EP4544462A1 - Methode de validation des predictions d'un modele supervise d'analyse quantitative multivariee de donnees spectrales - Google Patents
Methode de validation des predictions d'un modele supervise d'analyse quantitative multivariee de donnees spectralesInfo
- Publication number
- EP4544462A1 EP4544462A1 EP23730364.9A EP23730364A EP4544462A1 EP 4544462 A1 EP4544462 A1 EP 4544462A1 EP 23730364 A EP23730364 A EP 23730364A EP 4544462 A1 EP4544462 A1 EP 4544462A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- prediction
- model
- species
- predictions
- learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/71—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light thermally excited
- G01N21/718—Laser microanalysis, i.e. with formation of sample plasma
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1293—Using chemometrical methods resolving multicomponent spectra
Definitions
- the invention relates to the field of quantitative analysis, for example the determination of the concentration of chemical species contained in a sample, and supervised qualitative analysis, for example the classification of samples, from spectral data , that is to say data which presents a plurality of intensity values in different wavelength channels or spectral bands.
- the data can be both multi- or hyperspectral data, where the number of spectral bands varies from a few dozen to hundreds, and data from emission or absorption spectra of a chemical species, containing thousands of channels of wavelength.
- the invention relates to a new multivariate analysis method for the quantitative analysis of chemical species contained in a sample from spectral data, acquired by a spectroscopy technique.
- An objective of the invention concerns the determination of a confidence measure of the predictions of the model used for quantification, using a multi-output algorithm. More precisely, in the context of the invention, the model gives both the main prediction, for example the value of the concentration of a species from the spectrum, and secondary outputs on tasks related to this prediction, d 'where the need for a multi-output system. These secondary predictions are then used to measure confidence in the primary prediction.
- the invention aims to simultaneously predict directly verifiable quantities (i.e. i.e., present in the experimental data both during learning and both in inference) in the experimental data and which thus make it possible to ensure the reliability of the predicted concentration.
- a possible application of the invention concerns the determination of the concentration of chemical elements and an indicator of reliability of predictions from spectral data, for example acquired by means of a measurement technique.
- laser-induced plasma atomic emission spectroscopy or “Laser-Induced Breakdown Spectroscopy” (LIBS) in English.
- LIBS Laser-Induced Breakdown Spectroscopy
- the invention is not limited to this particular technique, it can be applied to any type of spectroscopy technique which produces multi- or hyperspectral data or spectral data of emission or absorption of chemical species.
- LIBS technology makes it possible to carry out a material analysis by focusing a laser beam on the surface of a sample.
- the plasma emission resulting from this focusing is collected by a spectrometer.
- the data acquired by this method are spectral data which correspond, for each focus point on the surface, to an emission spectrum comprising atomic and molecular lines characteristic of the elementary chemical composition of the sample. The intensity of these lines increases non-trivially with the concentration of chemical elements present in the sample.
- the calibration is carried out using several standards, that is to say samples whose species concentrations are previously known, to obtain a model which makes it possible to link the spectral signatures to the concentration of the species. This model can then be used to predict the unknown concentration of a species from a spectrum.
- the model there are methods for defining a measure of uncertainty around the predictions made by the model (e.g., confidence intervals).
- the calibration is characterized by relative uncertainties which increase as the concentration of species in the standards decreases, up to the detection limit. In fact, a level of 100% relative uncertainty is sometimes used as the definition of the detection limit.
- the limitations of the LIBS technique are typical of any spectroscopy technique: validation of the reliability of predictions is always problematic and relative uncertainties close to the detection limit are by definition important.
- the invention aims to overcome these limits by proposing: • A multivariate and multi-output quantification model to obtain both predictions of species concentrations and values that can be used to determine their confidence. To this end, the invention introduces a technique for validating the predictions of quantitative models and for establishing a confidence measure of the predictions or for identifying the presence of anomalies (i.e. predictions which do not have a good level of confidence).
- the invention aims to solve problems linked to the determination of the concentration of species from spectral data which can be produced by LIBS spectroscopy or by other spectroscopy methods (for example, multi- or hyperspectral imaging). .
- this type of data is characterized by specific spectra of the species present in a sample.
- the quantitative analysis of spectral signatures (for example, using the intensity of emission or absorption lines of chemical elements) finally makes it possible to determine the concentration of species.
- several problems can arise.
- the trained model provides in all cases a prediction based on a measurement, without verifying that the data used for learning are actually representative of the real data. There is therefore a need for a tool to verify the reliability of predictions when conditions of use of the model vary slightly compared to the learning conditions or on the contrary in order to ensure that the conditions of use of the model have not changed.
- This information can then be used to obtain a calibration function (for example, a straight line) which links the concentration of the species considered to the signal intensity for each of the standards, for which the concentration of the species is known .
- a calibration function for example, a straight line
- this procedure defines a relationship between concentration and spectral intensity.
- the calibration function can then be used to obtain predictions of the concentration of a species in an unknown sample by inverting this relationship, for example by means of interpolation as described in reference [2],
- the invention deals with the validation of predictions, rather than that of the model.
- the invention relates to a technique for associating a confidence measure of predictions using information available at any time, even in unknown data, and thus comparing it directly to a ground truth value. In the invention, this is achieved by the introduction of multi-output models, and in particular by the introduction of deep learning architectures, capable of efficiently processing the information contained in the data.
- the invention relates to a method for validating predictions from a multi-output model. That is to say, the secondary outputs are measurable experimentally to allow the relevance of the main output, assumed to be unknown during inference, to be assessed.
- the proposed invention includes an additional step compared to state-of-the-art methods. Multi-output algorithms are used to predict secondary outputs that are verifiable in the experimental data, so that a degree of confidence in the predictions can be provided. This is not possible when only predicting the concentration (or concentrations) of the chemical species.
- the invention proposes to use trained algorithms to predict both the concentration of a species from spectral data, but also to provide secondary outputs predicting additional quantities, such as the emission or absorption intensity.
- additional quantities such as the emission or absorption intensity.
- This additional data must be measurable and experimentally verifiable during inference.
- the prediction of these values must be sufficiently complicated for the model, compared to the main prediction. That is to say, the prediction must be made on a non-trivial task (whose complexity is comparable to that of the main prediction) from the input data to avoid an imbalance during the learning process. For example, it is possible to use the intensity of spectral lines or bands integrated on neighboring wavelength channels.
- the invention makes it possible to provide, for any concentration prediction, an indicator of reliability of the prediction which makes it possible, for example, to rule out certain measurements subject to hazards which will render the prediction provided by the model on these measurements , unreliable.
- the subject of the invention is a method, implemented by computer, for automatically learning a multi-output prediction model configured to jointly determine, from a set of spectral data characteristic of a sample, at least one main prediction of at least one first physical quantity characterizing a given species in the sample and at least one secondary prediction of at least one second physical quantity characterizing said species, the multi-output prediction model being trained from a set of annotated spectral data.
- the multi-output prediction model is a multi-task neural network and is implemented by means of a first common learning engine configured to extract sets of spectral data received as input, representations common to the different tasks to be solved and several learning engines specific to each task to be solved, each receiving said common representations as input and delivering as output a prediction corresponding to the task to be solved.
- the common learning engine is a convolutional neural network and the specific neural networks are convolutional neural networks supplemented by layers of fully connected neurons.
- said species is a chemical species
- the main prediction is a concentration value of the chemical species
- the secondary prediction is an intensity value of a spectral line for at least a given wavelength or at least one wavelength band of given width.
- the invention also relates to a prediction model implemented by computer, obtained according to the automatic learning method according to the invention.
- the invention also relates to a method, implemented by computer, for quantitative analysis of spectral data comprising the implementation of the prediction model according to the invention to determine, from a spectrum measured on a sample, at least one main prediction of at least one first physical quantity characterizing a given species in the sample and at least one secondary prediction of at least one second physical quantity characterizing said species, the method further comprising a step of calculation of a reliability indicator of the at least one main prediction from an indicator of the difference between at least one secondary prediction and a value of the second corresponding physical quantity measured on the spectrum.
- the reliability indicator is equal to the relative error between the intensity value predicted via the implementation of the prediction model and the corresponding intensity value measured on the spectrum.
- the reliability indicator is equal to the difference, in absolute value, between the intensity value predicted via the implementation of the prediction model and the intensity value corresponding measured on the spectrum divided by the standard deviation of this deviation.
- the quantitative analysis method according to the invention further comprises the implementation of a classification model configured to classify the predictions of concentration of chemical species into two classes corresponding to values normals and anomalies, based on predictions of spectral line intensity values or reliability indicators.
- the measured spectrum is acquired by means of a laser-induced plasma atomic emission spectroscopy method.
- the invention also relates to a computer program comprising instructions for the execution of a method according to the invention, when the program is executed by a processor as well as a recording medium readable by a processor on which a program is recorded comprising instructions for executing a method according to the invention, when the program is executed by a processor.
- FIG. 1 represents an example of spectral data characterizing a sample containing different chemical species
- FIG. 2 represents a diagram illustrating the learning and use of a multi-output prediction model according to the invention
- FIG. 3 represents a diagram of an example of implementation of the prediction model according to the invention by means of one or more convolutional neural networks
- FIG. 4 represents an example table of results of learning and analysis of secondary outputs.
- LIBS technology makes it possible to carry out material analysis by laser ablation and spectroscopy.
- the data acquired via this technique are spectral data which correspond, for each point in an area, to an emission spectrum comprising atomic and molecular lines characteristic of the elementary chemical composition of the sample.
- the LIBS spectral data are obtained by focusing a laser beam at a point on a surface to be analyzed.
- the plasma emission resulting from this focusing is collected and processed by spectroscopy to obtain a spectrum of atomic lines (or molecular bands) of emission. The process is iterated for each point in the area to be analyzed.
- Figure 1 represents, by way of illustration, an example of a spectrum of atomic lines 101 obtained for a sample having a certain chemical composition.
- Figure 1 we have identified the spectral signatures of certain chemical elements (Ca, Al) which correspond to atomic lines in given wavelength channels.
- An objective of the invention is to determine a prediction model capable of estimating a concentration of a chemical element from an automatic analysis of a spectrum such as that of Figure 1 and of further providing a indicator of reliability of the estimate provided.
- Figure 2 represents, in a diagram, a method of learning and using a prediction model according to the invention.
- the method consists, in a learning phase, of determining a prediction model 231 configured to determine, from spectra measured on given samples, a concentration of one or more chemical species on the one hand and a prediction of the intensity of one or more atomic lines (or molecular bands) in the spectrum on the other hand.
- the method then consists, in a use phase, of using the trained model to determine these predictions on new measured spectral data.
- the secondary outputs of the model are used to determine an indicator of prediction reliability.
- the method uses as input spectral input data 210 in the form of several sets of spectra obtained by LIBS spectroscopy in order to characterize a given sample.
- the spectral data 210 are labeled, that is to say, for example in the case of quantitative analysis, that we know the concentrations of the different chemical elements to be quantified in the sample.
- the input data 210 is constituted by a set of pairs each associating a spectrum with a concentration of one or more chemical elements in a sample of a given type.
- a sample is characterized, for example, by a type of material and a concentration of certain chemical elements in this material.
- the input data 210 are separated into a first subset of training data 221 and a second subset of evaluation data 222.
- the model 231 is trained from the training data 221 then it is optimized on the evaluation data 222 during an optimization cycle 232 to determine the best hyperparameters of the model 231.
- the choice of the percentage of realizations used for learning 221 may depend on the calculation support and the type of data to maximize the learning capacity of the architectures. For example, for a dataset containing 100,000 realizations typically we can use 80% for learning, but for datasets with millions of realizations the percentage can increase, unless the calculation support does not allow it.
- the learning data 221 is used directly for calculating the model parameters, while the evaluation data 222 is used to evaluate the predictions and optimize the model 231.
- the prediction model 231 is a multivariate and multi-output statistical model. It receives the entire spectra as input data and it is trained to predict on the one hand a first set of primary outputs 251 corresponding to one or more concentration predictions of one or more chemical species and on the other hand leaves a second set of secondary outputs 252 corresponding to one or more predictions of intensity values of atomic lines in certain wavelength ranges.
- the model 231 can be applied to a new set of spectral data 240.
- the first set of predictions 251 is used to determine the concentrations of the chemical species in the sample on which the spectrum d was measured. input 240.
- the second set of predictions 252 is processed to determine a confidence measure of the first predictions 251.
- the confidence measure is based on the interpretation in terms of probability (for example using the probability distribution of a given estimator, presented below) or relative error of the predictions of the secondary outputs 252 of the model 231. These outputs must predict quantities, linked to the prediction of species concentration, present in the unknown real data 240, so that a comparison between real values and predicted values makes it possible to quantify the reliability of the learning and predictions .
- the secondary outputs 252 were trained from the same representation of the input data 210, the secondary predictions 252 are linked to the species concentration predictions 251 and they share at least a subset of the model weights (implementations called “hard parameter sharing”, in English). We can then assume that a reliable result on the secondary outputs can lead to equally reliable predictions on the concentration, that is to say that the generalization capacity of the model for the main output and the secondary outputs is comparable.
- the input data 210 used for training the model are representative of the standard samples used for defining the model: several spectra can represent the same standard. This data can be preprocessed to reduce experimental fluctuations in spectral intensity values. For example, in one embodiment of the invention, each spectrum can be normalized by the intensity of a given line or band, or preprocessed by an SNV type method (“standard normal variate” in English) or any other pretreatment method. In another embodiment, for each standard, the average spectral intensity value of the spectrum at a given wavelength can be used to determine outliers. For example, before defining the model, we can reject spectra whose intensity value at a given wavelength is outside an arbitrary interval (for example outside percentiles 5 and 95, or 1 and 99 ).
- the interval depends on the measurement conditions and the number of realizations for each standard: if this number is high, a larger interval can be chosen (1 and 99 percentiles, for example). In another embodiment, these two types of pretreatment can be combined.
- the real data 240 used during the model use phase must be preprocessed in the same way as the input data, but it is possible to keep the outliers in the unknown real data 240. Indeed, if the model trained has a strong generalization ability, outliers will be properly handled during the inference phase. The goal is to learn the prediction task from a reliable representation of the standard: a good model must be able to extrapolate the necessary information in case the samples have defects.
- the evaluation of the model using evaluation data 222 can be done by directly using each spectrum as input data. We can then calculate the average of the predictions and associate a deviation with the predictions to better evaluate the performance of the algorithm (and train the algorithm on more complex cases where noise can produce significant differences between the spectra, even if coming from the same standard).
- the primary outputs 251 are composed of the concentrations of the chemical species analyzed.
- Outputs secondary 252 of the model contain the intensities of emission (or absorption) lines (or molecular bands) associated with the same chemical species in the case of spectral data, or by the intensity measured in a spectral band in the case of data multi- or hyperspectral.
- the intensities of the spectral lines or bands are used for the prediction of the concentration.
- the quantitative analysis method comprises a step of calculating a reliability indicator of the primary outputs of the model from the secondary outputs of the model.
- the reliability indicator corresponds to a measurement of the difference between the intensity predicted at the output of the model and the real intensity directly measured on the input data 240. For example, we can calculate the Student's t value.
- the invention thus proposes a method based on a multi-output automatic learning algorithm which allows the use of information explicitly contained in the experimental data (the intensities of the emission lines of the spectrum), therefore verifiable (by direct measurement on the spectrum), to obtain an interpretation of the concentration predictions in terms of probability.
- the invention thus proposes a method for quantitatively characterizing the confidence of these predictions using the error on the secondary outputs.
- these predictions must have elements in common with verifiable elements in the real data 240 to allow comparison. Rejection (or non-rejection) of the results can then be done using statistical tests.
- the secondary predictions or the result of the statistical tests can be coupled to a classification algorithm for the determination of anomalies in the samples analyzes. That is to say, the deviation between the predicted values and the actual values of the secondary outputs can be used as input to a classification algorithm trained to recognize normal values and anomalies of manufacturing, composition or measurement.
- a classification algorithm trained to recognize normal values and anomalies of manufacturing, composition or measurement.
- IR W -> A c [R L and /: A -> IR M , N is the dimension of the spectra, L is the dimension of the common representation and M is the number of predictions output from the model, such that M > 2 , using the same set of parameters for the model definition.
- the model first calculates a new representation of the input data (the spectra), which can then be used to generate the predictions.
- the coefficients ⁇ J £ represent hyperparameters of the model 231.
- model 231 is implemented using one or more neural networks.
- An example architecture of the Model 231 is shown in Figure 3.
- the neural networks constituting model 231 are convolutional neural networks.
- a convolutional neural network is implemented via a set of convolutions by vectors that extend towards the wavelength channels of the spectra.
- the input data is scanned using multiple convolution filters to obtain both primary predictions 251 and secondary predictions 252 from a multi-output model.
- neural networks can be deep, with several hidden layers, connect convolution and activation operations, with their trainable weights.
- the contribution of this type of algorithm is the direct use of the information contained in the spectra: by using a convolution operation, we can keep the relationships between neighboring wavelength channels and use the physical information contained in the profiles of spectral lines (or spectral bands) such as the emission or absorption intensity and the width of the profile.
- spectral lines or spectral bands
- a sufficient condition of the invention is the ability to form multiple output architectures, without specific restrictions on the type of neural network.
- a multitask network is shown (as described in reference [20]).
- the difference between a multi-task network and a multi-output network is related to the processing of data representations in the hidden layers: a multi-output network uses a common representation without subsequent transformations, but a multi-task network allows you to use a common representation and then use branches specialized in specific tasks. In this sense, a multi-task network is a multi-output network, but the reverse is not true.
- the use of a multitask architecture makes it possible to resolve many problems linked to overfitting of training data and to limit the impact of outliers in predictions (as mentioned for example in reference [18]).
- the algorithms used can be based on multitask deep convolutional neural networks. This gives the models the ability to directly analyze the spectra, formulating a new common representation of the data, and predict with great precision either the concentrations of the species of interest or, for example, the intensity of spectral emissions associated with the species. . These predictions can then be interpreted by probability models by studying the difference between the predicted values and the actual values.
- the input data 310 of the model have already been described: these are spectra acquired by a spectroscopy technique, possibly preprocessed.
- the first block of model 320 is represented by a neural network comprising one-dimensional convolutional layers.
- the task of this first block is to learn a common representation of the spectra using the same set of weights and parameters (“hard parameter sharing” in English).
- hard parameter sharing in English.
- Several examples of realizing multitask neural networks are given in reference [19]. It aims to transform the input data into a representation capable of providing predictions for all tasks at the same time.
- the outputs/tasks of the model 341 - 343, including the concentrations of the species analyzed and the values used for determining the confidence, are calculated using other specific convolutional networks 331 - 333: each network makes it possible to calculate subsequent transformations from the common representation and obtain a specific result.
- the The model calculates the secondary outputs (the intensities of the lines present in the spectra) starting from this common representation and then processing it in specific branches. So the model does not simply "read" the intensity in the spectra, but is forced to first find a suitable representation of the input data and then process this information.
- This type of network provides a significant improvement compared to traditional networks.
- the convolutional neural networks 331 -333 take as input a vector of common characteristics extracted from the input data via the first neural network 320, and produce as output a prediction in the form of a scalar. For this, each convolutional neural network 331 -333 has a final layer fully connected to an output.
- a first neural network 320 formed by 3 convolutional layers with respectively: i. 64, 40 and 16 filters, ii. kernel size: 5, 3 and 2 wavelength channels, iii. steps: 3, 2 and 1 wavelength channels,
- the activation function after each hidden layer is for example of the “LeakyReLU” type (with a slope of 0.03).
- the learning rate is for example set to 10 -3 .
- Figure 4 represents a table which shows the results of the calibration of the concentration of iron (Fe) in a nickel matrix (Ni) and a zirconium matrix (Zr). 25 laser shots were recorded for each sample and each spectrum is formed by 68,000 wavelength channels.
- the secondary outputs are the Fe emission line intensity values at 373.49 nm, 358.12 nm, 373.71 nm, 374.56 nm, 382.04 nm, 385.99 nm, 404.58 nm and 438.35 nm.
- the training consists only of spectra of the Ni standards (results are given on an independent test set).
- the multivariate architecture based on multitask learning, is capable of providing more precise predictions than the classic univariate method (UVA) on matrices of the same nature as those used for learning .
- the table also shows the generalization capabilities on a matrix of Zr, of a different nature from the training samples: the neural network is capable of considerably reducing the error in the extrapolation regime.
- the multitasking architecture makes it possible to measure the reliability of the predictions: the relative errors of the emission line intensities (and t values) in the Zr samples are indeed very high. In this case, the Zr matrix was recognized as anomalous with respect to the training distribution.
- the invention makes it possible to quantify the confidence in the predicted concentration using probability models on the basis of a set of predictions of a multi-output model.
- the model can predict both the concentration of species and other quantities (for example, the intensity of the emission lines of a chemical species). This makes it possible to have available both unknown quantities (e.g. species concentrations) and secondary quantities which can be directly verified in the spectra, and which are used to determine the confidence level of the quantitative analysis.
- the invention makes it possible to reduce uncertainties by using efficient algorithms capable of directly and completely automatically processing the spectral signatures of the species contained in the entire spectra.
- the steps of the invention can be implemented as a computer program comprising instructions for its execution.
- the computer program may be recorded on a processor-readable recording medium.
- the reference to a computer program which, when executed, performs any of the functions described above, is not limited to an application program executing on a single host computer. Rather, the terms computer program and software are used here in a general sense to refer to any type of computer code (e.g., application software, firmware, microcode, or any other form of computer instruction) which can be used to program one or more processors to implement aspects of the techniques described herein.
- the IT means or resources can in particular be distributed (“Cloud computing"), possibly using peer-to-peer technologies.
- the software code may be executed on any suitable processor (e.g., microprocessor) or processor core or set of processors, whether provided in a single computing device or distributed across multiple computing devices (e.g. example as possibly accessible in the environment of the device).
- the executable code of each program allowing the programmable device to implement the processes according to the invention can be stored, for example, in the hard disk or in read-only memory.
- the program(s) can be loaded into one of the storage means of the device before being executed.
- the central unit can control and direct the execution of the instructions or portions of software code of the program(s) according to the invention, instructions which are stored in the hard disk or in the read-only memory or in the other aforementioned storage elements.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Biochemistry (AREA)
- Plasma & Fusion (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Optics & Photonics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR2206060A FR3136856B1 (fr) | 2022-06-21 | 2022-06-21 | Méthode de validation des prédictions d’un modèle supervisé d’analyse quantitative multivariée de données spectrales |
| PCT/EP2023/063875 WO2023247127A1 (fr) | 2022-06-21 | 2023-05-24 | Methode de validation des predictions d'un modele supervise d'analyse quantitative multivariee de donnees spectrales |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4544462A1 true EP4544462A1 (fr) | 2025-04-30 |
Family
ID=85122664
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP23730364.9A Pending EP4544462A1 (fr) | 2022-06-21 | 2023-05-24 | Methode de validation des predictions d'un modele supervise d'analyse quantitative multivariee de donnees spectrales |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US20250378343A1 (fr) |
| EP (1) | EP4544462A1 (fr) |
| CN (1) | CN119404200A (fr) |
| FR (1) | FR3136856B1 (fr) |
| WO (1) | WO2023247127A1 (fr) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20250045619A1 (en) * | 2023-07-31 | 2025-02-06 | Maplebear Inc. | Conditional Loss Function for Training a Multitask Machine Learning Model |
| CN117929356B (zh) * | 2024-03-21 | 2024-07-02 | 沈阳尖科智能测控技术合伙企业(有限合伙) | 一种基于高斯过程回归的libs定量分析方法 |
| CN119044153B (zh) * | 2024-08-23 | 2025-09-05 | 浙江大学 | 一种土壤矿质营养元素的多变量检测方法、系统和设备 |
| CN119901727B (zh) * | 2025-01-15 | 2025-09-09 | 中国海洋大学 | 面向深海稀土元素libs定量探测的光谱筛选方法、设备及存储介质 |
| CN119741998A (zh) * | 2025-03-04 | 2025-04-01 | 中国地质大学(武汉) | 一种分析铀多金属矿元素的激光光谱数据处理方法 |
| CN120293874B (zh) * | 2025-06-12 | 2025-09-23 | 安徽新宇环保科技股份有限公司 | 一种基于全光谱分析的水质在线监测系统 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113196053A (zh) * | 2018-12-20 | 2021-07-30 | 佳能株式会社 | 信息处理装置、信息处理装置的控制方法及程序 |
-
2022
- 2022-06-21 FR FR2206060A patent/FR3136856B1/fr active Active
-
2023
- 2023-05-24 CN CN202380048593.6A patent/CN119404200A/zh active Pending
- 2023-05-24 WO PCT/EP2023/063875 patent/WO2023247127A1/fr not_active Ceased
- 2023-05-24 EP EP23730364.9A patent/EP4544462A1/fr active Pending
- 2023-05-24 US US18/877,218 patent/US20250378343A1/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| WO2023247127A1 (fr) | 2023-12-28 |
| CN119404200A (zh) | 2025-02-07 |
| FR3136856B1 (fr) | 2026-03-13 |
| FR3136856A1 (fr) | 2023-12-22 |
| US20250378343A1 (en) | 2025-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2023247127A1 (fr) | Methode de validation des predictions d'un modele supervise d'analyse quantitative multivariee de donnees spectrales | |
| CN109493287B (zh) | 一种基于深度学习的定量光谱数据分析处理方法 | |
| EP3846087A1 (fr) | Procede et systeme de selection d'un modele d'apprentissage au sein d'une pluralite de modeles d'apprentissage | |
| JPWO2019039313A1 (ja) | スペクトルの高速分類を利用したリアルタイムスペクトル解析 | |
| WO2010076469A1 (fr) | Identification de defaillances dans un moteur d'aeronef | |
| Cioccia et al. | Probabilistic-based identification of gunshot residues (GSR) using Laser-Induced Breakdown Spectroscopy (LIBS) and Support Vector Machine (SVM) algorithm | |
| EP3846047A1 (fr) | Procédé et système d'identification de variables pertinentes | |
| CN120009209B (zh) | 一种基于多波长分光光度的食品污染检测方法及系统 | |
| Chen et al. | A decision tree network with semi-supervised entropy learning strategy for spectroscopy aided detection of blood hemoglobin | |
| Zhang et al. | Raman spectroscopy combined with a novel explainable MSDCNN networks for avocado oil adulteration detection | |
| EP1007962B1 (fr) | Appareil et methode de classification utilisant une combinaison de methodes statistiques et de reseaux neuronaux | |
| FR3146004A1 (fr) | Procédé de contrôle non-destructif de pièces en matériau composite | |
| EP1845479B1 (fr) | L'analyse comparative d'un échantillon par rapport à une base de données | |
| Chia et al. | Neural network and extreme gradient boosting in near infrared spectroscopy | |
| WO2023247128A1 (fr) | Methode de generation de donnees spectrales synthetiques | |
| Lee et al. | Avocado ripeness classification using handheld Raman spectroscopy: addressing data imbalance with machine learning and resampling techniques | |
| EP4599242A1 (fr) | Procédé d'optimisation bi-niveau de la localisation de capteurs de détection de défaut(s) dans une structure par tomographie à ondes élastiques guidées | |
| Paiva-Peredo et al. | Optimizing 1D CNN models for accurate° Brix prediction | |
| FR2978577A1 (fr) | Dispositif et procede de diagnostic d'un processus industriel evolutif | |
| RU2844698C1 (ru) | Способ идентификации и скрининга качества лекарственных средств на основе рамановских спектров с применением искусственного интеллекта | |
| Sun et al. | NIRS regression model for dry matter content estimation in olive fruit with XGBoost pre-treatment method | |
| CN120275327B (zh) | 基于多种光谱融合的变压器油中多组分气体检测方法 | |
| FR3118244A1 (fr) | Procédé et dispositif de diagnostic d’anomalies | |
| WO2020002573A1 (fr) | Procede de modelisation pour le controle des resultats fournis par un reseau de neurones artificiels et autres procedes associes | |
| EP4660883A1 (fr) | Procédé d analyse et d ajustement d'un réseau de neurones artificiels |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20241210 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| GRAP | Despatch of communication of intention to grant a patent |
Free format text: ORIGINAL CODE: EPIDOSNIGR1 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: GRANT OF PATENT IS INTENDED |
|
| INTG | Intention to grant announced |
Effective date: 20260626 |