processes, and Markov models.

Now, when a user's environment changes from an old environment to a new environment (for example, moving from an old city to a new city, changing jobs, etc.), even if a model suitable for the user and the old environment (old model) is used, there is a risk that accurate calculation results associated with the user and the new environment will not be provided.

Therefore, in an embodiment of the present disclosure, when the user's environment is changed from an old environment to a new environment, the computation result providing device 10 creates a model (new model) suitable for the user and the new environment. Then, when the new model is completed, the computation result providing device 10 uses the new model to provide computation results associated with the user and the new environment.

The computation result providing device 10 of an embodiment of the present disclosure obtains data (teacher data) associated with the user and the new environment, and creates a new model using this teacher data. In one example, the teacher data is input via the input/output device 16 or obtained via the sensor 17. In another embodiment, the new model is created by a model creation device different from the computation result providing device 10.

However, it may take time from when the user's environment is changed until the new model is completed, especially before the training data required for model creation can be obtained. Until the new model is completed, accurate calculation results associated with the user and the new environment may not be provided.

Therefore, in an embodiment of the present disclosure, until the new model is completed, the calculation result providing device 10 provides calculation results associated with the user and the new environment using a model obtained by correcting the old model, i.e., the corrected old model. In one example, the corrected old model is formed from a model in which the calculation results obtained using the old model are corrected by a correction coefficient. In another example, the corrected old model is formed from a model in which the weighting coefficients of the old model are corrected by a correction coefficient. Examples of the correction coefficient include a correction coefficient that is determined in advance in association with at least one of the user and the new environment, a correction coefficient that is successively updated in association with at least one of the user and the new environment, etc.

That is, as shown in FIG. 2, until a change in the user's environment is input at time t1, the output of the old model is provided as the calculation result. When a change in the user's environment is input at time t1, the output of the corrected old model is provided as the calculation result. Meanwhile, creation of a new model begins at time t1. Then, when the new model is completed at time t2, the output of the new model is provided as the calculation result.

Figure 3 shows a calculation result providing routine in an embodiment according to the present disclosure. Referring to Figure 3, in step 100, it is determined whether a change in the user's environment has been received. When a change in the user's environment has been received, the process proceeds to step 101, where the current model is set to the old model. In the following step 102, the creation of a new model is started. In the following step 103, it is determined whether the new model has been completed. When the new model is not completed, the process proceeds to step 104, where the corrected old model is used to provide the calculation result. Then, the process returns to step 103. On the other hand, when the new model is completed, the process proceeds from step 103 to step 105, where the new model is used to provide the calculation result. When a change in the user's environment has not been received in step 100, the process proceeds to step 106, where the old model is used to provide the calculation result. In the embodiment according to the present disclosure, the current model (old model), the corrected old model, and the new model are stored in, for example, the storage device 13.

Fig. 4 is a functional block diagram of the processor 11 of the calculation result providing device 10 according to the embodiment of the present disclosure. Referring to Fig. 4, the processor 11 includes an output unit 11a. The output unit 11a is configured to provide, as a calculation result, an output of a corrected old model obtained by correcting an old model suitable for the user and the user's old environment when the user's environment is changed from an old environment to a new environment, until a new model suitable for the user and the new environment is completed, and to provide an output of the new model as a calculation result when the new model is completed.

10 Calculation result providing device 11 Processor 11a Output unit

Claims (1)

A calculation result providing device for providing a calculation result using a model suitable for a user and an environment of the user, comprising:
an output unit configured to start creating a new model suitable for the user and the new environment when the user's environment is changed from an old environment to a new environment, and to provide, as a calculation result, an output of a corrected old model obtained by correcting an old model suitable for the user and the user's old environment until the new model is completed, and to provide, as a calculation result, an output of the new model when the new model is completed;
An operation result providing device comprising:

Images