Information processing apparatus, information processing method, and program

Abstract

To improve the accuracy of prediction regarding a physical property value of a target object.SOLUTION: An information processing apparatus includes: acquisition means which acquires a first physical property value depending on one or more physical quantities; deriving means which combines the first physical property value with a second physical property value which depends on the physical quantity of at least one of the one or more physical quantities, to derive a feature quantity for which the dependency on the at least one of the physical quantities has been reduced compared to the first physical property value; and prediction means which executes prediction processing by referring to the feature quantity.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

[0002] Techniques for predicting the physical property values ​​of an object are known. For example, Patent Document 1 discloses a technique for determining an optimal solution for the physical property values ​​of plastics. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2023-62432 Summary of the Invention [Problem to be solved by the invention]

[0004] Generally, the physical properties of an object depend on one or more physical quantities. When predicting the physical properties of an object, the accuracy of the prediction may be significantly reduced due to the dependency of the physical properties on the one or more physical quantities. Furthermore, although these physical quantities can sometimes be measured directly, measurement is often difficult.

[0005] Therefore, there is a need to improve the accuracy of predicting the physical property values ​​of an object, regardless of whether or not one or more physical quantities on which the physical property values ​​depend can be measured.

[0006] The present disclosure has been made in view of the above-mentioned problems, and an exemplary purpose thereof is to provide a technique for improving prediction accuracy regarding physical property values ​​of an object. [Means for solving the problem]

[0007] An information processing device according to an exemplary aspect of the present disclosure includes an acquisition means for acquiring a first physical property value that depends on one or more physical quantities, a derivation means for combining the first physical property value with a second physical property value that depends on at least any of the one or more physical quantities to derive a feature value whose dependency on the at least any physical quantity is reduced compared to the first physical property value, and a prediction means for executing a prediction process using the feature value as a reference.

[0008] An information processing device according to an exemplary aspect of the present disclosure includes an acquisition means for acquiring a first physical property value that depends on one or more physical quantities, a derivation means for combining the first physical property value with a second physical property value that depends on at least any of the one or more physical quantities to derive a feature value whose dependency on the at least any physical quantity is reduced compared to the first physical property value, and a learning means for executing a learning process that references the feature value.

[0009] An information processing method according to an exemplary aspect of the present disclosure includes acquiring a first physical property value that depends on one or more physical quantities, combining the first physical property value with a second physical property value that depends on at least any of the one or more physical quantities to derive a feature value that has reduced dependency on the at least any physical quantity compared to the first physical property value, and performing a prediction process using the feature value as a reference.

[0010] An information processing method according to an exemplary aspect of the present disclosure includes: acquiring a first physical property value that depends on one or more physical quantities; combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities to derive a feature value having a reduced dependency on the at least one physical quantity compared to the first physical property value; and performing a learning process with reference to the feature value. Contains:

[0011] A program according to an exemplary aspect of the present disclosure is a program that causes a computer to function as an information processing device, and the program causes the computer to execute an acquisition process that acquires a first physical property value that depends on one or more physical quantities, a derivation process that combines the first physical property value with a second physical property value that depends on at least any of the one or more physical quantities to derive a feature value whose dependency on the at least any physical quantity is reduced compared to the first physical property value, and a prediction process that references the feature value.

[0012] A program according to an exemplary aspect of the present disclosure is a program that causes a computer to function as an information processing device, and the program causes the computer to execute an acquisition process that acquires a first physical property value that depends on one or more physical quantities, a derivation process that combines the first physical property value with a second physical property value that depends on at least any of the one or more physical quantities to derive a feature value whose dependency on the at least any physical quantity is reduced compared to the first physical property value, and a learning process that references the feature value. [Effects of the Invention]

[0013] According to an exemplary aspect of the present disclosure, an exemplary effect is achieved in that a technique for improving prediction accuracy regarding physical property values ​​of an object can be provided. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a block diagram illustrating a configuration of an information processing device according to the present disclosure. [Figure 2] FIG. 1 is a flow diagram showing the flow of an information processing method according to the present disclosure. [Figure 3] 1 is a block diagram illustrating a configuration of an information processing device according to the present disclosure. [Figure 4] FIG. 1 is a flow diagram showing the flow of an information processing method according to the present disclosure. [Figure 5] 1 is a block diagram illustrating a configuration of an information processing device according to the present disclosure. [Figure 6]FIG. 1 is a diagram for explaining information processing according to the present disclosure. [Figure 7] FIG. 1 is a diagram for explaining information processing according to the present disclosure. [Figure 8] FIG. 10 is a diagram illustrating an example of correspondence information according to the present disclosure. [Figure 9] FIG. 1 is a diagram for explaining information processing according to the present disclosure. [Figure 10] FIG. 1 is a diagram for explaining information processing according to the present disclosure. [Figure 11] FIG. 10 is a diagram illustrating a display example by an information processing device according to the present disclosure. [Figure 12] 1 is a block diagram illustrating a configuration of an information processing device according to the present disclosure. [Figure 13] FIG. 1 is a block diagram illustrating a configuration of a computer that functions as an information processing device according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0015] The following are examples of embodiments of the present invention. However, the present invention is not limited to the exemplary embodiments shown below, and various modifications are possible within the scope of the claims. For example, embodiments obtained by appropriately combining the technologies (part or all of the products or methods) employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, embodiments obtained by appropriately omitting some of the technologies employed in the exemplary embodiments shown below may also be included in the scope of the present invention. Furthermore, the effects mentioned in the exemplary embodiments shown below are examples of effects expected in the exemplary embodiments, and do not define the scope of the present invention. In other words, embodiments that do not exhibit the effects mentioned in the exemplary embodiments shown below may also be included in the scope of the present invention.

[0016] First Exemplary Embodiment A first exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. This exemplary embodiment is the basic form of each exemplary embodiment described later. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise. Furthermore, each technology shown in the drawings referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise.

[0017] (Configuration of information processing device 1) The configuration of an information processing device 1 according to this exemplary embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the configuration of the information processing device 1. As shown in Fig. 1, the information processing device 1 includes an acquisition unit 11, a derivation unit 12, and a prediction unit 13.

[0018] (Acquisition part 11) The acquisition unit 11 acquires a first physical property value, which is a physical property value related to an object and depends on one or more physical quantities. Here, the first physical property value acquired by the acquisition unit 11 refers to a first physical property value referenced in a prediction phase, for example, but this term does not limit the present exemplary embodiment.

[0019] Furthermore, the term "physical property value" refers to a property of an object that is expressed numerically using a predetermined scale. For example, the physical property value may be: (1) Physical properties related to the mechanical properties of the object (2) Physical properties related to the thermal properties of the object (3) Physical property values ​​related to the optical properties of the object (4) Physical property values ​​related to the electrical properties of the object (5) Physical properties related to the chemical properties of the object (6) Physical properties of the object However, these example classifications are not intended to limit the present instructional embodiments.

[0020] (1) Examples of physical property values ​​related to the mechanical properties of an object include: Tensile strength, flexural strength, tensile modulus, flexural modulus, impact fracture energy The following can be mentioned:

[0021] (2) Examples of physical property values ​​related to the thermal properties of an object include: Melting point, thermal expansion coefficient, thermal conductivity, specific heat, viscosity coefficient, melt flow rate, glass transition temperature The following can be mentioned:

[0022] (3) Examples of physical property values ​​related to the optical properties of the object include: Color, color development, refractive index, reflectance, transmittance The following can be mentioned:

[0023] (4) Examples of physical property values ​​related to the electrical properties of the object include: Dielectric constant, volume resistivity The following can be mentioned:

[0024] (5) Examples of physical property values ​​related to the chemical properties of the object include: Acid, alkali and oil resistance The following can be mentioned:

[0025] (6) Examples of physical property values ​​related to the physical properties of the object include: ·Density, specific gravity The following can be mentioned:

[0026] The above-mentioned physical property values ​​generally depend on one or more physical quantities. More specifically, the above-mentioned physical property values ​​of an object generally depend on the physical quantities of the constituent elements that make up the object, or on the physical quantities of the object itself. Here, the "constituent elements" may include atoms, molecules, monomers, polymers, etc., but these examples do not limit the present exemplary embodiment. Furthermore, the above-mentioned "physical quantity" refers to a quantity that has dimensions, at least in a physical theoretical system, and a quantity for which a unit quantity can be defined. Specific examples of physical quantities do not limit the present exemplary embodiment, but as an example, Length, mass, time, current, voltage, charge, temperature, amount of substance, molecular weight, density, specific gravity, luminous intensity, speed, acceleration, force, pressure, energy etc. may be included.

[0027] Note that, due to limitations of manufacturing equipment and measurement equipment, it may not be easy to directly measure the physical quantities described above. For example, it may not be as easy to acquire the physical quantities as compared to the first physical property value acquired by the acquisition unit 11 and the second physical property value described later. Even in such cases, according to the present exemplary embodiment, it is possible to improve the prediction accuracy in the prediction process, as described later.

[0028] Furthermore, the specific example of the "object" does not limit the present exemplary embodiment, but an example thereof is a substance containing resin. For example, resin materials such as virgin plastic, waste plastic, and recycled plastic can be cited. The first physical property value acquired by the acquisition unit 11 is stored in a storage unit (not shown), for example.

[0029] (Derivation part 12) The derivation unit 12 combines the first physical property value acquired by the acquisition unit 11, which is dependent on one or more physical quantities, with a second physical property value dependent on at least any of the one or more physical quantities, to derive (generate) a feature quantity whose dependency on at least any physical quantity is reduced compared to the first physical property value. Here, the second physical property value or a candidate for the second physical property value can be acquired by the acquisition unit 11 described above.

[0030] As an example, when a first physical property value dependent on a physical quantity x is expressed as p1(x) and a second physical property value dependent on the physical quantity x is expressed as p2(x), the derivation unit 12 derives (generates) a feature quantity F by combining p1(x) and p2(x) to obtain a feature quantity f. Here, the feature quantity F can be expressed as, for example, F=f(p1(x), p2(x)) Here, the function f(A, B) is, for example, a function defined by addition, subtraction, multiplication, division, exponential operation, logarithmic operation, power operation, etc., on at least two arguments A and B. The feature quantity F derived by the derivation unit 12 can be expressed by using the physical quantity x as follows: F=f(p1(x), p2(x))=g(x) Here, the function g(x) is a function whose dependency on the physical quantity x is reduced more than that of at least one of the first physical property value p1(x) and the second physical property value p2(x). As an example, the function g(x) is a function whose dependency on the physical quantity x is reduced more than that of the first physical property value p1(x). Hereinafter, when referring to at least one of the first physical property value p1(x) and the second physical property value p2(x), it may be written as the physical property value p(x).

[0031] The derivation unit 12 may, for example, The dependence of the first physical property p1(x) on the physical quantity x; The dependence of a plurality of physical property values ​​that are candidates for the second physical property value p2(x) on the physical quantity x; and selects the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information including the above.

[0032] In the above description, "g(x) having a reduced dependency on the physical quantity x compared to p(x)" refers to, for example, g(x) such that the degree of change in g(x) with a change in x is smaller than the degree of change in p(x) with a change in x. For example, when p(x) is expressed as a power function of x, p(x)=xn → g(x)=x m (0≦m <n) p(x)=x n → g(x)=x m (n <m≦0) If g(x) is expressed as a power function whose exponent is closer to 0 than p(x), or as a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0033] Another example is when p(x) is expressed as an exponential function of x: p(x)=exp(x) → g(x)=x m If g(x) is expressed as a power function or a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0034] As another example, if p(x) is expressed as a logarithmic function of x, then ·p(x)=log(x) → g(x)=constant If g(x) is expressed as a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0035] As another example, when p is expressed by a cross term between physical quantity x and physical quantity y, p(x,y)=xy → g(x,y)= x a ·y b (0 <a,0<b, a+b<2) If g(x, y) is expressed so that the sum of the exponents for x and y is smaller than the sum of the exponents in p(x, y), then it can be said that g(x, y) has a lower dependency on x and y than p(x, y). The feature amount derived (generated) by the derivation unit 12 is stored in a storage unit (not shown), for example.

[0036] In the above description, the matter of "reduced dependency on the physical quantity x" has been explained using mathematical expressions, but this does not limit the present exemplary embodiment. For example, the information processing device 1, or a device separate from the information processing device 1, the dependence of the first physical property on the physical quantity x, and Dependence of the second physical property on the physical quantity x The present invention may be configured to include a measurement unit that can directly or observationally measure the physical quantity x, and to verify that the "dependence on the physical quantity x has been reduced" through measurements by the measurement unit.

[0037] (Prediction Section 13) The prediction unit 13 performs a prediction process with reference to the feature quantities derived by the derivation unit 12. As an example, the prediction unit 13 performs a prediction process regarding the physical property values ​​of the object using a linear function or a nonlinear function with one or more feature quantities including the feature quantities derived by the derivation unit 12 as arguments. This process may include a process of predicting the physical property values ​​of the prediction target by a linear combination of one or more feature quantities including the feature quantities derived by the derivation unit 12.

[0038] As an example, if the plurality of feature quantities derived by the derivation unit 12 are expressed as F1, F2, . . . , the prediction unit 13 calculates the physical property value PP of the object to be predicted as follows: PP=h(F1,F2,...) Here, the function h is a function that takes the plurality of feature quantities F1, F2, ... as arguments. As one or more parameters of the function h, for example, those learned in advance by a learning process that refers to the feature quantities derived by the derivation unit 12 in the learning phase may be used. However, this example does not limit the present exemplary embodiment.

[0039] The function h may be a linear regression function or a nonlinear regression function. The function may represent a neural network having multiple layers, or may represent a support vector machine, a decision tree, Bayesian optimization, or the like. The prediction unit 13 may be configured to perform the prediction process by physical or chemical simulation with reference to one or more feature quantities including the feature quantity derived by the derivation unit 12.

[0040] As described above, F1, F2, ... derived by the derivation unit 12 have a reduced dependency on the feature value x compared to the first physical property value and the second physical property value acquired by the acquisition unit 11. For this reason, predicting the physical property value PP of the prediction target using the multiple feature values ​​F1, F2, ... derived by the derivation unit 12 improves prediction accuracy compared to predicting the physical property value PP of the prediction target using the first physical property value and the second physical property value directly.

[0041] The prediction result by the prediction unit 13 is - Presented to the user via a display unit (not shown) The information may be provided to an external device via a communication unit (not shown).

[0042] (Effects of information processing device 1) As described above, in the information processing device 1, an acquisition unit 11 that acquires a first physical property value that depends on one or more physical quantities; a derivation unit 12 that combines the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; a prediction unit 13 that executes a prediction process by referring to the feature amount; The configuration is adopted.

[0043] In other words, the information processing device 1 an acquisition unit 11 that acquires a first physical property value that depends on one or more physical quantities and a second physical property value that depends on at least one of the one or more physical quantities; a derivation unit 12 that derives a feature quantity having reduced dependency on at least one of the physical quantities by combining the first physical property value and the second physical property value; a prediction unit 13 that executes a prediction process by referring to the feature amount; It may also be expressed as a configuration comprising:

[0044] In this way, in the information processing device 1, by combining the first physical property value and the second physical property value, the prediction process is performed by referring to a feature whose dependency on the physical quantity is reduced, thereby improving the prediction accuracy compared to when the prediction process is performed by directly referring to the first physical property value and the second physical property value. (Flow of information processing method S1) Next, the flow of the information processing method S1 according to this exemplary embodiment will be described with reference to Fig. 2. Fig. 2 is a flow diagram showing the flow of the information processing method S1. As shown in Fig. 2, the information processing method S1 includes an acquisition process (acquisition step) S11, a derivation process (derivation step) S12, and a prediction process (prediction step) S13.

[0045] (Step S11) In step S11, the acquisition unit 11 acquires a first physical property value, which is a physical property value related to the object and depends on one or more physical quantities. A more specific description of the acquisition unit 11 has been given above, so a description thereof will be omitted here.

[0046] (Step S12) In step S12, the derivation unit 12 combines the first physical property value acquired by the acquisition unit 11, which is dependent on one or more physical quantities, with a second physical property value dependent on at least any of the one or more physical quantities, to derive (generate) a feature quantity whose dependency on the at least any physical quantity is reduced compared to the first physical property value. A more specific description of the derivation unit 12 has been given above, and therefore will not be repeated here.

[0047] (Step S13) In step S13, the prediction unit 13 executes a prediction process by referring to the feature amount derived by the derivation unit 12. A more specific description of the prediction unit 13 has been given above, and therefore will not be repeated here.

[0048] (Effect of information processing method S1) As described above, in the information processing method S1, Obtaining a first physical property value that depends on one or more physical quantities; deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value by combining the first physical property value with a second physical property value that depends on at least one physical quantity of the one or more physical quantities; - Execute prediction processing by referring to the feature values According to the above configuration, the same effects as those of the information processing device 1 are achieved.

[0049] (Configuration of information processing device 2) Next, the configuration of the information processing device 2 according to this exemplary embodiment will be described with reference to Fig. 3. Fig. 3 is a block diagram showing the configuration of the information processing device 2. As shown in Fig. 3, the information processing device 2 includes an acquisition unit 21, a derivation unit 22, and a learning unit 23.

[0050] (Acquisition part 21) The acquisition unit 21 acquires a first physical property value, which is a physical property value related to the object and depends on one or more physical quantities. Here, the first physical property value acquired by the acquisition unit 21 refers to the first physical property value referenced in the learning phase, for example, but this terminology does not limit this exemplary embodiment. Furthermore, the "physical property value" refers to a property of the object that is numerically expressed using a predetermined scale. The terms "physical property value," "physical quantity," and "object" are the same as those explained in the information processing device 1, so redundant explanations will be omitted.

[0051] The acquisition unit 21 may be configured to further acquire a ground truth label of the physical property value of the prediction target, which is referred to by a learning unit described later.

[0052] (Derivation part 22) The derivation unit 22 combines the first physical property value acquired by the acquisition unit 21, which is dependent on one or more physical quantities, with a second physical property value dependent on at least any of the one or more physical quantities, to derive (generate) a feature quantity whose dependency on the at least any physical quantity is reduced compared to the first physical property value. Here, the second physical property value or a candidate for the second physical property value can be acquired by the above-mentioned acquisition unit 21. The processing by the derivation unit 22 is similar to that of the derivation unit 12 included in the information processing device 1, and therefore a description thereof will be omitted here.

[0053] (Study Section 23) The learning unit 23 executes a learning process by referring to the feature amount derived by the derivation unit 22. As an example, the learning unit 23 Execute a prediction process for the physical property value of the object using a linear function or a nonlinear function with one or more feature quantities including the feature quantity derived by the derivation unit 22 as arguments; The learning process is performed by updating one or more parameters of the linear or nonlinear function so that the result of the prediction process approaches the correct label for the physical property value. Here, the prediction process may include a process of predicting the physical property value of the prediction target by a linear combination of one or more feature quantities including the feature quantity derived by the derivation unit 22. However, this does not limit the present exemplary embodiment.

[0054] As an example, if the plurality of feature quantities derived by the derivation unit 22 are expressed as F1, F2, . . . , the learning unit 23 calculates the physical property value PP of the object as follows: PP=h(F1,F2,...) Here, the function h is a function that takes the plurality of feature amounts F1, F2, ... as arguments. Then, the learning unit 23 updates at least one of one or more parameters of the function h so that the difference between the predicted physical property value PP and the correct label GL of the physical property value becomes smaller. The function h that has undergone the learning process by the learning unit 23 may also be referred to as a learned model.

[0055] The function h may be a linear regression function or a nonlinear regression function. The function may represent a neural network having multiple layers, or may represent a support vector machine, a decision tree, Bayesian optimization, or the like. The learning unit 23 may be configured to execute the prediction process by physical or chemical simulation with reference to one or more feature quantities including the feature quantity derived by the derivation unit 22.

[0056] The parameters of the above function learned by the learning unit 23 are -Stored in a memory unit (not shown) The prediction unit 13 included in the information processing device 1 described above may refer to the information.

[0057] (Effects of information processing device 2) As described above, in the information processing device 2, an acquisition unit 21 that acquires a first physical property value that depends on one or more physical quantities; a derivation unit (22) that combines the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; A learning unit 23 that executes a learning process by referring to the feature amount; The configuration is adopted.

[0058] In other words, the information processing device 2 an acquisition unit 21 that acquires a first physical property value that depends on one or more physical quantities and a second physical property value that depends on at least one of the one or more physical quantities; a derivation unit 22 that derives a feature quantity having reduced dependency on at least one of the physical quantities by combining the first physical property value and the second physical property value; A learning unit 23 that executes a learning process by referring to the feature amount; It may also be expressed as a configuration comprising:

[0059] In this way, in the information processing device 2, by combining the first physical property value and the second physical property value, the learning process is performed by referring to feature amounts whose dependency on the physical quantity is reduced, so that a trained model with higher prediction accuracy can be generated compared to when the learning process is performed by directly referring to the first physical property value and the second physical property value. Therefore, prediction accuracy can be improved compared to when the learning process is performed by directly referring to the first physical property value and the second physical property value.

[0060] (Flow of information processing method S2) Next, the flow of the information processing method S2 according to this exemplary embodiment will be described with reference to Fig. 4. Fig. 4 is a flow diagram showing the flow of the information processing method S2. As shown in Fig. 4, the information processing method S2 includes an acquisition process (acquisition step) S21, a derivation process (derivation step) S22, and a learning process (learning step) S23.

[0061] (Step S21) In step S21, the acquisition unit 21 acquires a first physical property value, which is a physical property value related to the object and depends on one or more physical quantities. A more specific description of the acquisition unit 21 has been given above, so a description thereof will be omitted here.

[0062] (Step S22) In step S22, the derivation unit 22 combines the first physical property value acquired by the acquisition unit 21, which is dependent on one or more physical quantities, with a second physical property value dependent on at least any of the one or more physical quantities, to derive (generate) a feature quantity whose dependency on the at least any physical quantity is reduced compared to the first physical property value. A more specific description of the derivation unit 22 has been given above, and therefore will not be repeated here.

[0063] (Step S23) In step S23, the learning unit 23 executes a learning process by referring to the feature amount derived by the derivation unit 22. A more specific description of the learning unit 23 has been given above, and therefore will not be repeated here.

[0064] (Effect of information processing method S2) As described above, in the information processing method S2, Obtaining a first physical property value that depends on one or more physical quantities; deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value by combining the first physical property value with a second physical property value that depends on at least one physical quantity of the one or more physical quantities; Execute the learning process by referring to the feature values The above configuration provides the same effects as the information processing device 2.

[0065] Second Exemplary Embodiment A second exemplary embodiment, which is one example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same functions as those described in the above exemplary embodiment will be assigned the same reference numerals, and their description will be omitted as appropriate. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs. Furthermore, each technology shown in each drawing referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical hindrance occurs.

[0066] (Configuration of information processing system 1A) The configuration of an information processing system 1A according to this exemplary embodiment will be described with reference to Fig. 5. Fig. 5 is a block diagram showing the configuration of the information processing system 1A. As shown in Fig. 5, the information processing system 1A includes an information processing device 100A and a measuring device 50 connected to the information processing device 100A via a network N. Here, the specific configuration of the network N does not limit this exemplary embodiment, but as an example, a wireless LAN (Local Area Network), a wired LAN, a WAN (Wide Area Network), a public line network, a mobile data communication network, or a combination of these networks can be used.

[0067] (Configuration of information processing device 100A) The configuration of an information processing device 100A according to this exemplary embodiment will be described with reference to Fig. 5. As shown in Fig. 5, the information processing device 100A includes a control unit 10, a storage unit 20, a communication unit 30, and an input / output unit 40.

[0068] (Communication unit 30) The communication unit 30 communicates with devices external to the information processing device 100A. As an example, the communication unit 30 communicates with the measurement device 50. The communication unit 30 transmits data supplied from the control unit 10 to the measurement device 50, and supplies data received from the measurement device 50 to the control unit 10.

[0069] (Input / output section 40) The input / output unit 40 is configured to include at least one of input / output devices such as a keyboard, a mouse, a display, a printer, and a touch panel. Alternatively, the input / output unit 40 may be configured to have input / output devices such as a keyboard, a mouse, a display, a printer, and a touch panel connected to it. In this configuration, the input / output unit 40 accepts various types of information input to the information processing device 100A from the connected input devices. Furthermore, the input / output unit 40 outputs various types of information to connected output devices under the control of the control unit 10. An example of the input / output unit 40 is an interface such as a USB (Universal Serial Bus).

[0070] (Storage unit 20) The storage unit 20 stores various data referenced by the control unit 10 and various data generated by the control unit 10. As an example, the storage unit 20 stores: ·Physical property value group PPVG Feature Group FVG Predictive Model PM Prediction result PRED ·Compatibility information CI is stored.

[0071] The physical property value group PPVG includes a plurality of physical property values, including a first physical property value PPV1 and a second physical property value PPV2. The physical property value group PPVG may also include one or more candidates for the first physical property value PPV1 and one or more candidates for the second physical property value PPV2. Each physical property value included in the physical property value group PPVG is, for example, acquired by the acquisition unit 11 (21).

[0072] The feature group FVG includes one or more feature values ​​derived by the derivation unit 12 (22).

[0073] The prediction model PM is a trained model that is trained by the training unit 23 and used in the prediction process by the prediction unit 13. The prediction result PRED is information indicating the prediction result by the prediction unit 13.

[0074] The correspondence information CI is information in which each of the plurality of physical property values ​​included in the physical property value group PPVG is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities. Specific examples of the correspondence information CI will be described later.

[0075] (Control unit 10) 5, the control unit 10 includes an acquisition unit 11, a derivation unit 12, a prediction unit 13, an output unit 14, and a learning unit 23. Here, the acquisition unit 11 also has the same functions as the acquisition unit 11 included in the information processing device 1 described in exemplary embodiment 1 and the acquisition unit 21 included in the information processing device 2 described in exemplary embodiment 1, and therefore the acquisition unit 11 may also be referred to as the acquisition unit 11 (21). Furthermore, the derivation unit 12 also has the same functions as the derivation unit 12 included in the information processing device 1 described in exemplary embodiment 1 and the derivation unit 22 included in the information processing device 2 described in exemplary embodiment 1, and therefore the derivation unit 12 may also be referred to as the derivation unit 12 (22).

[0076] (Acquisition part 11) The acquisition unit 11 acquires a first physical property value, which is a physical property value related to an object and depends on one or more physical quantities. Here, the first physical property value acquired by the acquisition unit 11 refers to a first physical property value referenced in a prediction phase, for example, but this term does not limit the present exemplary embodiment.

[0077] Furthermore, the term "physical property value" refers to a property of an object that is expressed numerically using a predetermined scale. For example, the physical property value may be: (1) Physical properties related to the mechanical properties of the object (2) Physical properties related to the thermal properties of the object (3) Physical property values ​​related to the optical properties of the object (4) Physical property values ​​related to the electrical properties of the object (5) Physical properties related to the chemical properties of the object (6) Physical properties of the object However, these example classifications are not intended to limit the present instructional embodiments.

[0078] (1) Examples of physical property values ​​related to the mechanical properties of an object include: Tensile strength, flexural strength, tensile modulus, flexural modulus The following can be mentioned:

[0079] (2) Examples of physical property values ​​related to the thermal properties of an object include: Melting point, thermal expansion coefficient, thermal conductivity, specific heat, viscosity coefficient, Melt Flow Rate The following can be mentioned:

[0080] (3) Examples of physical property values ​​related to the optical properties of the object include: Color, color development, refractive index, reflectance, transmittance The following can be mentioned:

[0081] (4) Examples of physical property values ​​related to the electrical properties of the object include: Dielectric constant, volume resistivity The following can be mentioned:

[0082] (5) Examples of physical property values ​​related to the chemical properties of the object include: Acid, alkali and oil resistance The following can be mentioned:

[0083] (6) Examples of physical property values ​​related to the physical properties of the object include: ·Density, specific gravity The following can be mentioned:

[0084] The above-mentioned physical property values ​​generally depend on one or more physical quantities. More specifically, the above-mentioned physical property values ​​of an object generally depend on the physical quantities of the constituent elements that make up the object, or on the physical quantities of the object itself. Here, the "constituent elements" may include atoms, molecules, monomers, polymers, etc., but these examples do not limit the present exemplary embodiment. Furthermore, the above-mentioned "physical quantity" refers to a quantity that has dimensions, at least in a physical theoretical system, and a quantity for which a unit quantity can be defined. Specific examples of physical quantities do not limit the present exemplary embodiment, but as an example, Length, mass, time, current, voltage, charge, temperature, amount of substance, molecular weight, density, specific gravity, luminous intensity, speed, acceleration, force, pressure, energy etc. may be included.

[0085] Note that, due to limitations of manufacturing equipment and measurement equipment, it may not be easy to directly measure the above-mentioned physical quantities. For example, it may not be easy to acquire the above-mentioned physical quantities compared to the first physical quantity and the second physical quantity acquired by the acquisition unit 11. Even in such cases, according to the present exemplary embodiment, it is possible to improve the prediction accuracy in the prediction process, as will be described later.

[0086] Furthermore, the specific example of the "object" does not limit the present exemplary embodiment, but an example thereof is a substance containing resin. For example, a resin material such as virgin plastic or waste plastic can be mentioned. The first physical property value acquired by the acquisition unit 11 is stored in the storage unit 20, for example.

[0087] (Derivation part 12) The derivation unit 12 combines the first physical property value acquired by the acquisition unit 11, which is dependent on one or more physical quantities, with a second physical property value dependent on at least any of the one or more physical quantities, to derive (generate) a feature quantity whose dependency on at least any physical quantity is reduced compared to the first physical property value. Here, the second physical property value or a candidate for the second physical property value can be acquired by the acquisition unit 11 described above.

[0088] In addition, as an example, the derivation unit 12 derives a feature quantity having reduced dependency on at least one of the physical quantities by combining the first physical property value and the second physical property value acquired by the acquisition unit 11.

[0089] As an example, when a first physical property value dependent on a physical quantity x is expressed as p1(x) and a second physical property value dependent on the physical quantity x is expressed as p2(x), the derivation unit 12 derives (generates) a feature quantity F by combining p1(x) and p2(x) to obtain a feature quantity f. Here, the feature quantity F can be expressed as, for example, F=f(p1(x), p2(x)) Here, the function f(A, B) is, for example, a function defined by addition, subtraction, multiplication, division, exponential operation, logarithmic operation, power operation, etc., on at least two arguments A and B. The feature quantity F derived by the derivation unit 12 can be expressed by using the physical quantity x as follows: F=f(p1(x), p2(x))=g(x) Here, the function g(x) is a function whose dependency on the physical quantity x is reduced compared to at least one of the first physical property value p1(x) and the second physical property value p2(x), more specifically, the first physical property value p1(x). Hereinafter, when referring to at least one of the first physical property value p1(x) and the second physical property value p2(x), it may be written as the physical property value p(x).

[0090] The derivation unit 12 may, for example, The dependence of the first physical property p1(x) on the physical quantity x; The dependence of a plurality of physical property values ​​that are candidates for the second physical property value p2(x) on the physical quantity x; and selects the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information including the above.

[0091] In the above description, "g(x) having a reduced dependency on the physical quantity x compared to p(x)" refers to, for example, g(x) such that the degree of change in g(x) with a change in x is smaller than the degree of change in p(x) with a change in x. For example, when p(x) is expressed as a power function of x, p(x)=x n → g(x)=x m (0≦m <n) p(x)=x n → g(x)=x m (n <m≦0) If g(x) is expressed as a power function whose exponent is closer to 0 than p(x), or as a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0092] Another example is when p(x) is expressed as an exponential function of x: p(x)=exp(x) → g(x)=x m If g(x) is expressed as a power function or a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0093] As another example, if p(x) is expressed as a logarithmic function of x, then ·p(x)=log(x) → g(x)=constant If g(x) is expressed as a constant function, then it can be said that g(x) has less dependency on x than p(x).

[0094] As another example, when p is expressed by a cross term between physical quantity x and physical quantity y, p(x,y)=xy → g(x,y)= x a ·y b (0 <a,0<b, a+b<2) If g(x, y) is expressed such that the sum of the exponents for x and y is smaller than the sum of the exponents in p(x, y), then it can be said that g(x, y) has a reduced dependency on x and y compared to p(x, y). The feature derived (generated) by the derivation unit 12 is stored in the storage unit 20, for example.

[0095] For example, the first physical property value p1(x) may include a property value related to the fluidity of the object (Melt Flow Rate), and the second physical property value p2(x) may include the specific gravity of the object. In this case, for example, the derivation unit 12 may derive the feature amount by multiplying the first physical property value p1(x) and the second physical property value p2(x).

[0096] (Prediction Section 13) The prediction unit 13 performs a prediction process with reference to the feature quantities derived by the derivation unit 12. As an example, the prediction unit 13 performs a prediction process regarding the physical property values ​​of the object using a linear function or a nonlinear function with one or more feature quantities including the feature quantities derived by the derivation unit 12 as arguments. This process may include a process of predicting the physical property values ​​of the prediction target by a linear combination of one or more feature quantities including the feature quantities derived by the derivation unit 12.

[0097] As an example, if the plurality of feature quantities derived by the derivation unit 12 are expressed as F1, F2, . . . , the prediction unit 13 calculates the physical property value PP of the object to be predicted as follows: PP=h(F1,F2,...) Here, the function h is a function that takes the plurality of feature quantities F1, F2, ... as arguments. As one or more parameters of the function h, for example, those learned in advance by a learning process that refers to the feature quantities derived by the derivation unit 12 in the learning phase may be used. However, this example does not limit the present exemplary embodiment.

[0098] The function h may be a linear regression function or a nonlinear regression function. The function may represent a neural network having multiple layers, or may represent a support vector machine, a decision tree, Bayesian optimization, or the like. The prediction unit 13 may be configured to perform the prediction process by physical or chemical simulation with reference to one or more feature quantities including the feature quantity derived by the derivation unit 12.

[0099] As described above, F1, F2, ... derived by the derivation unit 12 have a reduced dependency on the feature value x compared to the first physical property value and the second physical property value acquired by the acquisition unit 11. For this reason, predicting the physical property value PP of the prediction target using the multiple feature values ​​F1, F2, ... derived by the derivation unit 12 improves prediction accuracy compared to predicting the physical property value PP of the prediction target using the first physical property value and the second physical property value directly.

[0100] The prediction result by the prediction unit 13 is Presented to the user via the input / output unit 40, The information may be provided to an external device via the communication unit 30.

[0101] (Output unit 14) The output unit 14 outputs the processing result by the information processing device 100A. As an example, the output unit 14 presents the prediction result PRED by the prediction unit 13 to the user via the input / output unit 40. Furthermore, as an example, the output unit 14 presents candidate combinations of physical property values ​​in the learning process by the learning unit 23 to the user via the input / output unit 40, as shown in FIG.

[0102] (Study Section 23) The learning unit 23 executes a learning process by referring to the feature amount derived by the derivation unit 22. As an example, the learning unit 23 Execute a prediction process for the physical property value of the object using a linear function or a nonlinear function with one or more feature quantities including the feature quantity derived by the derivation unit 22 as arguments; The learning process is performed by updating one or more parameters of the linear or nonlinear function so that the result of the prediction process approaches the correct label for the physical property value. Here, the prediction process may include a process of predicting the physical property value of the prediction target by a linear combination of one or more feature quantities including the feature quantity derived by the derivation unit 22.

[0103] As an example, if the plurality of feature quantities derived by the derivation unit 22 are expressed as F1, F2, . . . , the learning unit 23 calculates the physical property value PP of the object as follows: PP=h(F1,F2,...) Here, the function h is a function that takes the plurality of feature amounts F1, F2, ... as arguments. Then, the learning unit 23 updates at least one of one or more parameters of the function h so that the difference between the predicted physical property value PP and the correct label GL of the physical property value becomes smaller. The function h that has undergone the learning process by the learning unit 23 may also be referred to as a learned model.

[0104] The function h may be a linear regression function or a nonlinear regression function. The function may represent a neural network having multiple layers, or may represent a support vector machine, a decision tree, Bayesian optimization, or the like. The learning unit 23 may be configured to execute the prediction process by physical or chemical simulation with reference to one or more feature quantities including the feature quantity derived by the derivation unit 22.

[0105] The parameters of the above function learned by the learning unit 23 are Stored in the memory unit 20, The prediction unit 13 may refer to the data.

[0106] (Measuring device 50) As shown in FIG. 5 , the measuring device 50 includes a control unit 51, a storage unit 52, a communication unit 53, and a measurement unit 54. The communication unit 53 communicates with devices external to the measuring device 50. As an example, the communication unit 53 communicates with an information processing device 100A included in the information processing system 1A. The communication unit 53 transmits data supplied from the control unit 51 to the information processing device 100A, and supplies data received from the information processing device 100A to the control unit 51. Note that the data received by the communication unit 53 from the information processing device 100A may include information regarding a measurement instruction for a physical property value used by the information processing device 100A. Furthermore, the data provided by the communication unit 53 to the information processing device 100A may include a measurement result by the measurement unit 54, which will be described later.

[0107] As an example, the measurement unit 54 measures one or more physical property values ​​for each of one or more objects via a sensor or the like. The measurement may include an analysis process that references output data from the sensor or the like. Details of the "object" and "physical property value" have been described above, so a detailed description thereof will be omitted here.

[0108] The control unit 51 controls each unit of the measurement device 50 in an integrated manner. As an example, the control unit 51 instructs the measurement unit 54 to measure the physical property values ​​of the object. As another example, the control unit 51 acquires the measurement results by the measurement unit 54. Here, as an example, the control unit 51 stores the measurement results by the measurement unit 54 in the memory unit 52. As another example, the control unit 51 provides the measurement results by the measurement unit 54 to the information processing device 100A via the communication unit 53.

[0109] The storage unit 52 stores the measurement results obtained by the measurement unit 54. That is, the storage unit 52 stores, as an example, the measurement results of the physical property values ​​of the object.

[0110] In this exemplary embodiment, the measuring device 50 is illustrated as a device separate from the information processing device 100A, but this does not limit the exemplary embodiment. The function of the control unit 51 provided in the measuring device 50 may be provided in the control unit of the information processing device 100A. Similarly, the measurement results of the physical property values ​​of the object stored in the memory unit 52 provided in the measuring device 50 may be stored in the memory unit of the information processing device 100A, and the measurement results may be referenced by the information processing device 100A itself.

[0111] (Prediction processing example 1) 6 shows prediction processing example 1 by information processing device 100A. This example shows, as an example, processing for predicting a physical property value of a prediction target (object physical property value TP) for an object using a plurality of physical property values ​​(physical property values ​​P11, P12, P21, P22) for the object. As shown in FIG. 6, prediction processing example 1 includes processing for acquiring physical property values ​​(S11-1, S11-2), processing for generating feature amounts (S12-1, S12-2), and processing for predicting physical property values ​​(S13).

[0112] (Process S11-1) In process S11-1, the acquisition unit 11 (21) acquires a first physical property value P11 and a second physical property value P12 of the object. Here, the first physical property value P11 and the second physical property value P12 acquired in this step are data (prediction data) referenced in the prediction phase. The first physical property value P11 depends on the physical quantity x1, and the second physical property value P12 also depends on the physical quantity x1. However, the dependencies of these physical properties on the physical quantity differ from each other, and in this example, The first physical property P11 is inversely proportional to the physical quantity x1, The second physical property P12 is proportional to the physical quantity x1.

[0113] It should be noted that the first physical property value P11 and the second physical property value P12 do not have to be simultaneous. The derivation unit 12 (22) identifies the dependency of the first physical property value P11 acquired by the acquisition unit 11 (21) on the physical quantity x1 by referring to the correspondence information CI; The derivation unit 12 (22) selects the second physical property value P12 that has a lower dependency on the physical quantity x1 than the first physical property value P11 by combining the first physical property value P11 and the second physical property value P12, The second physical property value P12 selected by the derivation unit 12 (22) is acquired by the acquisition unit 11 (21) in the process S11-1. The configuration may be as follows.

[0114] (Process S12-1) In process S12-1, the derivation unit 12 (22) combines the first physical property value P11 and the second physical property value P12 acquired in process S11-1 to derive a feature quantity that has reduced dependency on the physical quantity x1 compared to at least one of the first physical property value P11 and the second physical property value P12.

[0115] In this example, the derivation unit 12 (22) derives (generates) the feature quantity F1 by calculating the product of the first physical property value P11 and the second physical property value P12. More specifically, the derivation unit 12 (22) F1=P11·P12 The feature F1 is derived by performing the following calculation. Here, the symbol "·" represents a product. As mentioned above, in this example, The first physical property P11 is inversely proportional to the physical quantity x1, The second physical property P12 is proportional to the physical quantity x1. Therefore, the product of the first physical property value P11 and the second physical property value P12 is substantially independent of the physical quantity x1. In other words, the product of the first physical property value P11 and the second physical property value P12 has reduced dependency on the physical quantity x1 compared to either the first physical property value P11 or the second physical property value P12. Using such a quantity as the feature quantity F1 can contribute to improving the prediction accuracy in the prediction process described below.

[0116] As shown in this example, at least one of the first physical property value P11 and the second physical property value P12 is nonlinearly dependent on the physical quantity x1, and the derivation unit 12 (22) derives the feature value F1 whose nonlinear dependency on the physical quantity x1 is reduced compared to at least one of the first physical property value P11 and the second physical property value P12.

[0117] (Process S11-2) On the other hand, in process S11-2, the acquisition unit 11 (21) acquires a third physical property value P21 and a fourth physical property value P22 for the same object as the object targeted in process S11-1. Here, the third physical property value P21 depends on the physical quantity x2, and the fourth physical property value P22 also depends on the physical quantity x2. However, the dependencies of these physical properties on the physical quantities differ from each other, and in this example, The third physical property P21 is exponentially dependent on the physical quantity x2, as in P21=exp(x2), The fourth physical property P22 is inversely proportional to the physical quantity x2.

[0118] As in the process 11-1, the third physical property value P21 and the fourth physical property value P22 do not have to be calculated simultaneously. The derivation unit 12 (22) identifies the dependency of the third physical property value P21 acquired by the acquisition unit 11 (21) on the physical quantity x2 by referring to the correspondence information CI; The derivation unit 12 (22) selects the fourth physical property value P22 that has a lower degree of dependence on the physical quantity x2 than the third physical property value P21 by combining the third physical property value P21 and the fourth physical property value P22, The fourth physical property value P22 selected by the derivation unit 12 (22) is acquired by the acquisition unit 11 (21) in the process S11-2. The configuration may be as follows.

[0119] (Process S12-2) In process S12-2, the derivation unit 12 (22) combines the third physical property value P21 and the fourth physical property value P22 acquired in process S11-2 to derive a feature quantity that has reduced dependency on the physical quantity x2 compared to at least one of the third physical property value P21 and the fourth physical property value P22.

[0120] In this example, the derivation unit 12 (22) derives (generates) the feature quantity F2 by calculating the product of the logarithm of the third physical property value P21 and the fourth physical property value P22. More specifically, the derivation unit 12 (22) F1 = P22 log(P21) The feature quantity F2 is derived by performing the following calculation. Here, the symbol "·" represents a product. As mentioned above, in this example, The third physical property P21 is exponentially dependent on the physical quantity x2, The fourth physical property P22 is inversely proportional to the physical quantity x2. Therefore, the product of the logarithm of the third physical property value P21 and the fourth physical property value P22 is substantially independent of the physical quantity x2. In other words, the product of the logarithm of the third physical property value P21 and the fourth physical property value P22 has reduced dependency on the physical quantity x2 compared to both the third physical property value P21 and the fourth physical property value P22. Using such a quantity as the feature value F2 can contribute to improving the prediction accuracy in the prediction process described below.

[0121] As shown in this example, at least one of the third physical property value P21 and the fourth physical property value P12 is nonlinearly dependent on the physical quantity x2, and the derivation unit 12 (22) derives the feature value F2 whose nonlinear dependency on the physical quantity x2 is reduced compared to at least one of the third physical property value P21 and the fourth physical property value P22.

[0122] (Process S13) In process S13, the prediction unit 13 inputs the feature value F1 derived (generated) in process S12-1 and the feature value F2 derived (generated) in process S12-2 into a function (prediction model PM) having one or more trained parameters, thereby predicting a physical property value TP of the object to be predicted. More specifically, in this example, the prediction unit 13 uses a linear function having trained parameters α1 and α2 as the prediction model PM, TP = α1 F1 + α2 F2 The physical property value TP of the object to be predicted is predicted by the following equation. In this example, a function including only linear terms is used, but this does not limit the present exemplary embodiment. However, as described above, the derivation unit 12 (22) Deriving a feature value F1 in which the nonlinear dependency on the physical quantity x1 is reduced more than that of at least one of the first physical property value P11 and the second physical property value P12; Derive a feature value F2 in which the nonlinear dependency on the physical quantity x2 is reduced more than that of at least one of the third physical property value P21 and the fourth physical property value P22. The above configuration is adopted, and prediction processing is performed using the features F1 and F2 derived in this way, so the nonlinear effect on TP is reduced compared to when the above-mentioned physical property values ​​are directly input into the prediction model PM.

[0123] Therefore, in processing such as this example, predictions with sufficiently high prediction accuracy can be made even if a prediction model PM is used that does not include nonlinear effects or in which the nonlinear effects are suppressed to a predetermined level or below.

[0124] (Prediction processing example 2) 7 shows prediction processing example 1 by information processing device 100A. This example shows, as an example, processing for predicting a physical property value of a mixture C of objects A and B (physical property value TPC of mixture C) using a plurality of physical property values ​​of object A (physical property values ​​P1A, P2A) and a plurality of physical property values ​​of object B (physical property values ​​P1B, P2B). As shown in FIG. 7, prediction processing example 2 includes processing for acquiring physical property values ​​(S11A, S11B), processing for generating feature amounts (S12A, S12B), and processing for predicting physical property values ​​(S13C).

[0125] (Process S11A) In process S11A, the acquisition unit 11 (21) acquires a first physical property value P1A and a second physical property value P2A, which are physical property values ​​related to the object A. Here, the first physical property value P1A and the second physical property value P2A acquired in this step are data (prediction data) referenced in the prediction phase. The first physical property value P1A depends on the physical quantity x1, and the second physical property value P2A also depends on the physical quantity x1. However, the dependencies of these physical property values ​​on the physical quantity differ from each other, and in this example, The first physical property P1A is inversely proportional to the physical quantity x1, The second physical property P2A is proportional to the physical quantity x1.

[0126] It should be noted that the first physical property value P1A and the second physical property value P2A do not have to be simultaneous. The derivation unit 12 (22) identifies the dependency of the first physical property value P1A acquired by the acquisition unit 11 (21) on the physical quantity x1 by referring to the correspondence information CI; The derivation unit 12 (22) selects the second physical property value P2A that has a lower dependency on the physical quantity x1 than the first physical property value P1A by combining the first physical property value P1A and the second physical property value P2A, The second physical property value P2A selected by the derivation unit 12 (22) is acquired by the acquisition unit 11 (21) in the process S11A. The above configuration may also be used.

[0127] (Process S12A) In process S12A, the derivation unit 12 (22) combines the first physical property value P1A and the second physical property value P2A acquired in process S11A to derive a feature of the object A that has reduced dependence on the physical quantity x1 compared to at least one of the first physical property value P1A and the second physical property value P2A.

[0128] In this example, the derivation unit 12 (22) derives (generates) the feature amount FA of the object A by calculating the product of the first physical property value P1A and the second physical property value P2A. More specifically, the derivation unit 12 (22) FA=P1A·P2A By performing the above calculation, the feature quantity FA of the object A is derived. Here, the symbol "·" represents a product. As mentioned above, in this example, The first physical property P1A is inversely proportional to the physical quantity x1, The second physical property P2A is proportional to the physical quantity x1. Therefore, the product of the first physical property value P1A and the second physical property value P2A is substantially independent of the physical quantity x1. In other words, the product of the first physical property value P1A and the second physical property value P2A has reduced dependency on the physical quantity x1 compared to either the first physical property value P1A or the second physical property value P2A. Using such a quantity as the feature quantity FA of the object A can contribute to improving the prediction accuracy in the prediction process described below.

[0129] As shown in this example, at least one of the first physical property value P1A and the second physical property value P2A is nonlinearly dependent on the physical quantity x1, and the derivation unit 12 (22) derives a feature value FA whose nonlinear dependency on the physical quantity x1 is reduced more than that of at least one of the first physical property value P1A and the second physical property value P2A.

[0130] (Process S11B) In process S11B, the acquisition unit 11 (21) acquires a third physical property value P1B and a fourth physical property value P2B, which are physical property values ​​related to the object B. Here, the third physical property value P1B depends on the physical quantity x2, and the fourth physical property value P2B also depends on the physical quantity x2. This acquisition process is similar to the process of acquiring the first physical property value P1A and the second physical property value P2A, which are physical property values ​​related to the object A, in the above-mentioned process S11A, and therefore will not be described here.

[0131] (Process S12B) In process S12B, the derivation unit 12 (22) combines the third physical property value P1B and the fourth physical property value P2B acquired in process S11B to derive a feature of the object B that has reduced dependence on the physical quantity x2 compared to at least one of the third physical property value P1B and the fourth physical property value P2B.

[0132] In this example, the derivation unit 12 (22) derives (generates) the feature value FB of the object B by taking the product of the third physical property value P1B and the fourth physical property value P2B. This derivation (generation) process is similar to the process of deriving (generating) the feature value FA of the object A by taking the product of the first physical property value P1A and the second physical property value P2A in the above-mentioned process S12A, and therefore will not be described here.

[0133] (Process S13C) In process S13C, the prediction unit 13 inputs the feature values ​​FA of the object A derived (generated) in process S12A and the feature values ​​FB of the object B derived (generated) in process S12B into a function (prediction model PM) having one or more trained parameters, thereby predicting the physical property value TPC of the prediction target for the mixture C. In this example, more specifically, the prediction unit 13C - Trained parameters A1 and B1, and The composition ratio RA of the object A in the mixture C, and · Composition ratio RB of object B in mixture C Using a linear function with the following as the prediction model and PM, TPC = A1 FA RA + B1 FB RB In this example, a function including only linear terms is used, but this does not limit the present exemplary embodiment. More generally, the prediction unit 13 predicts the physical property value TPC of the prediction target for the mixture C by The product of the linear expression (FA) of the feature quantity for object A and the power of the composition ratio (RA) of object 1, and The product of the linear expression (FB) of the feature quantity for object 2 and the power of the composition ratio (RB) of object 2 Alternatively, the derivation unit 12 (22) may use a prediction model PM that uses a linear combination of the following: Deriving a feature value FA of the object A in which the nonlinear dependency on the physical quantity x1 is reduced more than that of at least one of the first physical property value P1A and the second physical property value P2A; Derive a feature value FB of the object B in which the nonlinear dependency on the physical quantity x2 is reduced more than that of at least one of the third physical property value P1B and the fourth physical property value P2B. The above configuration is adopted, and prediction processing is performed using the feature values ​​FA of object A and FB of object B derived in this way, so the nonlinear effect on the TPC is reduced compared to when the above-mentioned physical property values ​​are directly input into the prediction model PM.

[0134] Therefore, in processing such as this example, predictions with sufficiently high prediction accuracy can be made even if a prediction model PM is used that does not include nonlinear effects or in which the nonlinear effects are suppressed to a predetermined level or below.

[0135] (Correspondence information CI) 8 shows an example of the correspondence information CI referred to by the derivation unit 12 (22). The correspondence information CI is information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities. More specifically, in the example shown in FIG. 8, the correspondence information CI is Each of a plurality of physical property values ​​P1, P2, P3, The dependence of each of the plurality of physical property values ​​P1, P2, P3, on the plurality of physical quantities x1, x2, x3, and In the example shown in FIG. 8, the physical property value P1 is, for example, - Proportional to physical quantity x1, · Proportional to the square of the physical quantity x2, Proportional to the logarithm of the physical quantity x3 This is included in the corresponding information CI.

[0136] Such correspondence information CI is Physical or chemical theory Models that simulate physical or chemical phenomena Empirical rules based on experimental results For example, the derivation unit 12 (22) can determine the correspondence information CI in advance by, for example, using data measured by the measurement unit 54 included in the measurement device 50 as the experimental data required to generate the correspondence information CI.

[0137] (Learning process example 1) 9 shows learning process example 1 by the information processing device 100A. This example is a learning process corresponding to the prediction process example 1 described above, and is executed, for example, prior to the execution of the prediction process example 1 described above.

[0138] As shown in FIG. 9 , this example includes the processes S11-1, S11-2, S12-1, S12-2, and S13 described in the prediction process example 1. However, in this example, the multiple physical property values ​​(physical property values ​​P11, P12, P21, and P22) related to the object acquired in processes S11-1 and S11-2 are data (learning data) referenced in the learning phase. Also, in this example, at least one of processes S11-1 and S11-2 further acquires a ground truth label for the physical property values ​​of the object. The ground truth label is referenced in process S14, which will be described later, as an example. Also, in this example, process S13 may be configured to be executed by the learning unit 23. The details of processes S11-1, S11-2, S12-1, S12-2, and S13 have been described above, so a repeated description will be omitted here, and the following description will focus on differences from the already described matters.

[0139] (Process S14) In process S14, at least one of one or more parameters of the prediction model PM is updated so as to reduce the difference between the object property value TP predicted in process S13 and the above-mentioned correct label. As an example, at least one of the above-mentioned parameters α1 and α2 is updated so as to reduce the difference between the object property value TP predicted in process S13 and the above-mentioned correct label. As an example, the prediction model PM (trained prediction model PM) having the parameters updated in process S14 is stored in the storage unit 20 and is referenced in the above-mentioned prediction process example 1. Note that in this example, processes S13 and S14 may be configured to be repeatedly executed until the difference between the object property value TP predicted in process S13 and the above-mentioned correct label satisfies a predetermined convergence condition. In other words, the prediction unit 13 or the learning unit 23 performs the following steps until the predetermined convergence condition is satisfied. Selection of the second physical property value (P21 or P22) by the derivation unit 12; Determining how to combine the first physical property value (P11 or P12) and the second physical property value (P21 or P22), and - Derive (generate) features using the determined combination method However, this is not a limitation of the present exemplary embodiment.

[0140] (Learning process example 2) 10 shows learning process example 2 by the information processing device 100A. This example is a learning process corresponding to the above-mentioned prediction process example 2, and is executed, for example, prior to the execution of the above-mentioned prediction process example 2.

[0141] As shown in FIG. 10 , this example includes the processes S11A, S11B, S12A, S12B, and S13C described in the prediction process example 2. However, in this example, the multiple physical property values ​​(physical property values ​​P1A, P2A, P1B, and P2B) for objects A and B acquired in processes S11A and S11B are data (learning data) referenced in the learning phase. Also, in this example, at least one of processes S11A and S11B further acquires a ground truth label for the physical property values ​​of mixture C. The ground truth label is referenced in process S14C, which will be described later, as an example. Also, in this example, process S13C may be configured to be executed by the learning unit 23. The details of processes S11A, S11B, S12A, S12B, and S13C have been described above, and therefore a repeated description will be omitted here.

[0142] (Process S14C) In process S14C, at least one of one or more parameters of the prediction model PM is updated so as to reduce the difference between the physical property value TPC of the mixture C predicted in process S13C and the above-mentioned correct label. As an example, at least one of the above-mentioned parameters A1 and B1 is updated so as to reduce the difference between the physical property value TPC predicted in process S13C and the above-mentioned correct label. As an example, the prediction model PM (trained prediction model PM) having the parameters updated in process S14 is stored in the storage unit 20 and referenced in the above-mentioned prediction process example 2. Note that in this example, processes S13C and S14C may be configured to be repeatedly executed until the difference between the physical property value TPC predicted in process S13C and the above-mentioned correct label satisfies a predetermined convergence condition. In other words, the prediction unit 13 or the learning unit 23 performs the following steps until the predetermined convergence condition is satisfied. Selection of the second physical property value (P1B or P2B) by the derivation unit 12; Determining how to combine the first physical property value (P1A or P2A) and the second physical property value (P1B or P2B), and - Derive (generate) features using the determined combination method However, this is not a limitation of the present exemplary embodiment.

[0143] (Display example) Fig. 11 shows an example of display data generated by the output unit 14 included in the information processing device 100A and presented to the user via the input / output unit 40. As shown in Fig. 11, the output unit 14 may generate display data including a suggestion to the user according to the result of the analysis process by the derivation unit 12 (22).

[0144] 11 shows an example of display data when the derivation unit 12 (22) determines that a feature quantity F1 in which the nonlinearity of the physical quantity x1 is reduced can be generated by calculating the product of a physical property value P1 and a physical property value P2 among a plurality of physical property values ​​related to the object. In this way, the output unit 14 may present the display data according to the determination result by the derivation unit 12 (22) to the user via the input / output unit 40.

[0145] If the user selects "YES", the feature F1 is adopted as the feature to be referenced by the prediction unit 13 and the learning unit 23, and if the user selects "NO", the feature F1 is not adopted as the feature to be referenced by the prediction unit 13 and the learning unit 23, and other candidate features are presented.

[0146] Note that the display examples by the information processing device 100A are not limited to the above examples. For example, when there are three or more physical property values ​​that are candidates for combination, there are multiple ways to combine the physical property values. In such a case, The derivation unit 12 (22) evaluates the nonlinearity of the physical quantity for each possible combination of physical property values, The output unit 14 displays these combinations in an order according to the nonlinearity of the physical quantities. For example, a combination with weaker nonlinearity may be displayed at a higher rank. For example, if there are candidate physical quantities P1, P2, and P3, ·P1·P2: Nonlinearity to physical quantity x1: Large ·P2·P3: Nonlinearity to physical quantity x1: Small ·P3·P1: Nonlinearity to physical quantity x1: Medium If the derivation unit 12 (22) evaluates that the answer is yes, the output unit 14 notifies the user via the input / output unit 40 that the answer is yes. ·P2·P3: Nonlinearity to physical quantity x1: Small ·P3·P1: Nonlinearity to physical quantity x1: Medium ·P1·P2: Nonlinearity to physical quantity x1: Large It is also possible to present display data including combinations of physical property values ​​listed in the following order: By using such a display method, the user can easily determine what combination of physical property values ​​should be used.

[0147] (Application example) An application example of the information processing device 100A according to this exemplary embodiment will be described below. In this example, the information processing device 100A is applied to the recycling process of waste plastic. When recycling waste plastic, the physical property values ​​of the product must match the required characteristics. However, it is not easy to obtain the history of the waste plastic, and it is also not easy to measure the physical property values ​​and physical quantities of the waste plastic. According to the information processing device 100A according to this exemplary embodiment, even in such a situation, it is possible to preferably execute a prediction process regarding the physical property values ​​of the target object (product). In this example, the information processing device 100A executes the following processes.

[0148] (S101: Output property value selection process) The input / output unit 40 of the information processing device 100A acquires user selection information regarding the type of physical property value to be output. As an example, the input / output unit 40 may display a user interface screen including a plurality of types of physical property values ​​that can be output, and acquire the user selection information as an operation on the screen. The type of physical property value indicated by the selection information is stored in the storage unit 20, for example, and used as the type of target physical property value in the prediction process by the prediction unit 13 or the learning process by the learning unit 23.

[0149] (S102: Input data reception process) Furthermore, the acquisition unit 11 of the information processing device 100A acquires input data via the input / output unit 40 or the communication unit 30. Here, the input data may be data input by a user via the input / output unit 40, or may be data provided from the measurement device 50 via the communication unit 30.

[0150] In this process, the input data acquired by the acquisition unit 11 includes: Basic physical properties of waste plastics to be recycled Qualitative data on the plastic waste to be recycled The blending ratio of each of the waste plastics to be recycled, and the blending ratio in the product produced from the waste plastics to be recycled The acquisition unit 11 or the derivation unit 12 of the information processing device 100A refers to at least one of the above data to identify physical property values ​​related to each of the plastics or waste plastics to be recycled, and refers to the physical property values ​​in the feature amount derivation process by the derivation unit 12.

[0151] The qualitative data may include a character string related to the waste plastic. The character string may also include a product name containing the waste plastic. For example, the acquisition unit 11 or the derivation unit 12 of the information processing device 100A can identify the physical property values ​​of the waste plastic to be recycled by referring to correspondence information in which each of the qualitative data described above is associated with a physical property value.

[0152] It should be noted that the data acquired by the acquisition unit 11 in this process is not limited to the above example. As an example, the acquisition unit 11 may further acquire information regarding the physical property values ​​of virgin plastics used in producing the product. For example, the information processing device 100A: ·Waste plastic material WP1 and Virgin plastic VP1 and When making a prediction regarding a product (mixture) obtained by mixing the above, the acquisition unit 11 may be configured to acquire the physical property values ​​of the virgin plastic VP1 in addition to the physical property values ​​of the waste plastic material WP1, and refer to these physical property values ​​in the feature derivation process in the derivation unit 12.

[0153] (S103: Feature generation process) The derivation unit 12 of the information processing device 100A combines the first physical property value obtained in process S102, which is dependent on one or more physical quantities, with a second physical property value dependent on at least one of the one or more physical quantities, to derive (generate) a feature whose dependence on at least one physical quantity is reduced compared to that of the first physical property value.

[0154] In this example, the process involves estimating the physical quantity of mixed plastics. The target physical property value for which a nonlinear alligation rule originating from the physical quantity that is relatively more difficult to measure than the first physical property value and the second physical property value is established is estimated using the feature value obtained by a conversion process that eliminates the nonlinearity originating from the physical quantity that is relatively more difficult to measure. It can also be expressed as something.

[0155] Such processing is, for example, S103-1: Identify the nonlinear effect of the target physical quantity (corresponding to the first physical property value) on the above-mentioned difficult-to-measure physical quantity, S103-2: Identify the nonlinear effects of other measurable physical quantities (corresponding to the second physical property) on the above-mentioned difficult-to-measure physical quantities, S103-3: The nonlinear effects derived in the above steps S103-1 and S103-2 are combined to derive a feature (generated feature), and a prediction model PM is constructed by, for example, transforming the formula as follows so that the target property value is proportional only to the power term of the composition ratio. Predicted physical property value = A1 × generated feature_1 × composition ratio_1 + ··· + AN × generated feature_N × composition ratio_N + B1 × generated feature_1 × composition ratio_1^2 + … Here, A1, AN, B1, . . . indicate parameters to be learned (updated) by the learning unit 23, and are examples of parameters that the prediction model PM has. Furthermore, the above processes S103-1 and S103-2 may be executed by the derivation unit 12 with reference to the above-mentioned correspondence information CI, for example. Alternatively, they may be configured to be executed with reference to the measurement results by the measurement device 50.

[0156] (S104: Prediction processing) The derivation unit 12 of the information processing device 100A executes a prediction process using, for example, a prediction model PM that refers to the feature amount derived as described above.

[0157] (Specific examples of this application) A more specific example of processing related to this application example will be described below. For example, the measurement device 50 measures the target physical property value E from a plurality of waste plastics WP i By measuring each of the waste plastics WP i Physical property value E i Here, i is an index for distinguishing multiple waste plastics from each other. i The ratio R i Let us consider the case where the ratio R i As an example, the values ​​obtained in the output property value selection process S101 described above can be used.

[0158] Here, it is assumed that the target physical property value E depends on a physical quantity M that is relatively difficult to measure as follows:

number

number

number

number

[0159] An example of the above-mentioned target physical property value E or other physical property value K is MFR (Melt Flow Rate). Here, MFR (Melt Flow Rate) is the amount of resin (converted to weight) that flows per unit time when the thermoplastic resin is in a high-temperature state. From a statistical mechanical perspective, MFR can be understood in terms of molecular weight, specific gravity, and intermolecular interactions, and can be reproduced by simulations such as molecular dynamics. In addition, However, measuring molecular weight is costly and not easy in facilities that recycle waste plastics.

[0160] The derivation unit 12 according to this example is Each waste plastic WP i The first physical property E i (molecular weight) and the second physical property K i By multiplying this by (MFR), each waste plastic WP i Feature A of i of A i = E i K i The calculated feature amount is then referenced to perform the prediction process by the prediction unit 13 or the learning process by the learning unit 23. Alternatively, if the first physical property value and the second physical property value are interchanged, the following expression can be expressed: The first physical property value includes a physical property value related to the fluidity of the object (MFR), The second physical property value includes the specific gravity of the object, The derivation unit 12 derives the feature amount by multiplying the first physical property value and the second physical property value. This may be expressed as the configuration.

[0161] If the configuration does not include the derivation unit 12, the relative error in the prediction process is approximately 100%, but the inventors have verified that by having the derivation unit 12 perform the process described above, the relative error can be reduced to approximately 15%.

[0162] In addition to the first and second physical properties, or in place of some of them, a prediction model PM that refers to temperature dependency can also be used. The above-described processing may also be applied to measurements of nonlinear physics, such as an Izod impact test. In such tests, the nonlinear effects of the first and second physical properties (measured values) are significant, but the derivation unit 12 can derive feature values ​​that suitably improve prediction accuracy.

[0163] Furthermore, this example can be suitably applied to situations where strict physical or chemical theories are difficult to apply. For example, even in a situation where the dependency of the first physical property value and the second physical property value on the feature quantity cannot be precisely determined, the above-mentioned various effects can be achieved as long as the dependency of the first physical property value and the second physical property value on the feature quantity can be roughly grasped or specified.

[0164] Third Exemplary Embodiment A third exemplary embodiment, which is an example of an embodiment of the present invention, will be described in detail with reference to the drawings. Components having the same functions as those described in the above exemplary embodiment will be assigned the same reference numerals, and their description will be omitted as appropriate. The scope of application of each technology employed in this exemplary embodiment is not limited to this exemplary embodiment. That is, each technology employed in this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise. Furthermore, each technology shown in each drawing referenced to explain this exemplary embodiment can also be employed in other exemplary embodiments included in the present disclosure, to the extent that no particular technical obstacles arise.

[0165] (Configuration of information processing device 100B) The configuration of an information processing device 100B according to this exemplary embodiment will be described with reference to Fig. 12. Fig. 12 is a block diagram showing the configuration of the information processing device 100B. As shown in Fig. 12, the information processing device 100B does not include the learning unit 23, which is one of the components included in the information processing device 100A according to exemplary embodiment 2. The other components are the same as those of the information processing device 100A.

[0166] In this way, the information processing device 100B: An acquisition means (acquisition unit 11 (21)) for acquiring a first physical property value that depends on one or more physical quantities; a derivation means (deriving unit 12 (22)) that combines the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; A prediction means (prediction unit 13) that executes a prediction process by referring to the feature amount; output means (output unit 14) for outputting the processing result by the information processing device 100B; This configuration also provides the various effects of the exemplary embodiments described above.

[0167] [Software implementation example] Some or all of the functions of the information processing devices 1, 2, 100A, and 100B (hereinafter also referred to as "each of the above devices") may be realized by hardware such as an integrated circuit (IC chip), or by software.

[0168] In the latter case, each of the above devices is realized by, for example, a computer that executes instructions of a program, which is software that realizes each function. An example of such a computer (hereinafter referred to as computer C) is shown in Figure 13. Figure 13 is a block diagram showing the hardware configuration of computer C that functions as each of the above devices.

[0169] The computer C includes at least one processor C1 and at least one memory C2. The memory C2 stores a program P for causing the computer C to operate as each of the above-mentioned devices. In the computer C, the processor C1 reads and executes the program P from the memory C2, thereby realizing the functions of each of the above-mentioned devices.

[0170] The processor C1 may be, for example, a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), a micro processing unit (MPU), a floating point number processing unit (FPU), a physics processing unit (PPU), a tensor processing unit (TPU), a quantum processor, a microcontroller, or a combination thereof. The memory C2 may be, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or a combination thereof.

[0171] The computer C may further include a RAM (Random Access Memory) for expanding the program P during execution and for temporarily storing various data. The computer C may also include a communication interface for transmitting and receiving data to and from other devices. The computer C may also include an input / output interface for connecting input / output devices such as a keyboard, mouse, display, and printer.

[0172] Furthermore, the program P can be recorded on a non-transitory tangible recording medium M that can be read by the computer C. Such a recording medium M can be, for example, a tape, a disk, a card, a semiconductor memory, or a programmable logic circuit. The computer C can acquire the program P via such a recording medium M. The program P can also be transmitted via a transmission medium. Such a transmission medium can be, for example, a communication network or broadcast waves. The computer C can also acquire the program P via such a transmission medium.

[0173] Furthermore, the functions of each of the devices may be realized by a single processor provided in a single computer, by multiple processors provided in a single computer working in cooperation, or by multiple processors provided in each of multiple computers working in cooperation. Furthermore, the programs for causing each of the devices to realize the functions may be stored in a single memory provided in a single computer, or may be distributed and stored in multiple memories provided in a single computer, or may be distributed and stored in multiple memories provided in each of multiple computers.

[0174] [Appendix A] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0175] (Appendix A1) an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; a prediction means for executing a prediction process by referring to the feature amount; An information processing device comprising:

[0176] (Appendix A2) at least one of the first physical property value and the second physical property value nonlinearly depends on at least one of the physical quantities; The deriving means derives the feature quantity in which nonlinear dependency on the at least one physical quantity is reduced compared to the at least one physical property value. 10. The information processing device according to claim 1,

[0177] (Appendix A3) The derivation means selecting the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities; 10. The information processing device according to claim 9, wherein the information processing device is a

[0178] (Appendix A4) the first physical property value includes a physical property value related to the fluidity of the object, the second physical property value includes a specific gravity of the object; The deriving means derives the feature quantity by multiplying the first physical property value and the second physical property value. An information processing device according to any one of appendices A1 to A3.

[0179] (Appendix A5) The prediction means The physical property value of the prediction target is predicted by a linear combination of one or more feature quantities including the feature quantity derived by the derivation means. An information processing device according to any one of appendices A1 to A3.

[0180] (Appendix A6) an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; learning means for executing a learning process by referring to the feature amount; An information processing device comprising:

[0181] [Appendix B] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0182] (Appendix B1) an acquisition process in which at least one processor acquires a first physical property value that depends on one or more physical quantities; a derivation process in which the at least one processor combines the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; a prediction process in which the at least one processor executes a prediction process by referring to the feature amount; An information processing method comprising:

[0183] (Appendix B2) at least one of the first physical property value and the second physical property value nonlinearly depends on at least one of the physical quantities; In the derivation process, the at least one processor derives the feature quantity in which nonlinear dependency on the at least one physical quantity is reduced compared to the at least one physical property value. The information processing method described in Appendix B1.

[0184] (Appendix B3) In the derivation process, the at least one processor selecting the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities; 1. The information processing method described in Appendix B2.

[0185] (Appendix B4) the first physical property value includes a physical property value related to the fluidity of the object, the second physical property value includes a specific gravity of the object; In the derivation process, the at least one processor derives the feature quantity by multiplying the first physical property value and the second physical property value. 1. An information processing method according to any one of appendices B1 to B3.

[0186] (Appendix B5) In the prediction process, the at least one processor The physical property value of the prediction target is predicted by a linear combination of one or more feature quantities including the feature quantity derived by the derivation process. 1. An information processing method according to any one of appendices B1 to B3.

[0187] (Appendix B6) an acquisition process in which the at least one processor acquires a first physical property value that depends on one or more physical quantities; a derivation process in which the at least one processor combines the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; a learning process in which the at least one processor executes a learning process with reference to the feature amount; An information processing method comprising:

[0188] [Appendix C] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0189] (Appendix C1) A program that causes a computer to function as an information processing device, The computer an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; a prediction means for executing a prediction process by referring to the feature amount; An information processing program that functions as a

[0190] (Appendix C2) at least one of the first physical property value and the second physical property value nonlinearly depends on at least one of the physical quantities; The deriving means derives the feature quantity in which nonlinear dependency on the at least one physical quantity is reduced compared to the at least one physical property value. An information processing program as described in Appendix C1.

[0191] (Appendix C3) The derivation means selecting the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities; An information processing program as described in Appendix C2.

[0192] (Appendix C4) the first physical property value includes a physical property value related to the fluidity of the object, the second physical property value includes a specific gravity of the object; The deriving means derives the feature quantity by multiplying the first physical property value and the second physical property value. An information processing program according to any one of appendices C1 to C3.

[0193] (Appendix C5) The prediction means The physical property value of the prediction target is predicted by a linear combination of one or more feature quantities including the feature quantity derived by the derivation means. An information processing program according to any one of appendices C1 to C3.

[0194] (Appendix C6) The computer an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; a learning process that executes a learning process by referring to the feature amount; An information processing program that functions as a

[0195] [Appendix D] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0196] (Appendix D1) at least one processor, an acquisition process for acquiring a first physical property value that depends on one or more physical quantities; a derivation process of combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; a prediction process that executes a prediction process by referring to the feature amount; An information processing device that executes the above.

[0197] The information processing device may further include a memory, and the memory may store a program for causing the at least one processor to execute each of the processes.

[0198] (Appendix D2) at least one of the first physical property value and the second physical property value nonlinearly depends on at least one of the physical quantities; In the derivation process, the at least one processor derives the feature quantity in which nonlinear dependency on the at least one physical quantity is reduced compared to the at least one physical property value. 10. The information processing device according to claim 9, wherein said information processing device is a

[0199] (Appendix D3) In the derivation process, the at least one processor selecting the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities; 10. The information processing device according to claim 9, wherein the information processing device is an information processing device according to claim 1, wherein

[0200] (Appendix D4) the first physical property value includes a physical property value related to the fluidity of the object, the second physical property value includes a specific gravity of the object; In the derivation process, the at least one processor derives the feature quantity by multiplying the first physical property value and the second physical property value. An information processing device according to any one of appendices D1 to D3.

[0201] (Appendix D5) In the prediction process, the at least one processor The physical property value of the prediction target is predicted by a linear combination of one or more feature quantities including the feature quantity derived by the derivation process. An information processing device according to any one of appendices D1 to D3.

[0202] (Appendix D6) The at least one processor: an acquisition process for acquiring a first physical property value that depends on one or more physical quantities; a derivation process of combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; a learning process that executes a learning process by referring to the feature amount; An information processing device that executes the above.

[0203] [Appendix E] This disclosure includes the techniques described in the following appendices. However, the present invention is not limited to the techniques described in the following appendices, and various modifications are possible within the scope of the claims.

[0204] (Appendix E1) A program that causes a computer to function as an information processing device, The computer, an acquisition process for acquiring a first physical property value that depends on one or more physical quantities; a derivation process of combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; a prediction process that executes a prediction process by referring to the feature amount; A non-transitory recording medium on which an information processing program for executing the above is recorded. [Explanation of symbols]

[0205] 1, 2, 100A, 100B Information processing equipment 1A Information Processing System 10, 51 Control unit 11, 21 Acquisition Department 12, 22 Derivation part 13 Prediction Department 14 Output section 20, 52 storage section 23 Learning Department 30, 53 Communications Department 40 Input / output section 50 Measuring Equipment 54 Measuring part C1 processor C2 Memory

Claims

1. an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; a prediction means for executing a prediction process by referring to the feature amount; An information processing device comprising:

2. at least one of the first physical property value and the second physical property value nonlinearly depends on at least one of the physical quantities; The deriving means derives the feature quantity in which nonlinear dependency on the at least one physical quantity is reduced compared to the at least one physical property value.

2. The information processing device according to claim 1.

3. The derivation means and selecting the second physical property value to be combined with the first physical property value from the plurality of physical property values ​​by referring to correspondence information in which each of a plurality of physical property values ​​that are candidates for the first and second physical property values ​​is associated with the dependency of each of the plurality of physical property values ​​on one or more physical quantities. The information processing device according to claim 2 .

4. the first physical property value includes a physical property value related to the fluidity of the object; the second physical property value includes a specific gravity of the object; The deriving means derives the feature quantity by multiplying the first physical property value and the second physical property value. The information processing device according to claim 1 .

5. The prediction means The physical property value of the prediction target is predicted by a linear combination of one or more feature quantities including the feature quantity derived by the derivation means. The information processing device according to claim 1 .

6. an acquisition means for acquiring a first physical property value that depends on one or more physical quantities; a derivation means for combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value; learning means for executing a learning process by referring to the feature amount; An information processing device comprising:

7. Obtaining a first physical property value that depends on one or more physical quantities; deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value by combining the first physical property value with a second physical property value that depends on at least one physical quantity of the one or more physical quantities; Executing a prediction process with reference to the feature amount; An information processing method comprising:

8. Obtaining a first physical property value that depends on one or more physical quantities; deriving a feature quantity whose dependency on at least one physical quantity is reduced compared to that of the first physical property value by combining the first physical property value with a second physical property value that depends on at least one physical quantity of the one or more physical quantities; Executing a learning process with reference to the feature amount; An information processing method comprising:

9. A program that causes a computer to function as an information processing device, The program causes the computer to: an acquisition process for acquiring a first physical property value that depends on one or more physical quantities; a derivation process of combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; A prediction process that refers to the feature amount. A program that executes the following.

10. A program that causes a computer to function as an information processing device, The program causes the computer to: an acquisition process for acquiring a first physical property value that depends on one or more physical quantities; a derivation process of combining the first physical property value with a second physical property value that depends on at least one of the one or more physical quantities, thereby deriving a feature quantity whose dependency on the at least one physical quantity is reduced compared to that of the first physical property value; A learning process that refers to the feature amount. A program that executes the following.

Images