Molecular ground state electron density detection method and system and cloud computing platform

Abstract

The invention provides a molecular ground state electron density detection method, a molecular ground state electron density detection system and a cloud computing platform, which can be applied to the technical field of quantum chemistry. The molecular ground state electron density detection method comprises the following steps: acquiring first molecular data of a to-be-detected molecule; wherein the first molecular data comprises first molecular structure data and first electronic structure data; substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the to-be-detected molecule; inputting the first molecular data into a kinetic energy functional model to obtain a first electronic kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training; and performing iterative optimization on the initial electron density by taking minimization of the first electron kinetic energy functional as an optimization target to obtain the ground state electron density of the to-be-detected molecule. According to the method, the detection efficiency of the molecular ground state electron density is effectively improved through combination of deep learning and an optimization algorithm.

Description

Technical Field

[0001] The present application relates to the field of quantum chemistry technology, and in particular to a method, system and cloud computing platform for detecting molecular ground state electron density. Background Art

[0002] The ground state electron density refers to the distribution density of electrons in space in a system. In quantum chemistry, the ground state electron density describes the distribution of electrons in the lowest energy state in a molecule. The ground state electron density is of decisive significance for understanding the properties and behaviors of molecules because it directly determines the physical and chemical properties of molecules.

[0003] For the detection of molecular ground state electron density, the existing detection method is generally calculated through density functional theory (DFT). DFT efficiently calculates the electronic structure by simplifying the multi-body problem into a single-body problem. In DFT, the ground state electron density is the main physical quantity that determines all the properties of the system. By solving the ground state electron density, key information such as the system's ground state energy and Hamiltonian can be obtained.

[0004] However, existing methods for detecting molecular ground-state electron density still require huge computing resources and time when targeting large molecular systems, which limits their feasibility in practical applications. Summary of the invention

[0005] The present application provides a method, system and cloud computing platform for detecting the ground state electron density of a molecule, so as to solve the technical problem of low detection efficiency of the existing method for detecting the ground state electron density of a molecule.

[0006] According to the first aspect disclosed in the present application, the present application provides a method for detecting the ground state electron density of a molecule, comprising:

[0007] Acquire first molecular data of the molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data;

[0008] Substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected;

[0009] Inputting the first molecular data into a kinetic energy functional model to obtain a first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training;

[0010] Taking minimizing the first electron kinetic energy functional as an optimization goal, the initial electron density is iteratively optimized to obtain the ground state electron density of the molecule to be detected.

[0011] In a feasible implementation manner, the method further includes:

[0012] Collecting second molecular data of a plurality of sample molecules to construct a training set; wherein the second molecular data is annotated with a true label, and the true label includes a second electron kinetic energy functional of the sample molecule;

[0013] The training set is input into a pre-built deep learning model for iterative training to obtain the kinetic energy functional model; wherein the deep learning model includes a multi-layer perceptron, a convolutional neural network, a long short-term memory network and a Transformer neural network.

[0014] In a feasible implementation, the training set is input into a pre-built deep learning model for iterative training to obtain the kinetic energy functional model, including:

[0015] Inputting the second molecule data in the training set into the deep learning model to obtain an output result;

[0016] Calculating the loss value between the output result and the true label based on a preset loss function;

[0017] The parameters of the deep learning model are updated according to the loss value until the loss function converges or reaches a preset number of training iterations, thereby obtaining the kinetic energy functional model.

[0018] In a feasible implementation, collecting second molecule data of a plurality of sample molecules and constructing a training set includes:

[0019] Collecting second molecular data of a plurality of sample molecules; wherein the second molecular data includes second molecular structure data and second electronic structure data;

[0020] Performing data cleaning, data standardization, and real labeling on the second molecular data to obtain training data;

[0021] Based on the training data, a training set is constructed.

[0022] In a feasible implementation, the density initialization algorithm is a Hirshfeld algorithm or a Voronoi degeneration algorithm.

[0023] According to the second aspect disclosed in the present application, the present application provides a molecular ground state electron density detection system, comprising:

[0024] A data acquisition module, used to acquire first molecular data of the molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data;

[0025] A density initialization module, used for substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected;

[0026] A kinetic energy functional determination module, used to input the first molecular data into a kinetic energy functional model to obtain a first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training;

[0027] The density optimization module is used to iteratively optimize the initial electron density by minimizing the first electron kinetic energy functional to obtain the ground state electron density of the molecule to be detected.

[0028] According to the third aspect disclosed in the present application, the present application provides a cloud computing platform, including the molecular ground state electron density detection system as described in the second aspect.

[0029] In a feasible implementation manner, the cloud computing platform further includes a resource allocation system, and the resource allocation system is used to:

[0030] Based on the first molecular data of the molecule to be detected, evaluating the amount of computing resources required by the molecular ground state electron density detection system to perform ground state electron density detection on the molecule to be detected;

[0031] Based on the amount of computing resources, cloud computing resources are scheduled and allocated; wherein the cloud computing resources are used to deploy the molecular ground state electron density detection system.

[0032] According to a fourth aspect disclosed in the present application, the present application provides an electronic device, comprising a processor, and a memory communicatively connected to the processor;

[0033] The memory stores computer-executable instructions;

[0034] The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of the first aspects.

[0035] According to the fifth aspect disclosed in the present application, the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed, they are used to implement any method in the first aspect.

[0036] According to a sixth aspect disclosed in the present application, the present application provides a computer program product, including a computer program, and when the computer program is executed, it is used to implement any method in the first aspect.

[0037] Compared with the prior art, this application has the following beneficial effects:

[0038] The present application provides a method, system and cloud computing platform for detecting the ground state electron density of a molecule. The deep learning model obtained through training is used to approximate complex electron kinetic energy functionals, giving full play to its advantages in pattern recognition and complex data processing, which can significantly improve computing efficiency and reduce the demand for computing resources, and improve the automation and intelligence level of computing. The electron kinetic energy functional output by the model is then used to optimize the target electron density, and the ground state electron density of the molecule is accurately detected. Using deep learning and optimization algorithms, the detection efficiency of the ground state electron density of the molecule and the accuracy of the detection are effectively improved. At the same time, through the generalization ability of the deep learning model, the universality of the detection of the ground state electron density of the molecule can also be improved, so that it can be more widely used in different types of molecular systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0040] Figure 1 A schematic diagram of a method for detecting the ground state electron density of a molecule provided in an embodiment of the present application;

[0041] Figure 2 A flowchart of a model training method provided in an embodiment of the present application;

[0042] Figure 3 A flowchart of another model training method provided in an embodiment of the present application;

[0043] Figure 4 A schematic diagram of the structure of a molecular ground state electron density detection system provided in an embodiment of the present application;

[0044] Figure 5 A schematic diagram of the structure of a cloud computing platform provided in an embodiment of the present application;

[0045] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0046] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0047] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0048] The ground state electron density refers to the distribution density of electrons in space in a system. In quantum chemistry, the ground state electron density describes the distribution of electrons in the lowest energy state in a molecule. The ground state electron density is of decisive significance for understanding the properties and behaviors of molecules because it directly determines the physical and chemical properties of molecules.

[0049] Therefore, the ground state electron density is an important parameter for studying the properties and behaviors of molecules. By understanding the density distribution of electrons, key information such as the internal structure of molecules, the formation and breaking of chemical bonds, etc. can be revealed. The ground state electron density is also crucial for studying chemical reactions. It determines the ability of molecules to react with other substances. Therefore, it is an important tool for studying chemical reactions such as electron transfer reactions, photolysis reactions, and ion exchange reactions. For example, in the field of materials science, the prediction of molecular properties through the ground state electron density can optimize lead compounds in chemical synthesis, predict reaction mechanisms, and design efficient chemical materials. At the same time, in the field of material design, it can predict material properties and guide the discovery and performance optimization of new materials, especially in the development of high-performance energy materials. Through in-depth analysis of electronic structure characteristics, it can provide strong scientific guidance for experimental design and process optimization, thereby guiding the development and design of new materials.

[0050] For the detection of the ground state electron density of molecules, the existing detection method is generally calculated through density functional theory (DFT). DFT efficiently calculates the electronic structure by simplifying the multi-body problem into a single body problem. In DFT, the ground state electron density is the main physical quantity that determines all the properties of the system. By solving the ground state electron density, key information such as the ground state energy and Hamiltonian of the system can be obtained.

[0051] However, existing methods for detecting molecular ground-state electron density still require huge computing resources and time when targeting large molecular systems, which limits their feasibility in practical applications.

[0052] In response to the above technical problems, this application proposes a method for detecting the ground state electron density of a molecule, which effectively improves the detection efficiency of the ground state electron density of a molecule by combining deep learning with an optimization algorithm.

[0053] The technical solution of the molecular ground state electron density detection method provided by the present application is described in detail below through specific embodiments. It should be noted that the following embodiments can exist alone or in combination with each other, and the same or similar contents may not be repeated in different embodiments.

[0054] Figure 1 A schematic diagram of a method for detecting the ground state electron density of a molecule provided in an embodiment of the present application, see Figure 1 In some embodiments, the process of the molecular ground state electron density detection method includes the following steps:

[0055] S101, obtaining first molecular data of a molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data.

[0056] Among them, the molecular data of the molecules to be detected include molecular structure data and electronic structure data, specifically including the bond length, bond angle and dihedral angle between atoms of the molecules to be detected, as well as the distribution of electron clouds, molecular orbitals and electron density and other data information.

[0057] Specifically, the source of molecular data can be experimental measurements, previous calculation results or theoretical models, such as X-ray crystallography data, nuclear magnetic resonance (NMR) spectroscopy data or quantum chemical calculation results.

[0058] S102, substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected.

[0059] Among them, the density initialization algorithm is used to obtain the initial density of the molecules to be detected, providing an initial state for the subsequent optimization of the electron density.

[0060] Preferably, the density initialization algorithm is a Hirshfeld algorithm or a Voronoi degeneration algorithm.

[0061] Among them, the Hirshfeld algorithm is a molecule-based atomic decomposition method that can calculate the approximate value of the electron density from the molecular orbital wave function; the Voronoi degeneration method estimates the electron density through the geometric structure of the molecule, dividing the space around the molecule into Voronoi polyhedrons, and the volume of the polyhedron around each atom is related to the electron density of the atom.

[0062] S103, inputting the first molecule data into a kinetic energy functional model to obtain a first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training.

[0063] Among them, the kinetic energy functional model can learn the complex relationship of the electron kinetic energy functional and is used to approximate the electron kinetic energy functional of the molecule.

[0064] Specifically, the electron kinetic energy functional of a molecule is an important concept in quantum chemistry. It is based on density functional theory and is used to describe the kinetic energy of electrons in molecules. The electron kinetic energy functional of a molecule is a function that maps the electron density (or other related variables) to the kinetic energy value of the electron. In density functional theory, the kinetic energy of the electron is part of the total energy, so the electron kinetic energy functional is one of the key elements in constructing the total energy functional and plays an important role in molecular structure calculations. By optimizing the total energy functional, structural information such as the stable configuration, bond length, and bond angle of the molecule can be obtained.

[0065] S104, taking minimization of the first electron kinetic energy functional as the optimization goal, iteratively optimizing the initial electron density to obtain the ground state electron density of the molecule to be detected.

[0066] The optimization objective is to minimize the first electron kinetic energy functional, and the initial electron density is iteratively optimized until the initial electron density converges to the ground state to obtain the ground state electron density. In this process, the first electron kinetic energy functional output by the kinetic energy functional model is used as the optimization objective to guide the adjustment of the electron density to minimize the total energy of the system.

[0067] Specifically, the optimization algorithm for optimizing the initial electron density can use the gradient descent method or the Newton method. Among them, the gradient descent method is an iterative optimization algorithm used to solve the minimum or maximum value of a function. It calculates the gradient of the objective function to the parameter (that is, the rate of change of the function at the current parameter value), and then updates the parameters in the opposite direction of the gradient, so that the value of the objective function gradually decreases. Newton's method, also known as the Newton-Raphson method, is an iterative optimization algorithm for solving the root of an equation or the minimum value of a function. It uses the second-order derivative information of the function (Hessian matrix) to approximate the local properties of the function, and can converge to the optimal solution faster.

[0068] Preferably, the ground state electron density is substituted into a preset property calculation algorithm to obtain the physicochemical properties of the molecule to be detected; wherein the physicochemical properties include molecular energy, molecular force and charge distribution.

[0069] Among them, the property calculation algorithm can calculate the total energy of the molecule based on the optimized ground state electron density using quantum chemistry software packages such as Gaussian, which includes the calculation of electron kinetic energy, electron-nuclear attraction energy, electron-electron repulsion energy and nucleus-nucleus repulsion energy; calculate the physical and chemical properties of the molecule such as force and charge distribution to provide data support for further chemical analysis. The force of the molecule can be obtained by calculating the net force on each atom in the molecule, and the charge distribution can be obtained by the following methods: Mulliken charge analysis, a method of assigning electron density in a molecule to each atom; ESP (Electrostatic Potential) charge fitting: obtaining atomic charge by fitting the electrostatic potential of the molecule; RESP (Restrained ElectroStatic Potential) charge: a charge distribution method that takes into account changes in molecular conformation, suitable for the simulation of flexible molecules. These properties can be calculated by software such as Multiwfn, among which RESP charge calculation is a commonly used method, which fits the electrostatic potential of the molecule by the least squares method and imposes constraints to ensure the rationality of the charge distribution.

[0070] In addition, by conducting in-depth analysis of the calculation results, key electronic structure features are identified, and how these features affect the chemical reactivity and stability of molecules is explored to provide guidance for chemical synthesis and material design. Key electronic structure features include but are not limited to: electron density distribution, analyzing the distribution of electrons in molecules and identifying dense and sparse areas of electron clouds; chemical bond characteristics, analyzing covalent bonds, ionic bonds, and metallic bonds in molecules;

[0071] Preferably, the calculation results are compared with traditional methods or experimental data to verify the accuracy and reliability of the electronic structure calculation method of the present invention.

[0072] In this embodiment, the complex electron kinetic energy functional is approximated by the deep learning model obtained through training, giving full play to its advantages in pattern recognition and complex data processing, which can significantly improve the computing efficiency and reduce the demand for computing resources, and improve the automation and intelligence level of the calculation. The electron kinetic energy functional output by the model is then optimized as the optimization target electron density to accurately detect the ground state electron density of the molecule. Using deep learning and optimization algorithms, the detection efficiency of the molecular ground state electron density and the accuracy of the detection are effectively improved. At the same time, through the generalization ability of the deep learning model, the universality of the molecular ground state electron density detection can also be improved, so that it can be more widely used in different types of molecular systems. In addition, the present application can also pay special attention to the use of deep learning models to solve specific chemical problems that are difficult to accurately describe by traditional methods, such as transition state search, reaction kinetics simulation, etc.

[0073] exist Figure 1In the molecular ground state electron density detection method shown in the figure, it is necessary to train the kinetic energy functional model. Figure 2 , further introduces the content of training kinetic energy functional model in the above-mentioned molecular ground state electron density detection method.

[0074] Figure 2 A schematic diagram of a kinetic energy functional model training method provided in an embodiment of the present application, see Figure 2 In some embodiments, the process of the kinetic energy functional model training method includes the following steps:

[0075] S201, collecting second molecular data of multiple sample molecules to construct a training set; wherein the second molecular data is annotated with a true label, and the true label includes a second electron kinetic energy functional of the sample molecule.

[0076] Among them, the molecular data of the sample molecules include molecular structure data and electronic structure data, which specifically include the bond length, bond angle and dihedral angle between atoms of the sample molecules, as well as the distribution of electron clouds, molecular orbitals and electron density and other data information.

[0077] In order to guide model training, the molecular data of the sample molecules also need to be annotated with true labels, which are used as reference points when training deep learning models. The model gradually adjusts its internal parameters (such as weights and biases) by learning how to predict these labels so that its predictions are as close to these true labels as possible.

[0078] S202, inputting the training set into a pre-built deep learning model for iterative training to obtain a kinetic functional model; wherein the deep learning model includes a multi-layer perceptron, a convolutional neural network, a long short-term memory network, and a Transformer neural network.

[0079] Among them, the deep learning model extracts key features of the molecular data in the training set, such as electron density, interatomic distance, etc., as the input of the model, and uses these features for the training of the deep learning model to obtain a kinetic energy functional model for characterizing the electron kinetic energy functional.

[0080] Specifically, the architecture of deep learning models includes multi-layer perceptron, convolutional neural network, long short-term memory network and transformer neural network. Among them, multi-layer perceptron (MLP) is suitable for processing the mapping of molecular structure to electronic properties. Convolutional neural network (CNN) is suitable for extracting spatial features of image representation of molecules. Long short-term memory network (LSTM) is suitable for processing serialized molecular data, such as linear representation of molecules. Transformer neural network uses self-attention mechanism to process various parts of molecular structure and capture long-distance dependencies.

[0081] In this embodiment, the kinetic energy functional model is trained to improve the calculation efficiency of the electronic kinetic energy functional of the molecule to be detected.

[0082] exist Figure 2 Based on the embodiment shown, the following Figure 3 , further introduces the technical solution of the above model training method.

[0083] Figure 3 A flow chart of another model training method provided in the embodiment of the present application is shown in FIG. Figure 2 In some embodiments, the process of the model training method includes the following steps:

[0084] S301, collecting second molecular data of a plurality of sample molecules; wherein the second molecular data includes second molecular structure data and second electronic structure data.

[0085] S302, performing data cleaning processing, data standardization processing, and marking of real labels on the second molecular data to obtain training data.

[0086] Data cleaning is used to remove incomplete or erroneous data points, including verifying the integrity of the data, checking the consistency of the data points, and excluding those that do not meet the expected physical or chemical properties. In addition, the consistency of the data file format must be ensured, such as using MOL files or SDF file formats to store molecular structure data, and all electronic structure data should be stored in the same data structure so that the model can read and learn without error.

[0087] Data normalization is the process of normalizing data into a uniform format to ensure that it is suitable for training deep learning models. The uniform format usually includes: atomic coordinates, expressed as x, y, and z values ​​in three-dimensional space, and all data should be converted to the same unit system, such as angstroms ( Abbreviation: Angstrom, symbol ) as a unit of length; Molecular structure file format, it is recommended to use MOL file or SDF file format to store molecular structure data, these formats can contain graphical representations of molecules and related chemical information, such as atomic coordinates, bond types, charges, and functional group information; Electronic structure data should be represented in a compatible format, such as converting molecular orbitals into matrices of the same size to facilitate processing by deep learning models.

[0088] True labeling is a critical step in deep learning, especially in supervised learning tasks. This process involves assigning one or more true and accurate labels to each sample in the dataset. These labels represent the true characteristics or categories of the data, and the model will be used to predict these characteristics or categories after training. Specifically, the true label of the second molecule data is the true electron kinetic energy functional corresponding to the sample molecule.

[0089] S303: construct a training set based on the training data.

[0090] Among them, the processed training data is sorted and collected to form a training set.

[0091] S304, inputting the second molecule data in the training set into the deep learning model to obtain an output result.

[0092] The second molecule data in the training set is input into the deep learning model, and the output of each layer is calculated until the final prediction is obtained.

[0093] S305, calculating the loss value between the output result and the true label based on a preset loss function.

[0094] Among them, the loss function is used to calculate the difference between the predicted value and the true value.

[0095] Specifically, the loss function includes at least one of a mean square error loss function, a mean absolute error loss function, or a cross entropy loss function.

[0096] S306, updating the parameters of the deep learning model according to the loss value until the loss function converges or reaches a preset number of training iterations, thereby obtaining a kinetic functional model.

[0097] Among them, the gradient is calculated according to the loss function, and the network parameters are updated until the loss function converges or the preset number of training iterations is reached, and the kinetic energy functional model is obtained. In the process again, a large amount of molecular data is used to train the deep learning model so that it can learn the relationship between electron density and molecular energy. Through repeated forward propagation and backpropagation processes, the model gradually adjusts its parameters to minimize the prediction error.

[0098] Preferably, in order to further improve the performance of the model, methods such as cross-validation can also be used to evaluate the accuracy and generalization ability of the model, specifically including using an independent validation set to adjust the model parameters and using a test set to ultimately evaluate the performance of the model.

[0099] In this embodiment, the complex relationship of the electron kinetic energy functional is learned through a deep learning model to approximate the electron kinetic energy functional of the molecule.

[0100] Figure 4is a schematic diagram of a molecular ground state electron density detection system provided in an embodiment of the present application, see Figure 4 The molecular ground state electron density detection system includes various functional modules for implementing the aforementioned molecular ground state electron density detection method, and any functional module can be implemented by software and / or hardware.

[0101] In some embodiments, the molecular ground state electron density detection system 400 includes a data acquisition module 401, a density initialization module 402, a kinetic energy functional determination module 403 and a density optimization module 404. Among them:

[0102] The data acquisition module 401 is used to acquire first molecular data of the molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data;

[0103] The density initialization module 402 is used to substitute the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected;

[0104] The kinetic energy functional determination module 403 is used to input the first molecule data into the kinetic energy functional model to obtain the first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training;

[0105] The density optimization module 404 is used to iteratively optimize the initial electron density with minimization of the first electron kinetic energy functional as the optimization target, so as to obtain the ground state electron density of the molecule to be detected.

[0106] In some embodiments, the system 400 further includes a model training module 405, which is specifically used to:

[0107] Collecting second molecular data of multiple sample molecules to construct a training set; wherein the second molecular data is annotated with a true label, and the true label includes a second electron kinetic energy functional of the sample molecule;

[0108] The training set is input into a pre-built deep learning model for iterative training to obtain a kinetic energy functional model; the deep learning model includes a multi-layer perceptron, a convolutional neural network, a long short-term memory network, and a Transformer neural network.

[0109] In some embodiments, the model training module 405 is further used to:

[0110] Input the second molecule data in the training set into the deep learning model to obtain an output result;

[0111] Calculate the loss value between the output result and the true label based on the preset loss function;

[0112] The parameters of the deep learning model are updated according to the loss value until the loss function converges or reaches a preset number of training iterations, and the kinetic functional model is obtained.

[0113] In some embodiments, the model training module 405 is further used to:

[0114] Collecting second molecular data of a plurality of sample molecules; wherein the second molecular data includes second molecular structure data and second electronic structure data;

[0115] Perform data cleaning, data standardization, and real labeling on the second molecular data to obtain training data;

[0116] Based on the training data, a training set is constructed.

[0117] In some embodiments, the density initialization algorithm is a Hirshfeld algorithm or a Voronoi degeneration algorithm.

[0118] The molecular ground state electron density detection system 400 provided in the embodiment of the present application is used to execute the technical solution provided in the aforementioned molecular ground state electron density detection method embodiment. Its implementation principle and technical effect are similar to those in the aforementioned method embodiment and will not be repeated here.

[0119] It should be noted that it should be understood that the division of the various modules of the above device is only a division of a logical function. When actually implemented, it can be fully or partially integrated into a physical entity, or it can be physically separated. And these modules can all be implemented in the form of software called by a processing element, or they can all be implemented in the form of hardware. Some modules can also be implemented in the form of software called by a processing element, and some modules can be implemented in the form of hardware. For example, the data acquisition module 401 can be a separately established processing element, or it can be integrated in a certain chip of the above device to be implemented. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a certain processing element of the above device. The function of the above data acquisition module 401. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the hardware integrated logic circuit in the processor element or the instruction in the form of software.

[0120] Figure 5 This is a schematic diagram of the structure of a cloud computing platform provided in an embodiment of the present application. Figure 5 The cloud computing platform 500 includes the above-mentioned molecular ground state electron density detection system 501.

[0121] In this embodiment, the cloud computing platform provides extremely high flexibility and scalability, and can quickly adjust the scale of computing resources according to actual needs. Whether it is necessary to expand or reduce computing power, it can be easily achieved through the cloud platform. It can obtain and release resources at any time as needed to achieve elastic scaling to meet the ground state electron density detection needs of different molecules.

[0122] In some embodiments, the cloud computing platform 500 also includes a resource allocation system 502, which is used to: evaluate the amount of computing resources required by the molecular ground state electron density detection system 501 to perform ground state electron density detection on the molecule to be detected based on the first molecular data of the molecule to be detected; and schedule and allocate cloud computing resources based on the amount of computing resources; wherein the cloud computing resources are used to deploy the molecular ground state electron density detection system.

[0123] Among them, it is necessary to evaluate the required computing resources according to the complexity of the molecular system to be detected, which usually involves considering the size of the molecular system (such as the number of atoms), the complexity of intermolecular interactions, the required simulation time, and the expected simulation accuracy. For example, for large biological molecules or complex systems in materials science, more CPU / GPU cores and larger memory capacity may be required. In addition, the following existing algorithms can be used to assist in the evaluation: Gaussian ONIOM method, a multi-scale simulation method, which allows different levels of theoretical calculation methods to be used for different parts of the molecular system, thereby reducing the demand for computing resources while ensuring simulation accuracy; density matrix embedding theory (DMET), a quantum computing simulation method, divides the research system into spatial slices, performs lower-precision mean field calculations on the entire system, and performs higher-precision calculations on each small system after slicing, thereby reducing the number of quantum bits and circuit depth required.

[0124] By evaluating the amount of computing resources required by the molecular ground state electron density detection system to detect the ground state electron density of the molecule to be detected, computing resources can be dynamically allocated on the cloud computing platform based on the results of resource evaluation. This includes selecting the appropriate CPU / GPU type, memory size, and storage space. For example, for tasks that require a large amount of parallel computing, more GPU resources can be allocated; and for tasks that require processing large amounts of data, larger storage space is required. Cloud computing platforms usually provide flexible resource allocation strategies to ensure the efficient execution of computing tasks. In this way, the use of cloud computing resources can be optimized, cloud computing resources can be used more efficiently, and reasonable resource allocation and rapid execution of computing tasks can be achieved.

[0125] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application, see Figure 6The electronic device 600 includes a processor 601 and a memory 602 communicatively connected to the processor 601;

[0126] Memory 602 stores computer executable instructions;

[0127] The processor 601 executes the computer execution instructions stored in the memory 602 to implement the technical solution of the aforementioned molecular ground state electron density detection method.

[0128] In the above-mentioned electronic device 600, the memory 602 and the processor 601 are electrically connected directly or indirectly to realize data transmission or interaction. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines, such as through a bus connection. The bus can be an Industry Standard Architecture (Industry Standard Architecture, referred to as: ISA) bus, a Peripheral Component Interconnect (Peripheral Component Interconnect, referred to as: PCI) bus or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, referred to as: EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc., but it does not mean that there is only one bus or one type of bus. The computer execution instructions for realizing the above-mentioned molecular ground state electron density detection method are stored in the memory 602, including at least one software function module that can be stored in the memory 602 in the form of software or firmware. The processor 601 executes various functional applications and data processing by running the software programs and modules stored in the memory 602.

[0129] The memory 602 includes at least one type of readable storage medium, including but not limited to random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), and electrically erasable read-only memory (EEPROM). The memory 602 is used to store programs, and the processor 601 executes the programs after receiving the execution instruction. Furthermore, the software programs and modules in the above-mentioned memory 602 may also include an operating system, which may include various software components and / or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and may communicate with various hardware or software components to provide an operating environment for other software components.

[0130] The processor 601 may be an integrated circuit chip having the ability to process signals. The processor 601 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), etc. The methods, steps, and logic diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor, or the processor 601 may be any conventional processor, etc.

[0131] The electronic device 600 is used to execute the technical solution provided by the aforementioned molecular ground state electron density detection method embodiment. Its implementation principle and technical effect are similar to those in the aforementioned method embodiment and will not be repeated here.

[0132] An embodiment of the present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed, they are used to implement the technical solution of the aforementioned molecular ground state electron density detection method.

[0133] The computer-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The computer-readable storage medium may be any available medium that can be accessed by a general or special purpose computer.

[0134] An exemplary readable storage medium is coupled to the processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also be present as discrete components in the control device of the molecular ground state electron density detection system.

[0135] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the technical solution of the aforementioned molecular ground state electron density detection method.

[0136] In the above embodiments, it will be appreciated by those skilled in the art that the above-mentioned various method embodiments can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loading and executing computer program instructions on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless network, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or a data center that contains one or more available media integrations. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).

[0137] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0138] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the appended claims.

[0139] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for detecting the ground state electron density of a molecule, characterized in that: include: Acquire first molecular data of the molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data; Substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected; Inputting the first molecular data into a kinetic energy functional model to obtain a first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training; Taking minimizing the first electron kinetic energy functional as an optimization goal, the initial electron density is iteratively optimized to obtain the ground state electron density of the molecule to be detected.

2. The method according to claim 1, characterized in that The method further comprises: Collecting second molecular data of a plurality of sample molecules to construct a training set; wherein the second molecular data is annotated with a true label, and the true label includes a second electron kinetic energy functional of the sample molecule; The training set is input into a pre-built deep learning model for iterative training to obtain the kinetic energy functional model; wherein the deep learning model includes a multi-layer perceptron, a convolutional neural network, a long short-term memory network and a Transformer neural network.

3. The method according to claim 2, characterized in that Inputting the training set into a pre-built deep learning model for iterative training to obtain the kinetic energy functional model, including: Inputting the second molecule data in the training set into the deep learning model to obtain an output result; Calculating the loss value between the output result and the true label based on a preset loss function; The parameters of the deep learning model are updated according to the loss value until the loss function converges or reaches a preset number of training iterations, thereby obtaining the kinetic energy functional model.

4. The method according to any one of claim 2, characterized in that: Collect the second molecule data of multiple sample molecules and construct a training set, including: Collecting second molecular data of a plurality of sample molecules; wherein the second molecular data includes second molecular structure data and second electronic structure data; Performing data cleaning, data standardization, and real labeling on the second molecular data to obtain training data; Based on the training data, a training set is constructed.

5. The method according to any one of claims 1 to 4, characterized in that: The density initialization algorithm is the Hirshfeld algorithm or the Voronoi degeneration algorithm.

6. A molecular ground state electron density detection system, characterized in that: include: A data acquisition module, used to acquire first molecular data of the molecule to be detected; wherein the first molecular data includes first molecular structure data and first electronic structure data; A density initialization module, used for substituting the first molecule data into a preset density initialization algorithm to obtain the initial electron density of the molecule to be detected; A kinetic energy functional determination module, used to input the first molecular data into a kinetic energy functional model to obtain a first electron kinetic energy functional output by the kinetic energy functional model; wherein the kinetic energy functional model is obtained based on deep learning model training; The density optimization module is used to iteratively optimize the initial electron density by minimizing the first electron kinetic energy functional to obtain the ground state electron density of the molecule to be detected.

7. A cloud computing platform, characterized in that: Including, the molecular ground state electron density detection system as described in claim 6.

8. The cloud computing platform according to claim 7, characterized in that: The cloud computing platform also includes a resource allocation system, which is used to: Based on the first molecular data of the molecule to be detected, evaluating the amount of computing resources required by the molecular ground state electron density detection system to perform ground state electron density detection on the molecule to be detected; Based on the amount of computing resources, cloud computing resources are scheduled and allocated; wherein the cloud computing resources are used to deploy the molecular ground state electron density detection system.

9. An electronic device, characterized in that: comprising a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed, are used to implement the method according to any one of claims 1 to 5.

Images