Processing method and device of organic molecule maximum absorption peak prediction model

Abstract

The embodiment of the invention relates to a processing method and device of an organic molecule maximum absorption peak prediction model. The method comprises the following steps: constructing a maximum absorption peak prediction model for predicting the maximum absorbance and the maximum absorption wavelength of a maximum absorption peak according to a molecular structure, environment temperature, environment illumination intensity and an environment PH value input by the model by taking an MLP model and a Un i-Mo < l > model as a core component, wherein the maximum absorption peak prediction model is used for predicting the maximum absorbance and the maximum absorption wavelength of the maximum absorption peak according to the molecular structure, the environment temperature, the environment illumination intensity and the environment PH value input by the model; constructing a first data set through data acquisition, and training a maximum absorption peak prediction model based on the first data set; and after the model training is finished, receiving a first molecular structure and a first environment parameter set input by the user, and predicting the maximum absorption peak information of the first molecular structure under all first environment parameters by using the maximum absorption peak prediction model to obtain a corresponding first prediction vector set to be fed back to the current user. According to the method, the task processing real-time performance of the maximum absorption peak recognition task can be improved, and the personalized prediction level can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to a method and device for processing a prediction model of the maximum absorption peak of organic molecules. Background Art

[0002] Organic photovoltaic (OPV) materials refer to organic materials that directly convert solar energy or other light energy into electrical energy through the photovoltaic effect. Organic photovoltaic materials are composed of organic molecules. The absorption spectral curve is an important optical property curve of such organic molecules. Under normal circumstances, the ordinate of this curve is absorbance and the abscissa is light wave wavelength. The maximum absorption peak is the position of the maximum peak point on the absorption spectral curve. The ordinate corresponding to this point is denoted as the maximum absorbance, and the abscissa is denoted as the maximum absorption wavelength. Currently, when the structure of an organic molecule is known, to identify its maximum absorption peak, either an experimental method or a simulation calculation method is used; the experimental method means obtaining the absorption spectral curve through a series of experimental means first, and then further extracting the maximum absorption peak from the curve; the simulation calculation method means obtaining the maximum absorption peak through a series of quantum chemical calculations (such as density functional theory calculations, time-dependent density functional theory calculations, quantum mechanics calculations, molecular mechanics calculations, etc.). However, from practical experience, both of these conventional processing methods have the problems of long processing cycles and poor real-time performance. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, device, electronic device and computer-readable storage medium for processing a prediction model of the maximum absorption peak of organic molecules in view of the defects of the prior art. The present invention constructs a deep learning model for predicting the maximum absorption peak parameters (maximum absorbance, maximum absorption wavelength) of the current molecule based on the molecular structure, ambient temperature, ambient light intensity, and ambient pH value input into the model, with the MLP model and the Uni-Mol model as the core components, denoted as the maximum absorption peak prediction model; and collects the molecular structure data of various organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions (temperature, light intensity, pH value) through experiments, simulation calculations, and big data collection methods to obtain a first data set for model training; and trains the maximum absorption peak prediction model based on the first data set; and after the model training is completed, uses the maximum absorption peak prediction model to predict the maximum absorption peak information of the molecular structure input by the user under multiple sets of environmental parameters (temperature, light intensity, pH value) specified by the user. On the one hand, the technical solution of the present invention can shorten the task processing cycle of the maximum absorption peak identification task and improve the task processing real-time performance based on the maximum absorption peak prediction model; on the other hand, it can also improve the personalized prediction level of the maximum absorption peak parameters (maximum absorbance, maximum absorption wavelength) by combining environmental parameters.

[0004] To achieve the above object, a first aspect of an embodiment of the present invention provides a processing method for an organic molecule maximum absorption peak prediction model, the method comprising:

[0005] Construct a deep learning model with an MLP model and a Uni-Mol model as core components, denoted as the corresponding maximum absorption peak prediction model; the maximum absorption peak prediction model is used to predict the maximum absorbance and maximum absorption wavelength of the maximum absorption peak based on the molecular structure X, ambient temperature, ambient light intensity, and ambient pH value input to the model and output the corresponding prediction vector Y; the prediction vector Y includes a predicted absorbance and a predicted wavelength, and the predicted absorbance and the predicted wavelength correspond to the maximum absorbance and the maximum absorption wavelength respectively;

[0006] Collect molecular structure data of various organic molecules of organic photovoltaic materials and maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations, and big data collection methods, and form a corresponding first data set based on the collected data; the environmental conditions include temperature, light intensity, and pH value;

[0007] Train the maximum absorption peak prediction model based on the first data set;

[0008] After the model training is completed, receive the first molecular structure and the first set of environmental parameters input by the user; and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first set of environmental parameters, and obtain the corresponding first set of prediction vectors and feedback to the current user; the first set of environmental parameters includes a plurality of the first environmental parameters; the first environmental parameter includes temperature, light intensity, and pH value; the first set of prediction vectors includes a plurality of first prediction vectors; the first prediction vector is composed of a first absorbance and a first wavelength and corresponds one-to-one to the first environmental parameter.

[0009] Preferably, the molecular structure X includes a plurality of atoms a i , 1 ≤ atomic index i ≤ A X , A X is the total number of atoms of the molecular structure X; the atomic parameters of each atom a i include the atomic type t i and the three-dimensional atomic coordinate c i (x, y, z);

[0010] The first data set includes a plurality of first data records; the first data records include a first training molecular structure, first training environmental parameters, and a first maximum absorption peak label; the first training molecular structure includes a plurality of first atoms, and the atomic parameters of each first atom include an atomic type and three-dimensional atomic coordinates; the first training environmental parameters include a first temperature, a first light intensity, and a first pH value; the first maximum absorption peak label includes a first maximum absorbance label and a first maximum absorption wavelength label.

[0011] Preferably, the maximum absorption peak prediction model includes a preprocessing module, a first MLP model, a Uni-Mol model, a feature fusion module, a second MLP model, a third MLP model, and a prediction output module;

[0012] The input end of the preprocessing module is connected to the model input end of the maximum absorption peak prediction model, and the first and second output ends are respectively connected to the input ends of the first MLP model and the Uni-Mol model; the output ends of the first MLP model and the Uni-Mol model are respectively connected to the first and second input ends of the feature fusion module; the output end of the feature fusion module is respectively connected to the input ends of the second MLP model and the third MLP model; the output ends of the second MLP model and the third MLP model are respectively connected to the first and second input ends of the prediction output module; the output end of the prediction output module is connected to the model output end of the maximum absorption peak prediction model;

[0013] The preprocessing module is used to perform embedding encoding on the environmental temperature, the environmental light intensity, and the environmental pH value input to the model respectively according to a preset embedding encoding method for temperature, light intensity, and pH value to obtain corresponding temperature encoding, light intensity encoding, and pH value encoding, which form a corresponding first environmental encoding vector; and based on the one-hot encoding rule of atomic types of the Uni-Mol model, for each atom a according to all atomic types of the molecular structure X iPerform atomic type one-hot encoding to obtain the corresponding first atomic one-hot encoding vector, and form the corresponding first atomic encoding tensor from all the obtained first atomic one-hot encoding vectors; and initialize the encoding rules based on the atomic pair features of the Uni-Mol model, and perform pairwise atomic feature initialization on each pair of atoms according to all the three-dimensional atomic coordinates of the molecular structure X to obtain the corresponding first atomic pair encoding tensor; and send the first environmental encoding vector to the first MLP model; and send the first atomic encoding tensor and the first atomic pair encoding tensor to the Uni-Mol model; the first environmental encoding vector is a real vector of length 3, composed of the temperature encoding, the light intensity encoding, and the pH value encoding; the first atomic one-hot encoding vector is a one-dimensional one-hot encoding vector, and the vector length is denoted as D1, and D1 matches the total number of atomic types of the molecular structure X; in the first atomic one-hot encoding vector, the one-hot encoding corresponding to the current atomic type t i of the atom a i is set to 1, and the remaining D1 - 1 one-hot encodings are set to 0; the tensor shape of the first atomic encoding tensor is D1×A X ; the tensor shape of the first atomic pair encoding tensor is D2×A X ×A X , where D2 is the preset atomic pair encoding feature dimension;

[0014] The first MLP model is composed of multiple fully connected layers connected in sequence; the first MLP model is used to perform environmental feature extraction processing on the input first environmental encoding vector to obtain the corresponding first environmental feature tensor and send it to the feature fusion module; the tensor shape of the first environmental feature tensor is D3×3, where D3 is the preset environmental feature dimension; the first environmental feature tensor is specifically composed of three environmental feature vectors with a vector length of D3 each, namely the first temperature feature vector, the first light intensity feature vector, and the first pH value feature vector;

[0015] The Uni-Mol model is used to perform atomic and atomic pair feature extraction processing on the input first atomic encoding tensor and the first atomic pair encoding tensor to obtain the corresponding first atomic feature tensor and first atomic pair feature tensor; and perform feature fusion on the first atomic feature tensor and the first atomic pair feature tensor to obtain the corresponding first molecular feature tensor and send it to the feature fusion module; the tensor shape of the first atomic feature tensor is D4×A X , where D4 is the preset atomic feature dimension; the first atomic feature tensor is specifically composed of A X first sub-feature vectors with a vector length of D4 each; the tensor shape of the first atomic pair feature tensor is D5×A X ×A X, D5 is a preset atomic pair feature dimension; the first atomic pair feature tensor is specifically composed of A X first sub-feature tensors each with a shape of D5×A X . Each of the first sub-feature tensors is specifically composed of A X second sub-feature vectors each with a vector length of D5; the tensor shape of the first molecular feature tensor is (D4 + D5×A X )×A X ; the first molecular feature tensor is specifically composed of A X first atomic fusion vectors each with a vector length of (D4 + D5×A X ); each of the first atomic fusion vectors is formed by sequentially splicing a corresponding first sub-feature vector and A X corresponding second sub-feature vectors;

[0016] The feature fusion module is used to sequentially splice each of the first atomic fusion vectors of the first molecular feature tensor and the three environmental feature vectors of the first environmental feature tensor to form a corresponding second atomic fusion vector; and all the obtained second atomic fusion vectors are used to form a corresponding first fusion feature tensor and send it to the second and third MLP models; the tensor shape of the first fusion feature tensor is (D3×3 + D4 + D5×A X )×A X ;

[0017] The second MLP model is composed of multiple fully connected layers connected in sequence; the second MLP model is used to perform maximum absorbance prediction processing on the input first fusion feature tensor to obtain the corresponding predicted absorbance and send it to the prediction output module;

[0018] The third MLP model is composed of multiple fully connected layers connected in sequence; the third MLP model is used to perform maximum absorption wavelength prediction processing on the input first fusion feature tensor to obtain the corresponding predicted wavelength and send it to the prediction output module;

[0019] The prediction output module forms a corresponding prediction vector Y from the obtained predicted absorbance and the predicted wavelength and outputs it.

[0020] Preferably, the model training of the maximum absorption peak prediction model based on the first data set specifically includes:

[0021] Step 41, dividing the first data set based on a preset first division ratio to obtain two sub-data sets denoted as the corresponding first training set and first evaluation set;

[0022] Among them, both the first training set and the first evaluation set are composed of multiple first data records; the ratio of the total number of records in the first training set and the first evaluation set satisfies the first splitting ratio;

[0023] Step 42: Take the first first data record in the first training set as the corresponding current training record;

[0024] Step 43: Input the first training molecular structure of the current training record, the first temperature, the first light intensity, and the first pH value of the first training environment parameters as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y;

[0025] Step 44: Extract the first maximum absorption peak label of the current training record as the corresponding current label vector, and take the current prediction vector Y as the corresponding current prediction vector; and substitute the current prediction vector and the current label vector into a preset first model loss function for calculation to obtain the corresponding first loss value;

[0026] Among them, the first model loss function is implemented based on the L1 loss function or the L2 loss function;

[0027] Step 45: Identify whether the first loss value meets a preset first loss value range; if the first loss value meets the first loss value range, identify whether the current training record is the last first data record in the first training set. If so, go to Step 46; if not, take the next first data record in the first training set as the new current training record and return to Step 43 to continue training; if the first loss value does not meet the first loss value range, based on a preset first model optimizer, modulate the model parameters of the first, second, and third MLP models and the Uni-Mol model of the maximum absorption peak prediction model in the direction of minimizing the first model loss function for one round, and return to Step 43 to continue training at the end of this round of modulation;

[0028] Among them, the first model optimizer includes at least the Adam optimizer and the SGD optimizer;

[0029] Step 46, perform a round of traversal on all the first data records in the first evaluation set; and during this round of traversal, use the currently traversed first data record as the corresponding current evaluation record; and use the first training molecular structure of the current evaluation record and the first temperature, the first light intensity, and the first pH value of the first training environment parameters as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value and input them into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; and extract the first maximum absorption peak label of the current evaluation record as the corresponding current label vector, use the current prediction vector Y as the corresponding current prediction vector, and form a corresponding prediction-label pair from the current prediction vector and the current label vector; and at the end of this round of traversal, input all the obtained prediction-label pairs into a preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0030] Among them, the first model evaluation function is implemented based on the MAE function, the MSE function, or the RMSE function;

[0031] Step 47, identify whether the first evaluation value meets a preset first evaluation value range; if not, return to Step 42 to continue training; if so, stop training and confirm that the model training is completed.

[0032] Preferably, the step of using the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first environmental parameter set and feeding back the corresponding first prediction vector set to the current user specifically includes:

[0033] Perform a round of traversal on all the first environmental parameters in the first environmental parameter set; and during this round of traversal, use the currently traversed first environmental parameter as the corresponding current environmental parameter; and use the first molecular structure and the temperature, light intensity, and pH value of the current environmental parameter as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value and input them into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; and use the predicted absorbance and the predicted wavelength of the current prediction vector Y as the corresponding first absorbance and the first wavelength to form a corresponding first prediction vector; and at the end of this round of traversal, form the corresponding first prediction vector set from all the obtained first prediction vectors and feed it back to the current user.

[0034] A second aspect of the embodiments of the present invention provides an apparatus for implementing the processing method of the organic molecule maximum absorption peak prediction model described in the first aspect above. The apparatus includes: a model construction module, a data acquisition module, a model training module, and a model application module;

[0035] The model construction module is used to construct a deep learning model with the MLP model and the Uni-Mol model as core components, denoted as the corresponding maximum absorption peak prediction model; the maximum absorption peak prediction model is used to predict the maximum absorbance and maximum absorption wavelength of the maximum absorption peak based on the molecular structure X, environmental temperature, environmental light intensity, and environmental pH value input to the model and output the corresponding prediction vector Y; the prediction vector Y includes a predicted absorbance and a predicted wavelength, and the predicted absorbance and the predicted wavelength correspond to the maximum absorbance and the maximum absorption wavelength respectively;

[0036] The data acquisition module is used to collect data on the molecular structure data of various organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations, and big data collection methods, and form a corresponding first data set based on the collected data; the environmental conditions include temperature, light intensity, and pH value;

[0037] The model training module is used to train the maximum absorption peak prediction model based on the first data set;

[0038] The model application module is used to, after the model training is completed, receive the first molecular structure and the first set of environmental parameters input by the user; and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first set of environmental parameters to obtain the corresponding first set of prediction vectors and feedback them to the current user; the first set of environmental parameters includes multiple first environmental parameters; the first environmental parameter includes temperature, light intensity, and pH value; the first set of prediction vectors includes multiple first prediction vectors; the first prediction vector is composed of a first absorbance and a first wavelength and corresponds one-to-one to the first environmental parameter.

[0039] A third aspect of the embodiments of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0040] The processor is used to be coupled with the memory, read and execute the instructions in the memory to implement the method steps described in the first aspect above;

[0041] The transceiver is coupled with the processor, and the processor controls the transceiver to perform message sending and receiving.

[0042] In the fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer is caused to execute the instructions of the method described in the first aspect above.

[0043] The embodiments of the present invention provide a processing method, apparatus, electronic device, and computer-readable storage medium for an organic molecule maximum absorption peak prediction model. As can be seen from the above, the embodiments of the present invention use an MLP model and a Uni-Mol model as core components to construct a deep learning model for predicting the maximum absorption peak parameters (maximum absorbance, maximum absorption wavelength) of the current molecule based on the molecular structure, ambient temperature, ambient light intensity, and ambient pH value input to the model, denoted as the maximum absorption peak prediction model; and collect molecular structure data of various organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions (temperature, light intensity, pH value) through experiments, simulation calculations, and big data collection methods to obtain a first data set for model training; and train the maximum absorption peak prediction model based on the first data set; and after the model training is completed, use the maximum absorption peak prediction model to predict the maximum absorption peak information of the molecular structure input by the user under multiple sets of environmental parameters (temperature, light intensity, pH value) specified by the user. On the one hand, the embodiments of the present invention shorten the task processing cycle of the maximum absorption peak identification task and improve the real-time performance of task processing through the maximum absorption peak prediction model; on the other hand, the personalized prediction level of the maximum absorption peak parameters is improved by combining environmental parameters during model prediction. Description of the Drawings

[0044] Figure 1 It is a schematic diagram of a processing method for an organic molecule maximum absorption peak prediction model provided in Embodiment 1 of the present invention;

[0045] Figure 2 It is a module structure diagram of the maximum absorption peak prediction model provided in Embodiment 1 of the present invention;

[0046] Figure 3 It is a module structure diagram of a processing apparatus for an organic molecule maximum absorption peak prediction model provided in Embodiment 2 of the present invention;

[0047] Figure 4 It is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. Detailed Embodiments

[0048] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0049] Embodiment 1 of the present invention provides a method for processing a prediction model of the maximum absorption peak of an organic molecule, as Figure 1 shown in the schematic diagram of a method for processing a prediction model of the maximum absorption peak of an organic molecule provided in Embodiment 1 of the present invention, this method mainly includes the following steps:

[0050] Step 1, construct a deep learning model with an MLP model and a Uni-Mol model as core components, denoted as the corresponding maximum absorption peak prediction model.

[0051] Here, the maximum absorption peak prediction model of the embodiment of the present invention is used to predict the maximum absorbance and the maximum absorption wavelength of the maximum absorption peak according to the molecular structure X, ambient temperature, ambient light intensity, and ambient pH value input into the model, and output the corresponding prediction vector Y.

[0052] The molecular structure X input into the model includes multiple atoms a i , 1 ≤ atomic index i ≤ A X , A X is the total number of atoms in the molecular structure X; the atomic parameters of each atom a i include the atomic type t i and the three-dimensional atomic coordinate c i (x, y, z). The ambient temperature input into the model is a temperature value, and the unit is defaulted to degrees Celsius. It should be noted that the unit of all temperatures mentioned in the embodiments of the present invention is defaulted to degrees Celsius. The ambient light intensity input into the model is the light intensity of a preset light type, such as ultraviolet light, infrared light, etc. It should be noted that all light intensities mentioned in the embodiments of the present invention are the light intensities of this preset light type. The ambient pH value input into the model is the acidity and alkalinity parameter of the chemical environment where the organic molecule corresponding to the current molecular structure X is located, and its value is any integer within the conventional pH value range, that is, between 0 and 14. The prediction vector Y output by the model includes the predicted absorbance and the predicted wavelength, the predicted absorbance corresponds to the maximum absorbance of the maximum absorption peak, and the predicted wavelength corresponds to the maximum absorption wavelength of the maximum absorption peak.

[0053] The model structure of the maximum absorption peak prediction model is as Figure 2As shown in the module structure diagram of the maximum absorption peak prediction model provided in the first embodiment of the present invention, it includes: a preprocessing module, a first MLP model, a Uni-Mol model, a feature fusion module, a second MLP model, a third MLP model, and a prediction output module.

[0054] The connection relationships of the components of the maximum absorption peak prediction model are as follows: the input end of the preprocessing module is connected to the model input end of the maximum absorption peak prediction model, and the first and second output ends are respectively connected to the input ends of the first MLP model and the Uni-Mol model; the output ends of the first MLP model and the Uni-Mol model are respectively connected to the first and second input ends of the feature fusion module; the output end of the feature fusion module is respectively connected to the input ends of the second MLP model and the third MLP model; the output ends of the second MLP model and the third MLP model are respectively connected to the first and second input ends of the prediction output module; the output end of the prediction output module is connected to the model output end of the maximum absorption peak prediction model.

[0055] The functions of the components of the maximum absorption peak prediction model are as follows.

[0056] 1) Preprocessing module:

[0057] The preprocessing module of the embodiment of the present invention is used to perform embedding encoding on the ambient temperature, ambient light intensity, and ambient pH value input to the model respectively according to the preset temperature, light intensity, and pH value embedding encoding method to obtain corresponding temperature encoding, light intensity encoding, and pH value encoding, which form a corresponding first environmental encoding vector; and based on the atomic type one-hot encoding rule of the Uni-Mol model, perform atomic type one-hot encoding on each atom a according to all atomic types of the molecular structure X i to obtain a corresponding first atomic one-hot encoding vector, and form a corresponding first atomic encoding tensor from all the obtained first atomic one-hot encoding vectors; and based on the atomic pair feature initialization encoding rule of the Uni-Mol model, perform pairwise atomic feature initialization on each pair of atoms according to all three-dimensional atomic coordinates of the molecular structure X to obtain a corresponding first atomic pair encoding tensor; and send the first environmental encoding vector to the first MLP model; and send the first atomic encoding tensor and the first atomic pair encoding tensor to the Uni-Mol model.

[0058] Here, the preset temperature, light intensity, and pH value embedding coding methods in the embodiments of the present invention are three preset embedding coding methods, which can be customized based on specific application requirements. One common customization method is as follows: The temperature and light intensity are processed based on the normalization method, and the obtained normalized values are used as the corresponding temperature coding and light intensity coding, and the value of the pH value is directly used as the corresponding pH value coding. In addition, the Uni-Mol model used in the embodiments of the present invention is an encoder model implemented based on the Encoder sub-model of the Transformer model and used for atomic-level feature coding of the molecular structure. The detailed model structure, model inference principle, one-hot encoding rule for atomic types, initialization encoding rule for atomic pair features, and model pre-training scheme of the Uni-Mol model have been clearly described in the public technical document A, "Uni-Mol: A Universal 3D Molecular Representation Learning Framework". Therefore, the one-hot encoding rule for atomic types and the initialization encoding rule for atomic pair features used in the preprocessing module will not be further elaborated here.

[0059] It should be noted that the first environmental encoding vector generated during the processing of the preprocessing module in the embodiments of the present invention is a real vector with a length of 3, which is composed of temperature coding, light intensity coding, and pH value coding. The first atomic one-hot encoding vector generated by the preprocessing module is a one-dimensional one-hot encoding vector, and the vector length is denoted as D1, and D1 matches the total number of atomic types of the molecular structure X; in this first atomic one-hot encoding vector, the one-hot encoding corresponding to the atomic type t i of the current atom a i is set to 1, and the remaining D1 - 1 one-hot encodings are set to 0. The tensor shape of the first atomic encoding tensor generated by the preprocessing module is D1 × A X . The tensor shape of the first atomic pair encoding tensor generated by the preprocessing module is D2 × A X × A X , where D2 is the preset atomic pair encoding feature dimension. If the preprocessing module only calculates and normalizes the Euclidean distance of atomic pairs, then D2 can be set to 1 at least.

[0060] 2) The first MLP model:

[0061] The first MLP model in the embodiments of the present invention is composed of multiple fully connected layers connected in sequence. This first MLP model is used to perform environmental feature extraction processing based on the input first environmental encoding vector and send the corresponding first environmental feature tensor to the feature fusion module.

[0062] Here, the tensor shape of the first environmental feature tensor in the embodiments of the present invention is D3×3, where D3 is a preset environmental feature dimension; the first environmental feature tensor can be specifically regarded as composed of three environmental feature vectors with a vector length of D3 each, namely the first temperature feature vector, the first light intensity feature vector, and the first pH value feature vector.

[0063] 3) Uni-Mol model:

[0064] The Uni-Mol model in the embodiments of the present invention is used to perform atomic and atomic pair feature extraction processing on the input first atomic encoding tensor and first atomic pair encoding tensor to obtain the corresponding first atomic feature tensor and first atomic pair feature tensor; and perform feature fusion on the first atomic feature tensor and the first atomic pair feature tensor to obtain the corresponding first molecular feature tensor and send it to the feature fusion module.

[0065] Here, the tensor shape of the first atomic feature tensor in the embodiments of the present invention is D4×A X , where D4 is a preset atomic feature dimension; the first atomic feature tensor can be specifically regarded as composed of A X first sub-feature vectors with a vector length of D4 each.

[0066] The tensor shape of the first atomic pair feature tensor in the embodiments of the present invention is D5×A X ×A X , where D5 is a preset atomic pair feature dimension; the first atomic pair feature tensor can be specifically regarded as composed of A X first sub-feature tensors with a shape of D5×A X each, and each first sub-feature tensor can be specifically regarded as composed of A X second sub-feature vectors with a vector length of D5 each.

[0067] The tensor shape of the first molecular feature tensor in the embodiments of the present invention is (D4 + D5×A X )×A X ; the first molecular feature tensor can be specifically regarded as composed of A X first atomic fusion vectors with a vector length of (D4 + D5×A X ) each; each first atomic fusion vector can be regarded as composed of a corresponding first sub-feature vector and A X corresponding second sub-feature vectors sequentially concatenated.

[0068] As described above, the detailed model structure, model inference principle, and model pre-training scheme of the Uni-Mol model used in the embodiments of the present invention have been published in the public technical literature A. Therefore, the encoding process of the Uni-Mol model will not be further elaborated here. However, it should be noted that the Uni-Mol model used in the embodiments of the present invention has been pre-trained in advance according to the model pre-training scheme given in the public technical literature A.

[0069] 4) Feature fusion module:

[0070] The feature fusion module of the embodiments of the present invention is used to sequentially splice each first atomic fusion vector of the first molecular feature tensor and the three environmental feature vectors of the first environmental feature tensor to form a corresponding second atomic fusion vector; and all the obtained second atomic fusion vectors are used to form a corresponding first fusion feature tensor and send it to the second and third MLP models.

[0071] Here, the tensor shape of the first fusion feature tensor of the embodiments of the present invention is (D3×3 + D4 + D5×A X )×A X .

[0072] 5) Second MLP model:

[0073] The second MLP model of the embodiments of the present invention is sequentially connected by multiple fully connected layers. The second MLP model is used to perform maximum absorbance prediction processing on the input first fusion feature tensor to obtain a corresponding predicted absorbance and send it to the prediction output module.

[0074] 6) Third MLP model:

[0075] The third MLP model of the embodiments of the present invention is sequentially connected by multiple fully connected layers. The third MLP model is used to perform maximum absorption wavelength prediction processing on the input first fusion feature tensor to obtain a corresponding predicted wavelength and send it to the prediction output module.

[0076] 7) Prediction output module:

[0077] The prediction output module of the embodiments of the present invention forms a corresponding prediction vector Y from the obtained predicted absorbance and predicted wavelength and outputs it.

[0078] Step 2, collect the molecular structure data of various organic molecules of the organic photovoltaic material and the maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations, and big data collection methods, and form a corresponding first data set based on the collected data;

[0079] Among them, the environmental conditions include temperature, light intensity, and pH value; the first data set includes a plurality of first data records; the first data record includes a first training molecular structure, first training environmental parameters, and a first maximum absorption peak label; the first training molecular structure includes a plurality of first atoms, and the atomic parameters of each first atom include atomic type and three-dimensional atomic coordinates; the first training environmental parameters include a first temperature, a first light intensity, and a first pH value; the first maximum absorption peak label includes a first maximum absorbance label and a first maximum absorption wavelength label;

[0080] Specifically, it includes: Step 21, measuring the maximum absorption peak information of a plurality of organic molecules with known molecular structures for organic photovoltaic materials under different environmental conditions by experimental methods; and taking the molecular structure of each organic molecule processed during the experiment as a corresponding first training molecular structure, and taking the temperature, light intensity, and pH value of each environmental condition set during the experiment as a corresponding set of first temperature, first light intensity, and first pH value to form a corresponding first training environmental parameter, and taking the maximum absorbance and maximum absorption wavelength of the maximum absorption peak information measured for each organic molecule under each environmental condition during the experiment as a corresponding set of first maximum absorbance label and first maximum absorption wavelength label to form a corresponding first maximum absorption peak label; finally, each first training molecular structure obtained by the experimental method and its corresponding first training environmental parameter and first maximum absorption peak label form a corresponding first data record;

[0081] Step 22, performing simulation calculations on the maximum absorption peak information of a plurality of organic molecules with known molecular structures for organic photovoltaic materials under different environmental conditions by simulation calculation methods; and taking the molecular structure of each organic molecule processed during the simulation calculation as a corresponding first training molecular structure, and taking the temperature, light intensity, and pH value of each environmental condition set during the simulation calculation as a corresponding set of first temperature, first light intensity, and first pH value to form a corresponding first training environmental parameter, and taking the maximum absorbance and maximum absorption wavelength of the maximum absorption peak information measured for each organic molecule under each environmental condition during the simulation calculation as a corresponding set of first maximum absorbance label and first maximum absorption wavelength label to form a corresponding first maximum absorption peak label; finally, each first training molecular structure obtained by the simulation calculation method and its corresponding first training environmental parameter and first maximum absorption peak label form a corresponding first data record;

[0082] Step 23: Through a variety of preset big data collection channels, collect big data on the maximum absorption peak information of multiple organic molecules used in organic photovoltaic materials with known molecular structures under different environmental conditions; and take the molecular structure of each collected organic molecule as a corresponding first training molecular structure, and take the temperature, light intensity, and pH value of each environmental condition corresponding to each collected organic molecule as a group of corresponding first temperature, first light intensity, and first pH value to form a corresponding first training environmental parameter, and take the maximum absorbance and maximum absorption wavelength of the maximum absorption peak information of each collected organic molecule under each corresponding environmental condition as a group of corresponding first maximum absorbance label and first maximum absorption wavelength label to form a corresponding first maximum absorption peak label; finally, each first training molecular structure obtained by big data collection and processing, its corresponding first training environmental parameter, and first maximum absorption peak label form a corresponding first data record.

[0083] Here, the variety of big data collection channels in the embodiments of the present invention at least include all publicly available molecular structure information databases and molecular optical property information databases of organic photovoltaic material molecules, all publicly available scientific literature, periodicals, magazines, papers, technical solutions, experimental reports, etc.

[0084] Step 24: All the first data records obtained by experimental methods, simulation calculation methods, and big data collection methods form a corresponding first record set; and perform duplicate removal and abnormal data filtering processing on the first data records in the first record set; finally, take the first record set after duplicate removal and abnormal data filtering as the corresponding first data set.

[0085] Step 3: Train the maximum absorption peak prediction model based on the first data set.

[0086] Specifically, it includes: Step 31: Divide the first data set based on a preset first division ratio to obtain two sub-data sets denoted as the corresponding first training set and first evaluation set.

[0087] Here, the first division ratio in the embodiments of the present invention is a preset ratio parameter, such as 8:2; both the first training set and the first evaluation set are composed of multiple first data records; the record total ratio of the first training set and the first evaluation set satisfies the first division ratio.

[0088] Step 32: Take the first first data record in the first training set as the corresponding current training record.

[0089] Step 33: Input the first training molecular structure of the current training record, as well as the first temperature, first light intensity, and first pH value of the first training environment parameters, as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y;

[0090] Step 34: Extract the first maximum absorption peak label of the current training record as the corresponding current label vector, and use the current prediction vector Y as the corresponding current prediction vector; then substitute the current prediction vector and the current label vector into the preset first model loss function for calculation to obtain the corresponding first loss value;

[0091] Here, the first model loss function of the embodiment of the present invention is implemented based on the L1 loss function or the L2 loss function;

[0092] Step 35: Identify whether the first loss value meets the preset first loss value range; if the first loss value meets the first loss value range, then identify whether the current training record is the last first data record of the first training set. If so, go to Step 36; if not, use the next first data record of the first training set as the new current training record and return to Step 33 to continue training; if the first loss value does not meet the first loss value range, then based on the preset first model optimizer, modulate the model parameters of the first, second, and third MLP models and the Uni-Mol model of the maximum absorption peak prediction model in the direction of minimizing the first model loss function for one round, and return to Step 33 to continue training at the end of this round of modulation;

[0093] Here, the first loss value range of the embodiment of the present invention is a pre-set loss value range; the first model optimizer includes at least the Adam optimizer and the SGD optimizer;

[0094] Step 36: Conduct one round of traversal of all first data records in the first evaluation set; during this round of traversal, use the currently traversed first data record as the corresponding current evaluation record; input the first training molecular structure of the current evaluation record, as well as the first temperature, first light intensity, and first pH value of the first training environment parameters, as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; extract the first maximum absorption peak label of the current evaluation record as the corresponding current label vector, and use the current prediction vector Y as the corresponding current prediction vector, and form a corresponding prediction-label pair from the current prediction vector and the current label vector; at the end of this round of traversal, substitute all the obtained prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value;

[0095] Here, the first model evaluation function of the embodiments of the present invention is implemented based on the MAE function, the MSE function, or the RMSE function;

[0096] Step 37, identify whether the first evaluation value meets a preset first evaluation value range; if not, return to Step 32 to continue training; if so, stop training and confirm that the model training is completed.

[0097] Here, the first evaluation value range of the embodiments of the present invention is a preset evaluation value range.

[0098] Step 4, after the model training is completed, receive the first molecular structure and the first set of environmental parameters input by the user; and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first set of environmental parameters to obtain the corresponding first prediction vector set and feedback it to the current user;

[0099] Specifically, it includes: Step 41, after the model training is completed, receive the first molecular structure and the first set of environmental parameters input by the user;

[0100] Here, the first molecular structure of the embodiments of the present invention is the molecular structure of an organic molecule for an organic photovoltaic material, specifically composed of multiple atoms, and the atomic parameters of each atom are composed of the corresponding atomic type and the atomic three-dimensional coordinates; the first set of environmental parameters includes multiple first environmental parameters; each first environmental parameter includes temperature, light intensity, and pH value;

[0101] Step 42, and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first set of environmental parameters to obtain the corresponding first prediction vector set and feedback it to the current user;

[0102] Among them, the first prediction vector set of the embodiments of the present invention includes multiple first prediction vectors; the first prediction vector is composed of the first absorbance and the first wavelength; the first prediction vectors in the first prediction vector set correspond one by one to the first environmental parameters in the first set of environmental parameters;

[0103] Specifically, it includes: performing a round of traversal on all the first environmental parameters in the first environmental parameter set; during this round of traversal, taking the currently traversed first environmental parameter as the corresponding current environmental parameter; taking the first molecular structure and the temperature, light intensity, and pH value of the current environmental parameter as the corresponding molecular structure X, environmental temperature, environmental light intensity, and environmental pH value and inputting them into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; taking the predicted absorbance and predicted wavelength of the current prediction vector Y as the corresponding first absorbance and first wavelength to form a corresponding first prediction vector; and at the end of this round of traversal, forming a corresponding first prediction vector set from all the obtained first prediction vectors and feeding it back to the current user.

[0104] Figure 3 FIG. 4 is a module structure diagram of a processing device for a maximum absorption peak prediction model of organic molecules provided in Embodiment 2 of the present invention. This device is a terminal device or a server for implementing the foregoing method embodiment, or can also be a device that enables the foregoing terminal device or server to implement the foregoing method embodiment. For example, this device can be a device or a chip system of the foregoing terminal device or server. As Figure 3 shown, this device includes: a model construction module 201, a data collection module 202, a model training module 203, and a model application module 204.

[0105] The model construction module 201 is used to construct a deep learning model with the MLP model and the Uni-Mol model as the core components and denote it as the corresponding maximum absorption peak prediction model; the maximum absorption peak prediction model is used to predict the maximum absorbance and maximum absorption wavelength of the maximum absorption peak based on the molecular structure X, environmental temperature, environmental light intensity, and environmental pH value input into the model and output the corresponding prediction vector Y; the prediction vector Y includes the predicted absorbance and the predicted wavelength, and the predicted absorbance and the predicted wavelength respectively correspond to the maximum absorbance and the maximum absorption wavelength.

[0106] The data collection module 202 is used to collect the molecular structure data of various organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations, and big data collection methods and form a corresponding first data set based on the collected data; the environmental conditions include temperature, light intensity, and pH value.

[0107] The model training module 203 is used to perform model training on the maximum absorption peak prediction model based on the first data set.

[0108] The model application module 204 is configured to, after the model training is completed, receive a first molecular structure and a first set of environmental parameters input by a user; and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first set of environmental parameters to obtain a corresponding first set of prediction vectors and feedback them to the current user; the first set of environmental parameters includes a plurality of first environmental parameters; the first environmental parameters include temperature, light intensity, and pH value; the first set of prediction vectors includes a plurality of first prediction vectors; the first prediction vector is composed of a first absorbance and a first wavelength and corresponds to the first environmental parameter one by one.

[0109] The processing device for the maximum absorption peak prediction model of organic molecules provided by the embodiments of the present invention can execute the method steps in the above method embodiments, and its implementation principle and technical effects are similar, which will not be elaborated here.

[0110] It should be noted that it should be understood that the division of each module of the above device is only a logical function division. In actual implementation, it can be fully or partially integrated into a physical entity, or physically separated. And these modules can all be implemented in the form of software called by a processing element; they can also all be implemented in the form of hardware; or some modules can 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 model construction module can be a separately established processing element, or can be integrated in a certain chip of the above device. 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 to perform the functions of the above determined modules. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together or independently implemented. The processing element described 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 integrated logic circuit in the processor element or the instruction in the form of software.

[0111] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as: one or more Application Specific Integrated Circuits (ASICs), or, one or more Digital Signal Processors (DSPs), or, one or more Field Programmable Gate Arrays (FPGAs), etc. For another example, when a certain above module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processors that can call program code. For another example, these modules may be integrated together and implemented in the form of a System-on-a-chip (SOC).

[0112] In the above embodiments, it 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 the computer program instructions are loaded and executed on a computer, the processes or functions described in the foregoing method embodiments are generated in whole or in part. The above computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above computer instructions may 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 above computer instructions may be transmitted from a website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.). The above computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The above available medium may be a magnetic medium (for example, floppy disk, hard disk, magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, solid state disk (SSD)), etc.

[0113] Figure 4 It is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present invention. The electronic device may be a terminal device or a server that implements the method of the foregoing embodiments, or may be a terminal device or a server that is connected to the foregoing terminal device or server and implements the method of the foregoing embodiments. As Figure 4As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver operations of the transceiver 303. Various instructions may be stored in the memory 302 for completing various processing functions and implementing the processing steps described in the method of the foregoing embodiments. Preferably, the electronic device according to the embodiment of the present invention further includes: a power supply 304, a system bus 305, and a communication port 306. The system bus 305 is used to implement communication connections between components. The above communication port 306 is used for the electronic device to connect and communicate with other peripherals.

[0114] In Figure 4 the system bus 305 mentioned may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus may be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to implement communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include a Random Access Memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0115] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), a Graphics Processing Unit (GPU), etc.; it may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0116] It should be noted that the embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored, and when they run on a computer, the computer is caused to execute the methods and processing procedures provided in the above embodiments.

[0117] An embodiment of the present invention provides a processing method, device, electronic device, and computer-readable storage medium for an organic molecule maximum absorption peak prediction model. From the above content, it can be seen that the embodiment of the present invention constructs a deep learning model for predicting the maximum absorption peak parameters (maximum absorbance, maximum absorption wavelength) of the current molecule based on the molecular structure, ambient temperature, ambient light intensity, and ambient pH value input into the model, denoted as the maximum absorption peak prediction model, with the MLP model and the Uni-Mol model as the core components; and collects the molecular structure data of various organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions (temperature, light intensity, pH value) through experiments, simulation calculations, and big data collection methods to obtain a first data set for model training; and trains the maximum absorption peak prediction model based on the first data set; and after the model training is completed, uses the maximum absorption peak prediction model to predict the maximum absorption peak information of the molecular structure input by the user under multiple sets of environmental parameters (temperature, light intensity, pH value) specified by the user. On the one hand, the embodiment of the present invention shortens the task processing cycle of the maximum absorption peak identification task and improves the real-time performance of task processing through the maximum absorption peak prediction model; on the other hand, it improves the personalized prediction level of the maximum absorption peak parameters by combining environmental parameters during model prediction.

[0118] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented in hardware, software modules executed by a processor, or a combination of both. The software modules may be placed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0119] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for processing a maximum absorption peak prediction model of an organic molecule, characterized in that: The method comprises: A deep learning model is constructed with the MLP model and the Uni-Mo l model as core components and recorded as the corresponding maximum absorption peak prediction model; the maximum absorption peak prediction model is used to predict the maximum absorbance and the maximum absorption wavelength of the maximum absorption peak according to the molecular structure X, ambient temperature, ambient light intensity and ambient pH value input by the model and output the corresponding prediction vector Y; the prediction vector Y includes the predicted absorbance and the predicted wavelength, and the predicted absorbance and the predicted wavelength correspond to the maximum absorbance and the maximum absorption wavelength respectively; Collecting molecular structure data of multiple types of organic molecules of organic photovoltaic materials and maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations and big data collection, and forming a corresponding first data set based on the collected data; the environmental conditions include temperature, light intensity and pH value; Performing model training on the maximum absorption peak prediction model based on the first data set; After the model training is completed, the first molecular structure and the first environmental parameter set input by the user are received; and the maximum absorption peak prediction model is used to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first environmental parameter set to obtain the corresponding first prediction vector set and feedback it to the current user; the first environmental parameter set includes multiple first environmental parameters; the first environmental parameters include temperature, light intensity and pH value; the first prediction vector set includes multiple first prediction vectors; the first prediction vector consists of a first absorbance and a first wavelength, and corresponds one-to-one to the first environmental parameter.

2. The method for processing the maximum absorption peak prediction model of organic molecules according to claim 1, characterized in that: The molecular structure X includes a plurality of atoms a i , 1≤atomic index i≤A X , A X is the total number of atoms in the molecular structure X; each of the atoms a i The atomic parameters include the atomic type t i and the three-dimensional atomic coordinates c i (x,y,z); The first data set includes a plurality of first data records; the first data record includes a first training molecular structure, a first training environment parameter and a first maximum absorption peak label; The first training molecular structure includes multiple first atoms, and the atomic parameters of each of the first atoms include an atomic type and three-dimensional atomic coordinates; the first training environment parameters include a first temperature, a first light intensity and a first pH value; the first maximum absorption peak label includes a first maximum absorbance label and a first maximum absorption wavelength label.

3. The method for processing the maximum absorption peak prediction model of organic molecules according to claim 2, characterized in that: The maximum absorption peak prediction model includes a preprocessing module, a first MLP model, a Uni-Mo l model, a feature fusion module, a second MLP model, a third MLP model and a prediction output module; The input end of the preprocessing module is connected to the model input end of the maximum absorption peak prediction model, and the first and second output ends are respectively connected to the input ends of the first MLP model and the Uni-Mo l model; the output ends of the first MLP model and the Uni-Mo l model are respectively connected to the first and second input ends of the feature fusion module; the output end of the feature fusion module is respectively connected to the input ends of the second MLP model and the third MLP model; the output ends of the second MLP model and the third MLP model are respectively connected to the first and second input ends of the prediction output module; the output end of the prediction output module is connected to the model output end of the maximum absorption peak prediction model; The preprocessing module is used to embed the ambient temperature, ambient light intensity and ambient pH value of the model input according to the preset temperature, light intensity and pH value embedding coding method to obtain the corresponding temperature code, light intensity code and pH value code to form a corresponding first environmental coding vector; and based on the atomic type one-hot coding rule of the Uni-Mo l model, according to all the atomic types of the molecular structure X, each of the atomic a i Perform one-hot encoding of the atom type to obtain the corresponding first atom one-hot encoding vector, and form the corresponding first atom encoding tensor by all the first atom one-hot encoding vectors obtained; and based on the atom pair feature initialization encoding rule of the Uni-Mo l model, perform paired atomic feature initialization on each pair of atoms according to all three-dimensional atomic coordinates of the molecular structure X to obtain the corresponding first atom pair encoding tensor; and send the first environment encoding vector to the first MLP model; and send the first atom encoding tensor and the first atom pair encoding tensor to the Uni-Mo l model; the first environment encoding vector is a real number vector of length 3, which is composed of the temperature encoding, the light intensity encoding and the pH value encoding; the first atom one-hot encoding vector is a one-dimensional one-hot encoding vector, and the vector length is recorded as D1, D1 matches the total number of atom types of the molecular structure X; the first atom one-hot encoding vector contains the number of atoms corresponding to the current atom a i The atom type t i The corresponding one-hot encoding is set to 1, and the remaining D1-1 one-hot encodings are set to 0; the tensor shape of the first atomic encoding tensor is D1×A X ; The tensor shape of the first atom pair encoding tensor is D2×A X ×A X , D2 is the preset atomic pair encoding feature dimension; The first MLP model is formed by sequentially connecting multiple layers of fully connected layers; the first MLP model is used to perform environmental feature extraction processing according to the input first environmental coding vector to obtain a corresponding first environmental feature tensor and send it to the feature fusion module; the tensor shape of the first environmental feature tensor is D3×3, where D3 is a preset environmental feature dimension; the first environmental feature tensor is specifically composed of three environmental feature vectors with a vector length of D3, which are respectively a first temperature feature vector, a first light intensity feature vector, and a first pH value feature vector; The Uni-Mo l model is used to perform atomic and atomic pair feature extraction processing according to the input first atomic encoding tensor and the first atomic pair encoding tensor to obtain the corresponding first atomic feature tensor and the first atomic pair feature tensor; and perform feature fusion on the first atomic feature tensor and the first atomic pair feature tensor to obtain the corresponding first molecular feature tensor and send it to the feature fusion module; the tensor shape of the first atomic feature tensor is D4×A X , D4 is the preset atomic feature dimension; The first atomic eigentensor is specifically composed of A X The first atom pair eigenvector is composed of the first sub-eigenvectors of length D4; the tensor shape of the first atom pair eigentensor is D5×A X ×A X , D5 is the preset atomic pair feature dimension; the first atomic pair feature tensor is specifically composed of A X The shapes are all D5×A X The first sub-feature tensor is composed of A X The first sub-element tensor is composed of a second sub-element vector whose vector length is D5; the tensor shape of the first sub-element tensor is (D4+D5×A X )×A X ; The first molecular feature tensor is specifically composed of A X The length of each vector is (D4+D5×A X ) of the first atomic fusion vector; each of the first atomic fusion vectors is composed of a corresponding first sub-eigenvector and A X The corresponding second sub-feature vectors are sequentially concatenated; The feature fusion module is used to sequentially splice each of the first atomic fusion vectors of the first molecular feature tensor and the three environmental feature vectors of the first environmental feature tensor to form a corresponding second atomic fusion vector; and all the obtained second atomic fusion vectors form a corresponding first fusion feature tensor and send it to the second and third MLP models; the tensor shape of the first fusion feature tensor is (D3×3+D4+D5×A X )×A X ; The second MLP model is formed by sequentially connecting multiple layers of fully connected layers; the second MLP model is used to perform maximum absorbance prediction processing according to the first fusion feature tensor input to obtain the corresponding predicted absorbance and send it to the prediction output module; The third MLP model is formed by sequentially connecting multiple layers of fully connected layers; the third MLP model is used to perform maximum absorption wavelength prediction processing according to the first fusion feature tensor input to obtain the corresponding predicted wavelength and send it to the prediction output module; The prediction output module composes the corresponding prediction vector Y from the obtained predicted absorbance and the predicted wavelength and outputs it.

4. The method for processing the maximum absorption peak prediction model of organic molecules according to claim 2, characterized in that: The performing model training on the maximum absorption peak prediction model based on the first data set specifically includes: Step 41, segmenting the first data set based on a preset first segmentation ratio to obtain two sub-data sets recorded as a corresponding first training set and a first evaluation set; Wherein, both the first training set and the first evaluation set are composed of a plurality of the first data records; the ratio of the total number of records in the first training set and the first evaluation set satisfies the first segmentation ratio; Step 42, taking the first first data record of the first training set as the corresponding current training record; Step 43, inputting the first training molecular structure of the current training record and the first temperature, the first light intensity and the first pH value of the first training environment parameters as the corresponding molecular structure X, the environment temperature, the environment light intensity and the environment pH value into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; Step 44, extracting the first maximum absorption peak label of the current training record as the corresponding current label vector, and taking the current prediction vector Y as the corresponding current prediction vector; and bringing the current prediction vector and the current label vector into a preset first model loss function to calculate and obtain a corresponding first loss value; Wherein, the first model loss function is implemented based on the L1 loss function or the L2 loss function; Step 45, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, turning to step 46, if not, taking the next first data record of the first training set as the new current training record and returning to step 43 to continue training; if the first loss value does not satisfy the first loss value range, performing a round of modulation on the model parameters of the first, second, third MLP models and the Uni-Mo l model of the maximum absorption peak prediction model in the direction of minimizing the first model loss function based on a preset first model optimizer, and returning to step 43 to continue training at the end of this round of modulation; Wherein, the first model optimizer includes at least an Adam optimizer and an SGD optimizer; Step 46, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and input the first training molecular structure of the current evaluation record and the first temperature, the first light intensity and the first pH value of the first training environment parameter as the corresponding molecular structure X, the ambient temperature, the ambient light intensity and the ambient pH value into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; and extract the first maximum absorption peak label of the current evaluation record as the corresponding current label vector, and use the current prediction vector Y as the corresponding current prediction vector, and form a corresponding prediction-label pair by the current prediction vector and the current label vector; and at the end of this round of traversal, bring all the obtained prediction-label pairs into the preset first model evaluation function for calculation to obtain the corresponding first evaluation value; Wherein, the first model evaluation function is implemented based on the MAE function, the MSE function or the RMSE function; Step 47, identifying whether the first evaluation value satisfies a preset first evaluation value range; if not, returning to step 42 to continue training; if satisfied, stopping training and confirming the end of model training.

5. The method for processing the maximum absorption peak prediction model of organic molecules according to claim 1, characterized in that: The using the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first environmental parameter set to obtain a corresponding first prediction vector set and feed it back to the current user specifically includes: A round of traversal is performed on all the first environmental parameters of the first environmental parameter set; and in this round of traversal, the first environmental parameter currently traversed is used as the corresponding current environmental parameter; and the first molecular structure and the temperature, light intensity and pH value of the current environmental parameter are used as the corresponding molecular structure X, the ambient temperature, the ambient light intensity and the ambient pH value to input into the maximum absorption peak prediction model for prediction processing to obtain the corresponding prediction vector Y; and the predicted absorbance and the predicted wavelength of the current prediction vector Y are used as the corresponding first absorbance and the first wavelength to form a corresponding first prediction vector; and at the end of this round of traversal, all the obtained first prediction vectors form a corresponding first prediction vector set to feedback to the current user.

6. A device for executing the processing method of the organic molecule maximum absorption peak prediction model according to any one of claims 1 to 5, characterized in that: The device comprises: a model building module, a data acquisition module, a model training module and a model application module; The model building module is used to build a deep learning model with the MLP model and the Uni-Mo l model as core components, recorded as the corresponding maximum absorption peak prediction model; the maximum absorption peak prediction model is used to predict the maximum absorbance and the maximum absorption wavelength of the maximum absorption peak according to the molecular structure X, ambient temperature, ambient light intensity and ambient pH value input by the model and output the corresponding prediction vector Y; the prediction vector Y includes the predicted absorbance and the predicted wavelength, and the predicted absorbance and the predicted wavelength correspond to the maximum absorbance and the maximum absorption wavelength respectively; The data acquisition module is used to collect data on the molecular structure data of multiple types of organic molecules of organic photovoltaic materials and the maximum absorption peak information of each organic molecule under different environmental conditions through experiments, simulation calculations and big data acquisition, and to form a corresponding first data set based on the collected data; the environmental conditions include temperature, light intensity and pH value; The model training module is used to perform model training on the maximum absorption peak prediction model based on the first data set; The model application module is used to receive the first molecular structure and the first environmental parameter set input by the user after the model training is completed; and use the maximum absorption peak prediction model to predict the maximum absorption peak information of the first molecular structure under all the first environmental parameters in the first environmental parameter set to obtain the corresponding first prediction vector set and feedback it to the current user; the first environmental parameter set includes multiple first environmental parameters; the first environmental parameters include temperature, light intensity and pH value; the first prediction vector set includes multiple first prediction vectors; the first prediction vector consists of a first absorbance and a first wavelength, and corresponds one-to-one to the first environmental parameter.

7. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 5; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 5.

Images