System and method for analyzing spectral data using artificial intelligence
A system provides for an ability to automatically identify one or more chemical components of a sample, especially a biological fluid sample, based on analysis of spectral data by at least one artificial intelligence module. The artificial intelligence module is able to be trained on a plurality of spectral data samples having known concentrations of individual chemicals and elements. The system is preferably used for spectral techniques such as UV/vis and near-infrared (NIR) spectrophotometry, but is also able to be used for other forms of spectroscopic techniques, including mass spectrometry, infrared (IR) spectrometry, X-ray spectroscopy, Raman spectrometry, or nuclear magnetic resonance (NMR) spectrometry.
1 . An artificial intelligence (AI)-based system for automatically identifying molecules in a fluid sample, comprising:
one or more spectrometers or chemical analysis devices;
at least one reactor;
one or more servers configured to receive multi-dimensional training data from the one or more spectrometers or chemical analysis devices; and
an AI module on the one or more servers;
wherein the one or more servers are configured to train the AI module using the multi-dimensional training data to automatically develop characteristic profiles for a plurality of known molecules or sets of molecules;
wherein the multi-dimensional training data corresponds with samples of the plurality of known molecules or sets of molecules;
wherein the at least one reactor is operable to ionize a fluid sample via a non-thermal plasma source;
wherein the one or more servers receive experimental data for the ionized fluid sample from a testing spectrometer;
wherein the AI module is operable to compare the experimental data to the multi-dimensional training data;
wherein the AI module is operable to analyze the experimental data, wherein the AI module automatically correlates spectrographic data from the experimental data with the characteristic profiles to identify molecules in the fluid sample; and
wherein the AI module automatically generates a report based on comparison of the experimental data to the characteristic profiles for the plurality of molecules or sets of molecules.
2 . The system of claim 1 , wherein the one or more servers include cloud servers, and wherein the multi-dimensional training data received from the one or more spectrometers or chemical analysis devices is combined with public datasets and/or additional third party datasets.
3 . The system of claim 1 , wherein the multi-dimensional training data further includes synthetic data.
4 . The system of claim 1 , wherein the multi-dimensional training data includes data related to multiple different modalities of spectroscopic or other chemical analysis.
5 . The system of claim 1 , wherein the multi-dimensional training data includes data for at least one baseline of at least one sample.
6 . The system of claim 1 , wherein the AI module includes one or more machine learning (ML) models, deep learning (DL) models, neural networks (NNs), support vector machines (SVMs), and/or transformers.
7 . The system of claim 1 , wherein the one or more spectrometers includes one or more ultraviolet-visual (UV-VIS) spectrometers, one or more Raman spectrometers, one or more infrared (IR) spectrometers, one or more mass spectrometers, one or more nuclear magnetic resonance (NMR) spectrometers, one or more X-ray spectrometers, one or more optical emission spectrometers (OES), one or more atomic emission spectrometers (AES), and/or one or more Mossbauer spectrometers.
8 . The system of claim 1 , wherein the report includes indications of present identified molecules, indications of concentrations of the present identified molecules, and certainty values associated with each of the present identified molecules.
9 . An artificial intelligence (AI)-based method for automatically identifying molecules in a fluid sample, comprising:
receiving multi-dimensional training data for one or more servers from one or more spectrometers or chemical analysis devices;
training an AI module using the multi-dimensional training data to automatically develop characteristic profiles for a plurality of known molecules or sets of molecules via the one or more servers;
ionizing a fluid sample via a non-thermal plasma source using at least one reactor;
receiving experimental data for the ionized fluid sample from a testing spectrometer via the one or more servers;
automatically analyzing the experimental data by automatically correlating spectrographic data from the experimental data with the characteristic profiles to identify biological molecules in the fluid sample using the AI module; and
automatically generating a report with indications of present identified molecules based on comparison of the experimental data to the characteristic profiles for the plurality of molecules or sets of molecules using the AI module;
wherein the report includes indications of present identified molecules, indications of concentrations of the present identified molecules, and certainty values associated with each of the present identified molecules.
10 . The method of claim 9 , wherein the one or more servers include cloud servers, and further comprising combining the multi-dimensional training data received from the one or more spectrometers or chemical analysis devices with public datasets and/or additional third party datasets.
11 . The method of claim 9 , wherein the multi-dimensional training data includes data for at least one baseline of at least one sample.
12 . The method of claim 9 , wherein the multi-dimensional training data includes data related to multiple different modalities of spectroscopic or other chemical analysis.
13 . The method of claim 9 , wherein the multi-dimensional training data corresponds with samples of known molecules and/or compositions.
14 . The method of claim 9 , wherein the AI module includes one or more machine learning (ML) models, deep learning (DL) models, neural networks (NNs), support vector machines (SVMs), and/or transformers.
15 . The method of claim 9 , wherein the one or more spectrometers includes one or more ultraviolet-visual (UV-VIS) spectrometers, one or more Raman spectrometers, one or more infrared (IR) spectrometers, one or more mass spectrometers, one or more nuclear magnetic resonance (NMR) spectrometers, one or more X-ray spectrometers, one or more optical emission spectrometers (OES), one or more atomic emission spectrometers (AES), and/or one or more Mossbauer spectrometers.
16 . The method of claim 9 , wherein the multi-dimensional training data further includes synthetic data.
17 . An artificial intelligence (AI)-based system for automatically identifying molecules in a fluid sample, comprising:
one or more spectrometers or chemical analysis devices;
at least one reactor;
one or more cloud servers configured to receive multi-dimensional training data from the one or more spectrometers or chemical analysis devices; and
an AI module on the one or more cloud servers;
wherein the one or more cloud servers are configured to train the AI module using the multi-dimensional training data to automatically develop characteristic profiles for a plurality of known molecules or sets of molecules;
wherein the multi-dimensional training data received from the one or more spectrometers or chemical analysis devices is combined with public datasets and/or additional third party datasets;
wherein the at least one reactor is operable to ionize a fluid sample via a non-thermal plasma source;
wherein the one or more cloud servers receive experimental data for the ionized fluid sample from a testing spectrometer;
wherein the AI module is operable to compare the experimental data with the multi-dimensional training data;
wherein the AI module is operable to analyze the experimental data, and wherein the AI module automatically correlates spectrographic data from the experimental data with the characteristic profiles to identify molecules in the fluid sample; and
wherein the AI module automatically generates a report based on comparison of the experimental data to the multi-dimensional training data for the plurality of molecules or sets of molecules.
18 . The system of claim 17 , wherein the multi-dimensional training data corresponds with samples of known molecules and/or compositions.
19 . The system of claim 17 , wherein the multi-dimensional training data includes data related to multiple different modalities of spectroscopic or other chemical analysis.
20 . The system of claim 17 , wherein the report includes indications of present identified molecules, indications of concentrations of the present identified molecules, and certainty values associated with each of the present identified molecules.