Apparatus and method for analyzing characteristics of superconductor on basis of artificial intelligence technology

Abstract

An apparatus and a method for analyzing the characteristics of a superconductor on the basis of artificial intelligence technology are disclosed. The analysis apparatus comprises: a spectrum generation unit for generating a plurality of point-contact andreev reflection spectroscopy by combining a plurality of variables including temperature, superconducting gap, interfacial barrier strength, and broadening factor of a superconductor; and a training unit for training, on the basis of training data including the point-contact andreev reflection spectroscopy and the variables, an analysis model to which artificial intelligence technology is applied.

Description

Device and method for analyzing the characteristics of superconductors based on artificial intelligence technology

[0001] The present invention relates to a technology for analyzing superconducting characteristics, and more specifically, to a device and method for analyzing the characteristics of a superconductor using artificial intelligence technology that quickly and objectively analyzes characteristics such as gap and symmetry of a superconductor based on artificial intelligence technology.

[0002] Point-contact Andreev Reflection Spectroscopy (PCAR) is a spectroscopic method that checks the current that changes due to Andreev reflection that occurs when current flows at the junction of a normal metal and a superconductor (Fig. 1 (a)). Here, Andreev reflection refers to the phenomenon in which holes are reflected when electrons are incident between the normal metal and the superconductor (Fig. 2). In other words, the point-contact spectroscopy shows that the propagation of holes has the same effect as the propagation of electrons in the opposite direction, so when theoretically simulated, it shows a phenomenon in which the conductivity doubles (Fig. 1 (b)). However, this changes depending on the characteristics of the superconductor, temperature, and the state of the interface, and is based on the Blonder, Tinkham, and Klapwijk (BTK) theory.

[0003] The BTK theory, as illustrated in Fig. 3, calculates the probability of occurrence of each phenomenon A, B, C, and D for variables such as temperature (T), superconducting gap (△), interface barrier strength (Z), and broadening factor (Γ). Through this, the theoretical PCAR spectrum according to the variables can be simulated. In other words, by comparing the spectrum simulated with the BTK theory with experimental results, it is possible to analyze the characteristics of superconductors (especially, the superconducting gap).

[0004] However, generating the theoretical spectra described above requires numerous parameters, and analyzing them individually against experimental data is time-consuming. Furthermore, the noise and errors inherent in experimental data often lead to significant subjectivity on the part of the researcher during the comparative analysis process with the theoretical spectra.

[0005] The problem to be solved by the present invention is to provide a device and method for analyzing the characteristics of a superconductor using artificial intelligence technology, which very quickly and objectively analyzes the characteristics of a superconductor, such as superconducting gap and symmetry information, using artificial intelligence technology.

[0006] In order to solve the above problem, the analysis device according to the present invention includes a spectrum generation unit that generates a plurality of point contact spectral spectra by combining a plurality of variables including a temperature of a superconductor, a superconducting gap, an interface barrier strength, and a broadening factor, and a learning unit that trains an analysis model to which artificial intelligence technology is applied based on the point contact spectral spectra and learning data including the variables.

[0007] In addition, the above spectrum generation unit is characterized in that it performs simulations under multiple conditions using the BTK (Blonder Tinkham Klapwijk) theory and generates multiple point contact spectral spectra through the simulations.

[0008] In addition, the spectrum generation unit is characterized in that it performs the simulation while fixing the values ​​of the remaining variables except for one variable among the plurality of variables and changing the value of the one variable.

[0009] In addition, the learning unit is characterized in that it sets the point contact spectral spectrum and the temperature used when generating the point contact spectral spectrum as input values ​​of the analysis model, and sets at least one of the superconducting gap, interface barrier strength, and broadening factor used when generating the point contact spectral spectrum as an output value of the analysis model, and then performs the learning.

[0010] In addition, the learning unit is characterized in that it divides the learning data into a training set, a verification set, and a test set, performs the learning using the training set, verifies the accuracy of the analysis model being learned using the verification set, adjusts the weights, and evaluates the performance of the analysis model for which learning has been completed using the test set.

[0011] In addition, it is characterized by further including an analysis unit that analyzes the characteristics of a specific superconductor by inputting a point contact spectral spectrum and a preset temperature of the specific superconductor into the learned analysis model.

[0012] In addition, the analysis unit is characterized by analyzing at least one of the superconducting gap, symmetry, interface barrier strength, and broadening factor of the specific superconductor.

[0013] An analysis method performed by an analysis device for analyzing the characteristics of a superconductor according to the present invention includes a step of generating a plurality of point contact spectral spectra by combining a plurality of variables including a temperature, a superconducting gap, an interface barrier strength, and a broadening factor of the superconductor, and a step of training an analysis model to which artificial intelligence technology is applied based on the point contact spectral spectra and training data including the variables.

[0014] In addition, the above generating step is characterized by performing a simulation under multiple conditions using the BTK theory and generating multiple point contact spectral spectra through the simulation.

[0015] In addition, the generating step is characterized by fixing the values ​​of the remaining variables except for one variable among the plurality of variables, and performing the simulation while changing the value of the one variable.

[0016] In addition, the learning step is characterized in that the point contact spectral spectrum and the temperature used when generating the point contact spectral spectrum are set as input values ​​of the analysis model, and at least one of the superconducting gap, interface barrier strength, and broadening factor used when generating the point contact spectral spectrum is set as an output value of the analysis model, and then the learning is performed.

[0017] In addition, the learning step is characterized by dividing the learning data into a training set, a verification set, and a test set, performing the learning using the training set, verifying the accuracy of the analysis model being learned using the verification set, adjusting the weights, and evaluating the performance of the analysis model for which learning has been completed using the test set.

[0018] In addition, it is characterized by further including a step of analyzing the characteristics of the specific superconductor by inputting the point contact spectral spectrum and the preset temperature of the specific superconductor into the learned analysis model.

[0019] In addition, the analyzing step is characterized by analyzing at least one of the superconducting gap, symmetry, interface barrier strength, and broadening factor of the specific superconductor.

[0020] An analysis system according to the present invention includes an analysis device that analyzes the characteristics of a specific superconductor based on an analysis model applied with artificial intelligence technology, and a user terminal that receives the analyzed results from the analysis device and outputs the received results remotely, wherein the analysis device is characterized by including a spectrum generation unit that generates a plurality of point contact spectral spectra by combining a plurality of variables including a temperature of the superconductor, a superconducting gap, an interface barrier strength, and a broadening factor, and a learning unit that learns the analysis model based on the point contact spectral spectra and learning data including the variables.

[0021] According to an embodiment of the present invention, a large number of point contact spectral spectra for various variable combinations are generated based on the BTK theory, the spectra are learned through artificial intelligence technology to implement an analysis model, and the superconducting point contact spectral spectra can be analyzed very quickly and objectively using the implemented analysis model.

[0022] It can also be applied to automated or autonomous systems for research on superconducting materials, and its accuracy can be gradually improved through continuous learning.

[0023] This will accelerate research on superconducting materials, including topological superconductors, thereby increasing the efficiency of superconductivity research and contributing to the development of various cutting-edge technologies, including quantum computing.

[0024] Figure 1 is a drawing for explaining a point contact spectroscopy spectrum.

[0025] Figure 2 is a diagram explaining the Andreev reflex.

[0026] Figure 3 is a diagram for explaining the BTK theory.

[0027] Figure 4 is a schematic diagram illustrating an analysis system according to an embodiment of the present invention.

[0028] Figure 5 is a block diagram illustrating an analysis device according to an embodiment of the present invention.

[0029] Figure 6 is a block diagram for explaining a control unit according to an embodiment of the present invention.

[0030] FIG. 7 is a drawing for explaining a process for generating a point contact spectroscopy spectrum according to an embodiment of the present invention.

[0031] FIG. 8 is a drawing for explaining the modification of the point contact spectral spectrum for various variables according to an embodiment of the present invention.

[0032] FIG. 9 is a diagram for explaining learning data consisting of a training set, a verification set, and a test set according to an embodiment of the present invention.

[0033] Figure 10 is a drawing for explaining the structure of an analysis model according to an embodiment of the present invention.

[0034] Figure 11 is a drawing comparing an analysis method of the prior art and an analysis method according to the present invention.

[0035] Figure 12 is a flowchart for explaining a learning method according to an embodiment of the present invention.

[0036] Figure 13 is a flowchart for explaining an analysis method according to an embodiment of the present invention.

[0037] Figure 14 is a block diagram illustrating a computing device according to an embodiment of the present invention.

[0038] Below, with reference to the attached drawings, embodiments of the present invention are described in detail so that those skilled in the art can easily implement the invention. However, the present invention can be implemented in various different forms and is not limited to the embodiments described herein. In addition, in the drawings, parts irrelevant to the description are omitted for clarity of description, and similar parts are designated with similar reference numerals throughout the specification.

[0039] In this specification and drawings (hereinafter referred to as “this specification”), duplicate descriptions of identical components are omitted.

[0040] Additionally, when a component is referred to herein as being "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may be other components present in between. Conversely, when a component is referred to herein as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components present in between.

[0041] Additionally, the terms used herein are only used to describe specific embodiments and are not intended to limit the present invention.

[0042] Also, in this specification, singular expressions may include plural expressions unless the context clearly indicates otherwise.

[0043] In addition, in this specification, it should be understood that terms such as “include” or “have” are intended to specify only the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, and do not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0044] Also, in this specification, the term 'and / or' includes a combination of multiple listed items or any item among multiple listed items. In this specification, 'A or B' can include 'A', 'B', or 'both A and B'.

[0045] Additionally, in this specification, detailed descriptions of known functions and configurations that may obscure the gist of the present invention will be omitted.

[0046]

[0047] Figure 4 is a schematic diagram illustrating an analysis system according to an embodiment of the present invention.

[0048] Referring to FIG. 4, the analysis system (300) uses artificial intelligence technology to very quickly and objectively analyze superconductor properties, such as superconducting gap and symmetry information. The analysis system (300) includes an analysis device (100) and a user terminal (200).

[0049] The analysis device (300) generates multiple point contact spectral spectra for various variable combinations based on the BTK theory. The analysis device (300) learns an analysis model to which artificial intelligence technology is applied using the generated point contact spectral spectra. Here, the analysis model may include a deep learning-based neural network, and preferably, a convolutional neural network (CNN), but is not limited thereto. The analysis device (300) can analyze the point contact spectral spectrum of a specific superconductor to be analyzed based on the analysis model for which learning has been completed. Through this, the analysis device (300) can secure accuracy while drastically reducing the required time compared to conventional comparative analysis methods (e.g., trial and error, grid search, etc.).

[0050] The user terminal (200) is a terminal used by a user (or administrator) and communicates with the analysis device (300). The user terminal (200) receives the analyzed results from the analysis device (300) and outputs the received results remotely. In other words, the user terminal (200) helps the user to check the analyzed results in real time even from a remote location far away from the analysis device (100). In addition, the user terminal (200) can remotely transmit information related to a specific superconductor that the user wants to analyze to the analysis device (100) and control the analysis device (100) to perform analysis on the specific superconductor.

[0051] Meanwhile, the analysis system (300) establishes a communication network (350) between the analysis device (100) and the user terminal (200) to support communication between them. The communication network (350) may be composed of a backbone network and a subscriber network. The backbone network may be composed of one or more integrated networks among an X.25 network, a Frame Relay network, an ATM network, an MPLS (Multi-Protocol Label Switching) network, and a GMPLS (Generalized Multi-Protocol Label Switching) network. The subscriber network may be FTTH (Fiber To The Home), ADSL (Asymmetric Digital Subscriber Line), cable network, zigbee, bluetooth, Wireless LAN (IEEE 802.11b, IEEE 802.11a, IEEE 802.11g, IEEE 802.11n), Wireless Hart (ISO / IEC62591-1), ISA100.11a (ISO / IEC 62734), CoAP (Constrained Application Protocol), MQTT (Message Queuing Telemetry Transport), WIBro (Wireless Broadband), Wimax, 3G, HSDPA (High Speed ​​Downlink Packet Access), 4G, 5G, and 6G. In some embodiments, the communication network (350) may be an Internet network or a mobile communication network. In addition, the communication network (350) may include any other widely known or future-developed wireless or wired communication method.

[0052]

[0053] Figure 5 is a block diagram illustrating an analysis device according to an embodiment of the present invention.

[0054] Referring to FIGS. 4 and 5, the analysis device (100) includes a communication unit (10), an input unit (30), a control unit (50), an output unit (70), and a storage unit (90).

[0055] The communication unit (10) performs communication with the user terminal (200). The communication unit (10) transmits the results of the characteristic analysis of a specific superconductor to the user terminal (200) and supports the output of the results from the user terminal (200). In addition, the communication unit (10) receives information related to a specific superconductor that the user wishes to analyze from the user terminal (200) and supports the control of the characteristic analysis of the specific superconductor remotely.

[0056] The input unit (30) directly receives specific superconductor-related information desired to be analyzed by the user. In addition, the input unit (30) receives various conditional variables required for learning the analysis model and the spectrum generated using these variables.

[0057] The control unit (50) performs overall control of the analysis device (100). The control unit (50) analyzes the characteristics of a specific superconductor using a previously learned analysis model. This enables the control unit (50) to perform a fast and accurate analysis.

[0058] Here, the control unit (50) generates a plurality of point contact spectral spectra under various conditions using the BTK theory to learn a deep learning-based analysis model. The control unit (50) can learn the analysis model based on learning data including the generated plurality of point contact spectral spectra and the variables used to generate each point contact spectral spectra.

[0059] The output unit (70) outputs the analyzed results from the control unit (50). The output unit (70) may be a display, and may include a liquid crystal display (LCD), a thin film transistor-liquid crystal display (TFT LCD), an organic light-emitting diode (OLED), a flexible display, a 3D display, etc.

[0060] The storage unit (90) stores a program or algorithm to drive the analysis device (100). The storage unit (90) stores learning data, information related to a specific superconductor, and an analysis model. The storage unit (90) may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc.

[0061]

[0062] FIG. 6 is a block diagram illustrating a control unit according to an embodiment of the present invention, FIG. 7 is a diagram illustrating a process of generating a point contact spectral spectrum according to an embodiment of the present invention, FIG. 8 is a diagram illustrating a deformation of a point contact spectral spectrum according to various variables according to an embodiment of the present invention, FIG. 9 is a diagram illustrating learning data consisting of a training set, a verification set, and a test set according to an embodiment of the present invention, and FIG. 10 is a diagram illustrating a structure of an analysis model according to an embodiment of the present invention. Here, (a) of FIG. 8 is a diagram illustrating a spectral deformation that changes according to a change in the value of a temperature (T) variable, (b) of FIG. 8 is a diagram illustrating a spectral deformation that changes according to a change in the value of a superconducting gap (△) variable, (c) of FIG. 8 is a diagram illustrating a spectral deformation that changes according to a change in the value of an interface barrier strength (Z) variable, and (d) of FIG. 8 is a diagram illustrating a spectral deformation that changes according to a change in the value of a broadening factor (Γ) variable.

[0063] Referring to FIGS. 6 to 10, the control unit (50) includes a spectrum generation unit (51), a learning unit (53), and an analysis unit (55).

[0064] The spectrum generating unit (51) generates a plurality of point contact spectral spectra by combining a plurality of variables including at least one of the temperature (T), the superconducting gap (△), the interface barrier strength (Z), and the broadening factor (Γ) of the superconductor. In other words, the spectrum generating unit (51) can generate a point contact spectral spectra by combining various variables.

[0065] The spectrum generation unit (51) performs various simulations under multiple conditions using the BTK (Blonder Tinkham Klapwijk) theory. The spectrum generation unit (51) can generate multiple point contact spectral spectra through simulations under various conditions. At this time, the spectrum generation unit (51) can perform the simulation by fixing the values ​​of the remaining variables except for one variable among the multiple variables and changing the value of one variable that is not fixed (Fig. 8).

[0066] The learning unit (53) trains an analysis model to which artificial intelligence technology is applied. The learning unit (53) generates learning data including a plurality of point contact spectral spectra generated from the spectrum generation unit (51) and variables used in generating each point contact spectral spectra, and trains the analysis model based on the generated learning data.

[0067] The learning unit (53) divides the learning data into a training set, a validation set, and a test set. Specifically, the learning unit (53) can divide the total learning data into 80% and 20%. The learning unit (53) sets the learning data corresponding to 80% again as a training set (80%) and a validation set (20%), trains the analysis model, and sets the learning data corresponding to 20% as a test set to evaluate the performance of the analysis model. For example, the learning unit (53) performs learning of the analysis model using the training set, and verifies the accuracy of the analysis model being learned using the validation set. At this time, the learning unit (53) adjusts the weights according to the verification result to determine the coefficient of determination (R) of the analysis model. 2The learning unit (53) can be controlled so that the coefficient of determination (C) approaches 100%. The learning unit (53) evaluates the performance of the analysis model for which learning has been completed using the test set. If the evaluated performance is below a preset standard, the learning unit (53) can retrain the analysis model using the training set and the verification set. Here, the preset standard may be a coefficient of determination of 95% to 99% or more, but is not limited thereto and can be changed by the user.

[0068] Here, the analysis model may include a deep learning-based neural network. Preferably, the analysis model may include a convolutional neural network composed of a convolution layer, a max pooling layer, and a fully connected layer. The convolution layer performs a convolution operation using a convolution filter and an operation using an activation function. Here, the activation function may be a Rectified Linear Unit (ReLU) that can describe a nonlinear relationship. The max pooling layer performs a pooling (or sub-sampling) operation using a pooling filter. The fully connected layer consists of multiple operation nodes and performs an operation using an activation function.

[0069] That is, the learning unit (53) can set the point contact spectral spectrum and the temperature used when generating the contact spectral spectrum among the learning data as input values ​​of the analysis model, and set at least one of the superconducting gap, interface barrier strength, and broadening factor used when generating the point contact spectral spectrum as an output value of the analysis model, and then perform learning. At this time, the learning unit (53) can repeatedly perform learning of the analysis model until the user reaches a desired determination coefficient.

[0070] When information related to a specific superconductor that a user wants to analyze is input, the analysis unit (55) inputs the point contact spectral spectrum and the preset temperature included in the information into a pre-learned analysis model to analyze the characteristics of the specific superconductor. The analysis unit (55) can output at least one of the superconducting gap, interface barrier strength, and broadening factor of the point contact spectral spectrum through the analysis model, and can analyze the superconductor based on the output variable. At this time, the analysis unit (55) can obtain fast and accurate results by analyzing the characteristics of the superconductor based on the analysis model.

[0071] The analysis unit (55) controls the output unit (70) to output the analyzed results. In addition, the analysis unit (55) transmits the analyzed results to the user terminal (200) so that the user terminal (200) can output the results.

[0072]

[0073] Figure 11 is a drawing comparing an analysis method of the prior art and an analysis method according to the present invention.

[0074] Referring to Figure 11, after obtaining an experimental spectrum, a comparison analysis was conducted with the conventional trial and error method, grid search, and the present invention based on the obtained spectrum.

[0075] When conducting analysis using the trial and error method, it is difficult to predict how long it will take (it may take several minutes or several hours depending on each situation), and when conducting analysis using the grid search method, the analysis time depends on the range of variables being analyzed, so there is a time error for each user conducting the analysis (if the user is an expert, the analysis will proceed a little faster, and if the user is an ordinary person, the analysis will proceed for a long time).

[0076] However, the analysis method of the present invention utilizes artificial intelligence technology to perform analysis, thereby supporting very fast and accurate result analysis (less than 100 ms).

[0077] In addition, when the experimentally measured spectral deformation is complex or includes features unrelated to the characteristics of the superconductor (e.g., experimental errors, noise, etc.), analysis through deviations such as grid search has a problem in that it is difficult to derive accurate results. However, the present invention enables accurate analysis by taking these errors into account.

[0078]

[0079] Figure 12 is a flowchart for explaining a learning method according to an embodiment of the present invention.

[0080] Referring to Figures 4 and 12, the learning method generates a large number of point-contact spectral spectra for various variable combinations based on the BTK theory, and implements an analysis model by learning the spectra using artificial intelligence technology. The learning method repeatedly performs learning until the analysis model obtains a coefficient of determination exceeding a preset standard.

[0081] In step S110, the analysis device (100) receives inputs of various condition variables. The analysis device (100) receives multiple variables including the temperature of the superconductor, the superconducting gap, the interface barrier strength, and the broadening factor. At this time, the analysis device (100) can receive the variables remotely or directly.

[0082] In step S120, the analysis device (100) generates a plurality of point contact spectral spectra. The analysis device (100) generates a plurality of point contact spectral spectra by combining a plurality of input variables. At this time, the analysis device (100) performs simulations under a plurality of conditions using the BTK theory, and can generate a plurality of point contact spectral spectra through the performed simulations.

[0083] In step S130, the analysis device (100) learns an analysis model. The analysis device (100) learns an analysis model to which artificial intelligence technology is applied based on a plurality of point contact spectral spectra and learning data including the above variables. The analysis device (100) sets the point contact spectral spectra and the temperature used when generating the point contact spectral spectra as input values ​​of the analysis model, and sets at least one of the superconducting gap, the interface barrier strength, and the broadening factor used when generating the point contact spectral spectra as an output value of the analysis model, and then performs learning.

[0084] At step S140, the analysis device (100) determines whether learning of the analysis model is complete. If the determination coefficient of the analysis model is greater than or equal to a preset standard, the analysis device (100) determines that learning is complete and performs step S150. If the determination coefficient of the analysis model is less than the preset standard, the analysis device determines that learning is incomplete and re-performs step S130.

[0085] In step S150, the analysis device (100) performs a performance test of the analysis model. The analysis device (100) evaluates the performance of the analysis model for which learning has been completed. At this time, the analysis device (100) can evaluate the performance of the analysis model using learning data not used during the learning process.

[0086] At step S160, the analysis device (100) determines whether the performance of the analysis model is above a preset standard. If the performance of the analysis model is above the preset standard, the analysis device (100) determines that no further learning is necessary and terminates the learning-related algorithm. If the performance of the analysis model is below the preset standard, the analysis device determines that further learning is necessary and re-performs step S110.

[0087]

[0088] Figure 13 is a flowchart for explaining an analysis method according to an embodiment of the present invention.

[0089] Referring to Figures 4 and 13, the analysis method utilizes an analytical model to perform a very fast and objective analysis of superconducting point contact spectroscopy spectra. The analysis method can be applied to automated or autonomous systems for superconducting materials research, and its accuracy can be gradually improved through continuous learning. This analysis method can accelerate research on superconducting materials, including topological superconductors, thereby enhancing the efficiency of superconductivity research and contributing to the advancement of various cutting-edge technologies, including quantum computing.

[0090] In step S210, the analysis device (100) receives information related to a specific superconductor. Here, the specific superconductor refers to a superconductor that the user wishes to analyze, and the information related to the specific superconductor refers to a point contact spectroscopy spectrum obtained from the specific superconductor. The analysis device (100) can receive the information remotely or directly.

[0091] In step S220, the analysis device (100) analyzes the characteristics of a specific superconductor. The analysis device (100) inputs the point contact spectroscopy spectrum and the preset temperature included in the input information into a pre-learned analysis model to analyze the characteristics of the specific superconductor. At this time, the analysis device (100) can analyze at least one of the superconducting gap, symmetry, interface barrier strength, and broadening factor of the specific superconductor.

[0092]

[0093] Figure 14 is a block diagram illustrating a computing device according to an embodiment of the present invention.

[0094] Referring to FIG. 14, the computing device (TN100) may be a device (e.g., an analysis device, a user terminal, etc.) described herein.

[0095] The computing device (TN100) may include at least one processor (TN110), a transmitter / receiver device (TN120), and a memory (TN130). In addition, the computing device (TN100) may further include a storage device (TN140), an input interface device (TN150), an output interface device (TN160), and the like. The components included in the computing device (TN100) may be connected by a bus (TN170) to communicate with each other.

[0096] The processor (TN110) can execute program commands stored in at least one of the memory (TN130) and the storage device (TN140). The processor (TN110) may refer to a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor in which methods according to embodiments of the present invention are performed. The processor (TN110) may be configured to implement procedures, functions, methods, etc. described in relation to embodiments of the present invention. The processor (TN110) may control each component of the computing device (TN100).

[0097] The memory (TN130) and the storage device (TN140) can each store various information related to the operation of the processor (TN110). The memory (TN130) and the storage device (TN140) can each be configured with at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory (TN130) can be configured with at least one of a read-only memory (ROM) and a random access memory (RAM).

[0098] The transceiver (TN120) can transmit or receive wired or wireless signals. The transceiver (TN120) can be connected to a network to perform communication.

[0099]

[0100] Meanwhile, the embodiments of the present invention are not implemented only through the devices and / or methods described so far, but may also be implemented through a program that realizes a function corresponding to the configuration of the embodiments of the present invention or a recording medium on which the program is recorded, and such implementation can be easily implemented by a person skilled in the art to which the present invention pertains based on the description of the embodiments described above.

[0101]

[0102] Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims also fall within the scope of the present invention.

Claims

1. A spectrum generation unit that generates multiple point contact spectral spectra by combining multiple variables including the temperature of the superconductor, the superconducting gap, the interface barrier strength, and the broadening factor; and A learning unit that learns an analysis model using artificial intelligence technology based on the point contact spectral spectrum and learning data including the variables; An analysis device including:

2. In paragraph 1, The above spectrum generating unit, An analysis device characterized in that it performs simulations under multiple conditions using the BTK (Blonder Tinkham Klapwijk) theory and generates multiple point contact spectral spectra through the simulations.

3. In paragraph 2, The above spectrum generating unit, An analysis device characterized in that the simulation is performed while fixing the values ​​of the remaining variables except for one variable among the plurality of variables and changing the value of the one variable.

4. In paragraph 1, The above learning department, An analysis device characterized in that the point contact spectral spectrum and the temperature used when generating the point contact spectral spectrum are set as input values ​​of the analysis model, and at least one of the superconducting gap, interface barrier strength, and broadening factor used when generating the point contact spectral spectrum is set as an output value of the analysis model, and then the learning is performed.

5. In paragraph 1, The above learning department, An analysis device characterized in that the above learning data is divided into a training set, a verification set, and a test set, the learning is performed using the training set, the accuracy of the analysis model being learned is verified using the verification set, and then the weights are adjusted, and the performance of the analysis model for which learning has been completed is evaluated using the test set.

6. In paragraph 1, An analysis unit that analyzes the characteristics of a specific superconductor by inputting the point contact spectral spectrum and a preset temperature of the specific superconductor into the learned analysis model; An analysis device characterized by further including:

7. In paragraph 6, The above analysis unit, An analysis device characterized in that it analyzes at least one of the superconducting gap, symmetry, interface barrier strength and broadening factor of the above-mentioned specific superconductor.

8. In an analysis method performed by an analysis device that analyzes the characteristics of a superconductor, A step of generating multiple point contact spectral spectra by combining multiple variables including the temperature, superconducting gap, interface barrier strength, and broadening factor of the superconductor; and A step of training an analysis model using artificial intelligence technology based on the point contact spectral spectrum and learning data including the variables; Analysis method including.

9. In paragraph 8, The above generating steps are: An analysis method characterized in that a simulation is performed under multiple conditions using the BTK theory and a plurality of point contact spectral spectra are generated through the simulation.

10. In paragraph 8, The above generating steps are: An analysis method characterized in that the simulation is performed while fixing the values ​​of the remaining variables except for one variable among the plurality of variables and changing the value of the one variable.

11. In paragraph 8, The above learning steps are: An analysis method characterized in that the point contact spectral spectrum and the temperature used when generating the point contact spectral spectrum are set as input values ​​of the analysis model, and at least one of the superconducting gap, interface barrier strength, and broadening factor used when generating the point contact spectral spectrum is set as an output value of the analysis model, and then the learning is performed.

12. In paragraph 8, The above learning steps are: An analysis method characterized in that the learning data is divided into a training set, a verification set, and a test set, the learning is performed using the training set, the accuracy of the analysis model being learned is verified using the verification set, the weights are adjusted, and the performance of the analysis model for which learning has been completed is evaluated using the test set.

13. In paragraph 8, A step of analyzing the characteristics of a specific superconductor by inputting the point contact spectral spectrum and the preset temperature of the specific superconductor into the learned analysis model; An analysis method characterized by further including:

14. In paragraph 13, The above analysis steps are: An analysis method characterized by analyzing at least one of the superconducting gap, symmetry, interface barrier strength and broadening factor of the above-mentioned specific superconductor.

15. An analysis device that analyzes the characteristics of a specific superconductor based on an analysis model using artificial intelligence technology; and A user terminal that receives the analyzed results from the analysis device and outputs the received results remotely; The above analysis device, A spectrum generating unit that generates multiple point contact spectral spectra by combining multiple variables including the temperature of the superconductor, the superconducting gap, the interface barrier strength, and the broadening factor; and A learning unit that learns the analysis model based on the point contact spectral spectrum and learning data including the variables; An analysis system characterized by including:

Images