Method for obtaining model of spectrometer or spectroscope
The spectrum and parameters are encrypted through homomorphic encryption methods, allowing the model builder to analyze and create models of spectrometers or spectrometers, solving the problem of difficult protection of customer privacy in the existing technology and achieving efficient model creation and data privacy protection.
Patent Information
- Application Number
- CN202380059752.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-08-19
- Filing Date
- 2023-08-18
- Publication Date
- 2025-03-14
AI Technical Summary
The prior art is difficult to create models for spectrometers or spectrometers without compromising customer privacy, especially when processing encrypted customer data, requiring multiple manual steps and processing of sensitive information.
The homomorphic encryption method is used to encrypt spectra and parameters, allowing the model builder to analyze encrypted data and create models without decoding sensitive information. Ensure privacy of customer data by selecting, tuning, or recreating models.
The ability to create models for spectrometers or spectrometers without exposing customer sensitive information is achieved, improving data privacy protection and model creation efficiency.
Smart Images
Figure CN119631136A_ABST
Abstract
Description
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS
[0002] This application is related to and claims the benefit of priority from German Patent Application No. 102022121066.9 filed on August 19, 2022, the entire contents of which are incorporated herein by reference. Technical Field
[0003] The present disclosure relates to a method for obtaining a model of a spectrometer or a spectroscope, wherein the model contains a relationship between a spectrum of a gas mixture or a gas and a parameter of the gas mixture or the gas, in particular the composition of the gas mixture and / or the concentration of each component. The present disclosure also relates to a method for evaluating the spectrum of a gas or a gas mixture detected by a spectrometer or a spectroscope by means of the model obtained by the above method. Background Art
[0004] Induced radiation effects, such as Raman scattering and fluorescence, have become very valuable tools for the non-destructive determination of molecular composition. Raman spectroscopy is an established and practical chemical analysis and characterization method that can be applied to many different chemical fabrics. As a real-time, non-destructive technique, Raman spectroscopy is compatible with a wide range of samples, including opaque solids, aqueous solutions, emulsions and gases, without the need for sample preparation.
[0005] In addition to Raman spectrometers, various other types of spectrometers or spectroscopes exist, including IR spectroscopes and NIR spectroscopes.
[0006] Each type of spectrometer or spectroscope behaves uniquely in recording the spectrum of a gas mixture. A model is used to evaluate the spectrum, in particular to obtain parameters of the gas mixture, such as the composition and / or concentration of the individual components of the gas mixture. Such a model contains one or more regression algorithms or parameters for one or more such regression algorithms. The model has parameters for a regression algorithm, wherein the regression algorithm calculates the parameters from the spectrum.
[0007] Creating models for customers currently requires a lot of detailed manual work, which often uses sensitive customer data. The content of this data indicates which chemical components are present and in what amounts, especially to experienced chemometricians. Therefore, companies may be reluctant to disclose information that could reveal trade secrets. This makes it difficult to provide modeling as a service. Since this service can be valuable to companies and help expand the customer base of spectroscopic devices, there is a need to be able to create models without compromising the privacy of customers. There are multiple encryption methods to hide the content of the information, but none of these methods are designed to address the unique privacy needs of spectroscopic applications.
[0008] There are two main points to consider when dealing with encrypted customer data for spectroscopy applications, especially for Raman applications:
[0009] The model creation process requires many manual steps that are usually performed by chemometricians or expert modelers.
[0010] Additionally, some additional information must be hidden from the model builder to ensure that customer data is protected (specifically which wavelengths or frequencies of the spectrum are relevant to the model). Summary of the invention
[0011] Based on this problem, an object of the present disclosure is to propose a method that allows a model builder to create a model for a customer's spectrometer or spectroscope without knowing the customer's data.
[0012] The object is achieved by a method for obtaining a model of a spectrometer or spectroscope, wherein the model comprises a relationship between a spectrum of a gas mixture or a gas and a parameter of the gas mixture or the gas, in particular the composition and / or concentration of the individual components of the gas mixture:
[0013] - acquiring a reference spectrum of the predetermined gas or gas mixture by means of a spectrometer or a spectroscope and acquiring parameters of the predetermined gas or gas mixture, in particular the composition of the predetermined gas mixture and / or the concentrations of the individual components;
[0014] - Encryption of spectra and / or parameters by means of homomorphic encryption methods;
[0015] - Analyze spectra and parameters without decoding them;
[0016] - performing one or more of the following steps on a model based on the analysis, wherein the model includes a relationship between a spectrum of a gas mixture and a parameter:
[0017] i. Select a model from a large number of pre-created models,
[0018] ii. Adjust the previously created model, and
[0019] iii. Recreate the model.
[0020] Therefore, according to the present disclosure, the customer reference spectrum is hidden from the model builder using homomorphic encryption techniques. However, due to the properties of the reference spectrum encoded by homomorphic methods, the modeler can still analyze the spectrum and create a customized model - without the modeler being able to see the reference spectrum and the frequency content associated with the reference spectrum (and in some cases, even the parameters of the model itself).
[0021] Spectrometers and spectroscopes are measuring instruments that record the spectrum of a gas mixture and enable the individual components of the gas mixture to be identified with high precision. While spectrometers provide information about which components are present in the gas mixture, they also provide statements about the amounts of the respective components - in other words, they indicate for their measuring range how intense the radiation is at the respective observation wavelength.
[0022] According to an advantageous further development of the method disclosed herein, it is provided that the steps of recording and encryption are performed by a first instance (in particular a client), wherein the encrypted spectrum and / or the encrypted parameters are sent to a second instance (in particular a service provider), wherein the steps of analysis and steps regarding model execution are performed by the second instance. For this purpose, one or more computers are provided, with which the corresponding method steps are performed. It can also be provided that the cloud is accessed from the instances IN1, IN2, on which one or more method steps are performed.
[0023] According to an advantageous embodiment of the method disclosed herein, it is provided that the reference spectrum is shuffled by a first instance before encryption, wherein information about the type of shuffling is collected, wherein this information is not sent to the second instance. For example, the reference spectrum is divided into frequency intervals, which are then interchanged one after the other with respect to the order. This increases security, since even if the reference spectrum is decoded, the exact reference spectrum cannot be obtained.
[0024] According to an advantageous embodiment of the method disclosed herein, it is provided that a key pair comprising a private key and a public key is created on a first instance, wherein the public key is sent to a second instance, wherein the model is encrypted by the second instance with the aid of the public key and can be decrypted by the private key. This increases security by preventing a third party (e.g. in the case of a hacker) from interpreting the model.
[0025] According to an advantageous embodiment of the method disclosed herein, it is provided that the model has parameters for a regression algorithm, wherein the regression algorithm calculates the parameters from the spectrum.
[0026] According to an advantageous embodiment of the method of the present disclosure, a step is provided in which a machine learning or AI algorithm is used to select, adapt or recreate the model.
[0027] Furthermore, the object is solved by a method for evaluating a spectrum of a gas or a gas mixture detected by a spectrometer or spectroscope by means of a model obtained by the above-described method according to the present disclosure, wherein parameters of the gas or the gas mixture are obtained by evaluating the spectrum.
[0028] According to an advantageous embodiment of the method according to the present disclosure, it is provided that the model is sent to the first instance and wherein the evaluation is performed by the first instance. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] The present disclosure is explained in more detail with reference to the following drawings, in which:
[0030] Figure 1 A flow chart showing a first embodiment of a method according to the present disclosure; and
[0031] Figure 2 A flow chart of a second embodiment of a method according to the present disclosure is shown. DETAILED DESCRIPTION
[0032] exist Figure 1 In a first embodiment of the method according to the present disclosure, a suitable model is selected and adapted to a spectrometer or spectroscope. The model is specific to a spectrometer or spectroscope and describes the relationship between the spectrum detected by the spectrometer or spectroscope and the parameters of the gas mixture, in particular the composition of the gas mixture and / or the concentrations of the individual components whose spectrum is detected by means of the spectrometer. In other words, the model makes it possible to determine the parameters of the gas mixture from the spectrum detected by it.
[0033] The spectrometer or spectroscope is used by the customer and is referred to here as a first instance IN1.
[0034] The task of selecting and fitting an appropriate model is given to a model builder, here referred to as the second instance IN2.
[0035] In a first method step a), the customer acquires a reference spectrum of a known gas mixture containing at least two components by means of a spectrometer or a spectroscope. A known gas mixture means that the customer knows exactly the parameters of the gas mixture, such as the composition of the gas mixture and / or the concentrations of the individual components. A spectrum usually has frequency or wavelength on one axis, while the other axis represents intensity. The intensity is measured over a frequency or wavelength range by means of a spectrometer or a spectroscope. Different gas mixtures have different characteristic curves. With the aid of the model, statements can be made about how a spectrometer or a spectroscope captures and extends these characteristic developments - because each spectrometer or a spectroscope represents these developments differently. By applying the model to the spectra of unknown gas mixtures, their parameters can therefore also be determined. Thus, in particular, the model contains parameters for a regression algorithm that is able to calculate the parameters from the spectrum.
[0036] In method step b), a key pair is created in the first instance IN1. This consists of a public key and a private key. With the help of the public key, information or data can be encrypted. The public key is used to decrypt the encrypted data.
[0037] Since the reference spectrum and parameters (if applicable) are sensitive information that should not be disclosed to the model builder, the reference spectrum and parameters (if applicable) are encrypted in method step c) by means of a homomorphic encryption method. Homomorphic encrypted data can be processed or offset by its special properties, without the second instance IN2 performing the processing needing to know the contents of the unencrypted file. The result data that has been processed or applied can be decoded again by the client, whereby the processing or application remains valid.
[0038] For example, partially homomorphic encryption methods or fully homomorphic encryption methods can be considered for this purpose. Homomorphic encryption methods or cryptographic systems can be classified by their homomorphic properties:
[0039] Partially homomorphic encryption methods exist, for example, as (partial) additively homomorphic encryption methods with the following properties:
[0040] ;
[0041] Or as a (partial) multiplicative homomorphic encryption method with the following properties.
[0042] .
[0043] Additionally, there are fully homomorphic encryption methods that have both additive and multiplicative homomorphic properties.
[0044] The reference spectrum and the public key encrypted in this way and, if applicable, the parameters of the known gas mixture are sent in method step d) to the second instance IN2 and received by it.
[0045] Subsequently, the coded reference spectrum is analyzed in method step e) and, if necessary, associated with parameters. Based on this analysis, a suitable model for the spectrometer or spectroscope is selected from various models. If necessary, this is still adjusted in method step f).
[0046] For method steps e) and f), one or more AI or machine learning algorithms are used, which have been pre-learned on various spectra and, if necessary, on parameters. Due to the homomorphic encryption method used, the second instance IN2 does not have to parse the encryption in order to perform method steps e) and f). Subsequently, the instance IN2 generates the results of these method steps, in particular information about the model (e.g. information about the regression), encrypts them with the help of the public key and sends them to the first instance IN1.
[0047] After the results have been received by the first instance, they are decoded in method step g). Subsequently, their rationality is checked, in particular the model performance is tested. If the results are not rational for the client, a corresponding feedback is reported to the second instance IN2, thereby repeating method steps e) and f). In the event of confirmation of rationality by the client, a corresponding feedback is reported to the second instance IN2.
[0048] In a subsequent method step h), the selected or adapted model is encrypted by means of a public key and sent to the first instance IN 1. Encryption means that sensitive information contained in the model cannot be interpreted by unauthorized persons.
[0049] In the final method step i), the model is decrypted by the first instance IN1 with the aid of the private key. Subsequently, a final evaluation of the model is carried out. For this purpose, for example, additional spectra of other known gas mixtures are recorded. With the aid of the model, the parameters of the respective gas mixture are subsequently determined and the actually known parameters of the other gas mixtures are used to select the other known gas mixtures. If the final evaluation is successful, the spectrometer or spectroscope is operated with the model.
[0050] In the event that the final evaluation is unsuccessful, the method can be repeated from method step e).
[0051] exist Figure 2 In a second embodiment of the method according to the present disclosure, a suitable model is created for the spectrometer or spectroscope and adapted if necessary. Figure 1 , but differ in various method steps.
[0052] In a first method step a′), the customer acquires a reference spectrum of a known gas mixture containing at least two components by means of a spectrometer or a spectroscope.
[0053] In an optional method step b'), the reference spectra are mixed in the first instance IN1. This means, for example, that the frequencies are no longer arranged in ascending order, but are divided into frequency intervals which are interchanged with one another in sequence. Information about how exactly the mixing was done is stored by the first instance IN1. With the help of this information, the reference spectra can be sorted back into the correct order.
[0054] In method step c′), the first instance IN1 creates a key pair consisting of a public key and a private key.
[0055] The reference spectrum and the parameters (if necessary) in method step d′) are subsequently encrypted by means of a homomorphic encryption method.
[0056] The encrypted reference spectrum is then received in method step e') together with the public key and analyzed in method step f') and, if necessary, associated with the parameters. Based on this analysis, one or more suitable models for the spectrometer or spectroscope are created and, if necessary, adapted in method step f').
[0057] For method steps f') and g'), one or more AI or machine learning algorithms are used, which have been pre-learned on various spectra and, if necessary, on various parameters. Due to the homomorphic encryption method used, the second instance IN2 does not have to parse the encryption in order to perform method steps f') and g'). The AI or machine learning algorithm is also able to process the mixed and encrypted reference spectra accordingly. Subsequently, the instance IN2 generates the results of these method steps, in particular information about the model (e.g., information about the regression), encrypts them with the help of the public key and sends them to the first instance IN1.
[0058] After the results have been received by the first instance, they are decoded in method step g). Subsequently, it is checked which model has the highest model performance. To this end, the results are compared with the preferences. If the results are not satisfactory for the customer, corresponding feedback is reported to the second instance IN2, so that method steps f') and g') are repeated. The finally selected model is transmitted to the second instance IN2.
[0059] In a subsequent method step i'), the selected model is encrypted by means of the public key and sent to the first instance IN1. Encryption means that sensitive information contained in the model cannot be interpreted by unauthorized persons.
[0060] In a final method step j'), the model is decrypted by the first instance IN1 with the aid of the private key. Subsequently, a final evaluation of the model is carried out. For this purpose, for example, additional spectra of other known gas mixtures are recorded. With the aid of the model, the parameters of the respective gas mixture are subsequently determined and the actually known parameters of the other gas mixture are used to select the other known gas mixture. If the final evaluation is successful, the spectrometer or spectroscope is operated with the model.
[0061] In the event that the final evaluation is unsuccessful, the method can be repeated from method step f′).
[0062] The method according to the present disclosure is not limited to the two embodiments mentioned. It goes without saying for a person skilled in the art that he can also combine the various method steps mentioned only in one embodiment in each case.
[0063] On one side or in the instance IN1, IN2 means that there are one or more computers with which the method steps are performed. This also includes that the instance IN1, IN2 has access to a cloud on which one or more method steps can be performed.
Claims
1. A method for obtaining a model of a spectrometer or a spectroscope, wherein: The model comprises a relationship between a spectrum of a gas mixture or a gas and parameters of the gas mixture or the gas, in particular the composition of the gas mixture and / or the concentrations of individual components, the method comprising: - acquiring a reference spectrum of a predetermined gas or a gas mixture by means of the spectrometer or the spectroscope, and acquiring parameters of the predetermined gas or the gas mixture, in particular the composition of the predetermined gas mixture and / or the concentrations of the individual components; - encrypting the spectrum and / or the parameters by means of a homomorphic encryption method; - analyzing said spectrum and said parameter without decoding said spectrum or said parameter; - performing one or more of the following steps on a model based on the analysis, wherein the model comprises a relationship between the spectrum of the gas mixture and the parameter: i. Select a model from a large number of pre-created models, ii. Adjust the previously created model, and iii. Recreate the model.
2. The method according to claim 1, wherein: The steps of recording and of encrypting are performed by a first instance (IN1), in particular a client, wherein the encrypted spectrum and / or parameters are sent to a second instance (IN2), in particular a service provider, wherein the steps of analyzing and of executing the steps with respect to the model are performed by the second instance (IN2).
3. The method according to claim 2, wherein: The reference spectrum is mixed by the first instance (IN1) before scrambling, wherein information about the type of mixing is collected, wherein the information is not sent to the second instance (IN2).
4. The method according to claim 2 or 3, wherein: A key pair consisting of a private key and a public key is created on the first instance (IN1), wherein the public key is sent to the second instance (IN2), wherein the model is encrypted by the second instance (IN2) using the public key and can be decrypted using the private key.
5. The method according to one or more of the preceding claims, wherein: The model has parameters for a regression algorithm, wherein the regression algorithm calculates the parameters from the spectrum.
6. The method according to any one or more of the preceding claims, wherein: Machine learning or AI algorithms are used for said steps of selecting, said adjusting or said recreating said model.
7. A method for evaluating the spectrum of a gas or a gas mixture detected by a spectrometer or spectroscope with the aid of a model obtained by the method according to one or more of claims 1 to 6, wherein: By evaluating the spectrum, parameters of the gas or the gas mixture are obtained.
8. The method according to claim 7, wherein: The model is sent to the first instance (IN1), and wherein the evaluating is performed by the first instance (IN1).