Generation method and device of similar chemical reaction
By generating similar chemical reactions through generative models, the problems of time-consuming and labor-intensive technologies and low success rates in existing technologies are solved, and efficient and accurate chemical reaction generation is achieved.
Patent Information
- Application Number
- CN202510658646.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-05-21
- Filing Date
- 2025-05-21
- Publication Date
- 2025-09-19
AI Technical Summary
The existing methods for generating similar chemical reactions are time-consuming, labor-intensive, and have a low success rate. They rely on manual labor and experience, resulting in low efficiency.
The reference chemical reaction is converted into a reference vector through a generative model, and the similarity and yield are calculated using a scoring function to generate the target chemical reaction, reducing human participation and improving generation efficiency and success rate.
The generation model does not require human participation, which improves the generation efficiency, ensures the similarity and success rate between the target chemical reaction and the reference chemical reaction, and reduces the impact of chemical reaction characteristics on the generation results.
Smart Images
Figure CN120673873A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of chemical informatics, and in particular to a method and device for generating similar chemical reactions. Background Art
[0002] Whether a specific chemical reaction can occur depends on the suitability of the reaction template for a specific molecule under specific conditions. Small structural changes in the reactants can lead to significant changes in reactivity and yield, posing a significant challenge to chemical synthesis planning. To synthesize a specific compound, a chemist, after identifying potential starting materials based on their knowledge and experience, typically searches for similar reactions in relevant literature to estimate the likelihood of obtaining the product. However, due to the vastness of the reaction space, finding similar reactions can be challenging.
[0003] Currently, there are two methods for generating similar chemical reactions: the first involves manual enumeration based on chemists' experience and rules, taking into account chemical knowledge such as steric hindrance and electronic effects; the second involves collecting a large amount of existing reaction data and templates, and then using algorithms to randomly generate potential chemical reactions. However, the first method relies on manual effort and experience, consuming a significant amount of chemists' time and effort. Furthermore, the second method has a lower success rate for generated chemical reactions due to the characteristics of chemical reactions, such as the fact that small structural changes in reactants can lead to significant changes in reaction activity and yield, and that even very different reactant and product structures can be explained by the same reaction mechanism. Summary of the Invention
[0004] In view of this, the purpose of the embodiments of the present application is to provide a method and device for generating similar chemical reactions, so as to solve the problems of time-consuming and labor-intensive generation and low success rate in the prior art of generating chemical reactions.
[0005] In a first aspect, an embodiment of the present application provides a method for generating a similar chemical reaction, which includes:
[0006] Obtain reference chemical reactions;
[0007] converting the reference chemical reaction into a reference vector;
[0008] The reference vector is calculated by a generative model to generate a target chemical reaction; wherein the generative model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0009] In a possible implementation, calculating the reference vector by generating a model to generate a target chemical reaction includes:
[0010] Generate universal vectors based on the reference vectors using the generative model, each universal vector corresponding to a universal chemical reaction;
[0011] The target chemical reaction is determined from the general chemical reactions based on the scoring function.
[0012] In a possible implementation, determining the target chemical reaction from the general chemical reactions based on the scoring function includes:
[0013] Taking the reference vector as an anchor point, selecting the candidate vector from the space vector corresponding to the chemical space formed by the universal chemical reaction;
[0014] Based on the scoring function, selecting a target vector from the candidate vectors;
[0015] The target vector is converted into a target chemical reaction.
[0016] In a possible implementation, the selecting the candidate vector from the space vector formed by the general vector using the reference vector as an anchor point includes:
[0017] The reference vector and the spatial vector are input into a selection model so that the selection model calculates the reference vector and the spatial vector to obtain the candidate vector, and the selection model includes an autoencoder or a diffusion model.
[0018] In a possible implementation, screening out a target vector from the candidate vectors based on the scoring function includes:
[0019] calculating, for each candidate chemical reaction, a similarity between the candidate chemical reaction and the reference chemical reaction using the scoring function, and predicting a yield of each candidate chemical reaction, wherein the yield includes at least one or more of an existence probability, a reaction degree, and a reaction duration of the candidate chemical reaction;
[0020] A target vector is selected from the candidate vectors based on the similarity and the yield.
[0021] In a possible implementation, the selecting a target vector from the candidate vectors based on the similarity and the yield includes:
[0022] Determining a confidence value for each of the candidate chemical reactions based on the similarity and the yield;
[0023] From the plurality of candidate chemical reactions, candidate vectors corresponding to confidence values greater than a threshold are selected as the target vectors.
[0024] In one possible implementation, the generative model is trained by the following steps:
[0025] Build an initial model;
[0026] Training the initial model using training samples so that the initial model obtains chemical reaction rules, wherein the training samples include multiple chemical reaction samples;
[0027] Based on the chemical reaction rules, the model parameters of the initial model are adjusted to obtain the generation model.
[0028] In a second aspect, an embodiment of the present application further provides a device for generating a similar chemical reaction, comprising:
[0029] an acquisition module configured to acquire a reference chemical reaction;
[0030] a conversion module configured to convert the reference chemical reaction into a reference vector;
[0031] A generation module is configured to calculate the reference vector through a generation model to generate a target chemical reaction; wherein the generation model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0032] In a third aspect, an embodiment of the present application further provides a storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method for generating a similar chemical reaction as described in any one of the above are executed.
[0033] In a fourth aspect, an embodiment of the present application further provides an electronic device, comprising: a processor and a memory, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through a bus, and when the machine-readable instructions are executed by the processor, the steps of the method for generating similar chemical reactions as described in any one of the above are performed.
[0034] The present application generates a target chemical reaction similar to a reference chemical reaction by generating a model. Compared with the prior art which relies on manual and empirical determination of chemical reactions, this application does not require human involvement, saves time and effort, and improves generation efficiency. At the same time, generating the target chemical reaction based on similarity and yield not only reduces the impact of the chemical reaction characteristics on the generation results, ensuring the success rate of the generated target chemical reaction, but also ensures a high degree of similarity between the target chemical reaction and the reference chemical reaction, thereby improving the confidence level of the target chemical reaction.
[0035] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0037] Figure 1 A flow chart showing a method for generating similar chemical reactions provided in this application is shown;
[0038] Figure 2 A flowchart of determining a target chemical reaction from a general chemical reaction based on a scoring function in the method for generating similar chemical reactions provided by the present application is shown;
[0039] Figure 3 A flow chart showing the training and generation model in the method for generating similar chemical reactions provided in this application is shown;
[0040] Figure 4 A schematic diagram of the structure of a device for generating similar chemical reactions provided in this application is shown;
[0041] Figure 5 A schematic structural diagram of the electronic device provided in this application is shown. DETAILED DESCRIPTION
[0042] Various aspects and features of the present application are described herein with reference to the accompanying drawings.
[0043] It should be understood that various modifications may be made to the embodiments of the present application. Therefore, the above description should not be considered as limiting, but merely as an example of an embodiment. Other modifications within the scope and spirit of the present application will occur to those skilled in the art.
[0044] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.
[0045] These and other characteristics of the present application will become apparent from the following description of a preferred form of embodiment given as a non-limiting example with reference to the accompanying drawings.
[0046] It should also be understood that although the present application has been described with reference to certain specific examples, those skilled in the art will be able to implement many other equivalent forms of the present application that have the features described in the claims and are therefore within the scope of protection defined thereby.
[0047] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.
[0048] Specific embodiments of the present application will be described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments described are merely examples of the present application and may be implemented in a variety of ways. Familiar and / or repetitive functions and structures are not described in detail to avoid obscuring the present application with unnecessary or redundant details. Therefore, the specific structural and functional details described herein are not intended to be limiting, but rather serve merely as a basis and representative basis for the claims to teach those skilled in the art to variously utilize the present application with substantially any suitable detailed structure.
[0049] This specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may all refer to one or more of the same or different embodiments according to the present application.
[0050] First, in order to facilitate understanding of this application, a method for generating a similar chemical reaction provided by this application is first introduced in detail. Figure 1 As shown, a method for generating a similar chemical reaction provided in an embodiment of the present application, the specific steps include S101-S103.
[0051] S101, obtaining a reference chemical reaction.
[0052] Here, the reference chemical reaction is an existing chemical reaction, such as a chemical reaction that appears in literature or papers, a chemical reaction obtained during an experiment, etc.
[0053] The reference chemical reaction includes raw materials, reaction type, reaction conditions and generated chemicals.
[0054] S102, converting the reference chemical reaction into a reference vector.
[0055] After obtaining the reference chemical reaction, the reference chemical reaction is converted into a reference vector, which is a vector that can be recognized, calculated, and processed by an algorithm or model.
[0056] S103, calculating the reference vector through a generative model to generate a target chemical reaction; wherein the generative model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0057] Optionally, the reference vector is input into a generative model, and the reference vector is calculated by the generative model to generate a target chemical reaction. As one example, when the reference vector is calculated by the generative model to generate a target chemical reaction, the generative model first generates universal vectors based on the reference vector, each universal vector corresponding to a universal chemical reaction. Then, based on a scoring function, the target chemical reaction is determined from the universal chemical reactions. Based on this, the target chemical reaction can be one or more.
[0058] The generative model includes a scoring function, which can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0059] After obtaining the universal chemical reaction, the candidate chemical reaction is selected from the chemical space formed by the universal chemical reaction using the reference chemical reaction as an anchor point. Figure 2 A flow chart of a method for determining a target chemical reaction from general chemical reactions based on a scoring function is shown, and specific steps include S201-S203.
[0060] S201 , using the reference vector as an anchor point, selecting a candidate vector from the space vectors corresponding to the chemical space formed by the universal chemical reaction.
[0061] S202: Filter out a target vector from the candidate vectors based on the scoring function.
[0062] S203, converting the target vector into a target chemical reaction.
[0063] In a specific implementation, all universal chemical reactions can form a chemical space. Therefore, after generating a universal vector, that is, obtaining a universal chemical reaction, a chemical space is formed based on the universal chemical reaction, and the chemical space is converted into a space vector.
[0064] Afterwards, the reference vector and the spatial vector are input into the selection model, where the selection model includes an autoencoder or a diffusion model. Of course, the selection model also requires a large number of training samples for training. The training samples include reference vector samples, spatial vector samples and candidate vector samples. The specific training process will not be described in detail here.
[0065] After receiving the reference vector and the spatial vector, the selection model calculates the reference vector and the spatial vector to obtain a candidate vector. The candidate vector is a vector in the spatial vector whose distance from the reference vector is less than a certain distance, that is, the candidate chemical reaction corresponding to the candidate vector has a close correlation with the reference chemical reaction. The close correlation between the candidate chemical reaction and the reference chemical reaction includes at least a small difference in the raw material structure between the candidate chemical reaction and the reference chemical reaction, and a small difference in the reaction mechanism between the candidate chemical reaction and the reference chemical reaction.
[0066] After obtaining the candidate vectors, the target vector is screened out from the candidate vectors based on the scoring function, and the target vector is converted to obtain the target chemical reaction, which improves the accuracy of the target chemical reaction to a certain extent.
[0067] Optionally, when selecting a target vector from candidate vectors based on a scoring function, the scoring function is first used to calculate the similarity between each candidate chemical reaction and a reference chemical reaction, and to predict the yield of each candidate chemical reaction.
[0068] Wherein, the similarity is determined by at least one or more of the raw material sub-similarity, reaction type sub-similarity, reaction condition sub-similarity, and chemical sub-similarity between the candidate chemical reaction and the reference chemical reaction. Optionally, the raw material sub-similarity, reaction type sub-similarity, reaction condition sub-similarity, and chemical sub-similarity each correspond to a weight value. When determining the similarity between each candidate chemical reaction and the reference chemical reaction, for each candidate chemical reaction, the similarity between the candidate chemical reaction and the reference chemical reaction is calculated, that is, the raw material sub-similarity, reaction type sub-similarity, reaction condition sub-similarity, and chemical sub-similarity between the candidate chemical reaction and the reference chemical reaction are calculated, and the similarity between the candidate chemical reaction and the reference chemical reaction is calculated based on the raw material sub-similarity, reaction type sub-similarity, reaction condition sub-similarity, chemical sub-similarity, and the corresponding weight values.
[0069] The yield of a candidate chemical reaction includes at least one or more of the probability of existence of the candidate chemical reaction, the degree of reaction, and the reaction duration. In a specific implementation, when determining the yield of each candidate chemical reaction, a prediction model is used to predict the yield of each candidate chemical reaction. The prediction model is pre-trained, using existing chemical reactions, such as reference chemical reactions, and existing chemical reactions as training samples to train the prediction model. Existing chemical reactions are chemical reactions that cannot occur or complete.
[0070] As one example, when the yield of a candidate chemical reaction includes the probability of existence of the candidate chemical reaction, the prediction model only needs to predict a numerical value, and the yield of the candidate chemical reaction can be set as a percentage. For example, the prediction model predicts that the yield of the candidate chemical reaction is 95%, 78%, 43%, etc. When the yield is 95%, it represents that the probability of existence of the candidate chemical reaction is 95%. When the yield of a candidate chemical reaction includes the existence probability, reaction degree and reaction time of the candidate chemical reaction, the prediction model needs to predict three values, and then calculate the three predicted values to obtain the yield of the candidate chemical reaction. At this time, the yield of the candidate chemical reaction can be set as a score. For example, the prediction model predicts the existence probability of the candidate chemical reaction, the first prediction value corresponding to the reaction degree, and the second prediction value corresponding to the reaction time. The existence probability, the first prediction value and the second prediction value are calculated to obtain a score, which is the yield of the candidate chemical reaction. The yield of the candidate chemical reaction is 23 points, 75 points, 81 points, etc. When the yield is 81, it indicates that the existence probability of the candidate chemical reaction is high, the reaction degree is high, and the reaction time is short.
[0071] Accordingly, the yield of a candidate chemical reaction includes one or more of the probability of existence, reaction degree, and reaction duration of the candidate chemical reaction. Other examples will not be described in detail in this application.
[0072] After obtaining the similarity between the candidate chemical reaction and the reference chemical reaction, and predicting the yield of each candidate chemical reaction, the target vector is screened from the candidate vectors based on the similarity and yield to obtain the target chemical reaction corresponding to the target vector.
[0073] As one example, when generating at least one target chemical reaction based on similarity and yield, a confidence value is first determined for each candidate chemical reaction based on the similarity and yield. A higher confidence value indicates a closer correlation between the corresponding candidate chemical reaction and the reference chemical reaction, meaning that the corresponding candidate chemical reaction is a known chemical reaction and is relatively similar to the reference chemical reaction.
[0074] After obtaining the confidence value of each candidate chemical reaction, candidate vectors corresponding to confidence values greater than a threshold are selected from multiple candidate chemical reactions as target vectors, where the threshold is set by the user based on the current scenario and the currently existing chemical reactions.
[0075] Of course, a preset number of candidate vectors can also be selected as target vectors in descending order of confidence values. The preset number is determined based on the user's actual needs and / or the current scenario. For example, for a reference chemical reaction with two raw materials, the threshold is determined to be 10; for a reference chemical reaction with three raw materials, the threshold is determined to be 15, etc.
[0076] After screening and obtaining the target vector, i.e., the target chemical reaction, the user can use the target chemical reaction as a reference for determining new chemical reactions, for example, to verify the reliability of the target chemical reaction.
[0077] The present application generates a target chemical reaction similar to a reference chemical reaction by generating a model. Compared with the prior art which relies on manual and empirical determination of chemical reactions, this application does not require human involvement, saves time and effort, and improves generation efficiency. At the same time, generating the target chemical reaction based on similarity and yield not only reduces the impact of the chemical reaction characteristics on the generation results, ensuring the success rate of the generated target chemical reaction, but also ensures a high degree of similarity between the target chemical reaction and the reference chemical reaction, thereby improving the confidence level of the target chemical reaction.
[0078] Optionally, Figure 3 A flow chart of a method for training and obtaining a generative model is shown, wherein the specific steps include S301-S303.
[0079] S301, constructing an initial model.
[0080] S302: Train the initial model using training samples so that the initial model obtains chemical reaction rules. The training samples include multiple chemical reaction samples.
[0081] S303, based on the chemical reaction rules, the initial model adjusts the model parameters to obtain a generation model.
[0082] Optionally, an initial model is pre-constructed based on general chemical rules and a scoring function. Specifically, model parameters of the initial model are constructed based on the general chemical rules and the scoring function. The initial model is generated by combining the constructed model parameters with fine-tuning parameters, where the fine-tuning parameters are adjustable parameters in the initial model. The scoring function is determined based on similarity samples and yield samples between the chemical reaction sample and its corresponding similar chemical reaction. The scoring function enables the generated model to generate similar chemical reactions with high similarity to the chemical reaction sample and similar chemical reactions with high yield.
[0083] The initial model is then trained using training samples to enable the initial model to acquire chemical reaction rules. Specifically, the training samples include multiple chemical reaction samples, which can be character strings or molecular graphs corresponding to chemical reactions. The character strings or molecular graphs corresponding to the chemical reactions are converted into chemical vectors and input into the initial model to train the initial model. This allows the initial model to acquire chemical reaction rules, and then the model parameters of the initial model are adjusted based on the chemical reactions to obtain a generated model.
[0084] Reaction rules are the reaction rules followed when raw materials undergo chemical reactions, such as steric effects and electronic effects. Compared to general chemical rules, chemical reaction rules better reflect the actual reaction rules followed by chemical reactions. Generative models based on chemical reaction rules and general chemical rules generate chemical reactions with higher yields, thereby improving the success rate of the generated target chemical reactions.
[0085] Based on the same inventive concept, the second aspect of this application also provides a device for generating similar chemical reactions. Since the principle of solving the problem by the device in this application is similar to the above-mentioned method for generating similar chemical reactions in this application, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be repeated.
[0086] See also Figure 4 As shown, the generation device of similar chemical reactions includes:
[0087] an acquisition module 401 configured to acquire a reference chemical reaction;
[0088] A conversion module 402 configured to convert the reference chemical reaction into a reference vector;
[0089] A generation module 403 is configured to calculate the reference vector through a generation model to generate a target chemical reaction; wherein the generation model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0090] In another embodiment, the generating module 403 is specifically configured as follows:
[0091] Generate universal vectors based on the reference vectors using the generative model, each universal vector corresponding to a universal chemical reaction;
[0092] The target chemical reaction is determined from the general chemical reactions based on the scoring function.
[0093] In another embodiment, when determining the target chemical reaction from the general chemical reactions based on the scoring function, the generating module 403 is configured to:
[0094] Taking the reference vector as an anchor point, selecting the candidate vector from the space vector corresponding to the chemical space formed by the universal chemical reaction;
[0095] Based on the scoring function, selecting a target vector from the candidate vectors;
[0096] The target vector is converted into a target chemical reaction.
[0097] In yet another embodiment, when the generating module 403 selects the candidate vector from the space vector formed by the general vector with the reference vector as the anchor point, it is configured as follows:
[0098] The reference vector and the spatial vector are input into a selection model so that the selection model calculates the reference vector and the spatial vector to obtain the candidate vector, and the selection model includes an autoencoder or a diffusion model.
[0099] In yet another embodiment, when the generating module 403 selects the target vector from the candidate vectors based on the scoring function, it is configured to:
[0100] calculating, for each candidate chemical reaction, a similarity between the candidate chemical reaction and the reference chemical reaction using the scoring function, and predicting a yield of each candidate chemical reaction, wherein the yield includes at least one or more of an existence probability, a reaction degree, and a reaction duration of the candidate chemical reaction;
[0101] A target vector is selected from the candidate vectors based on the similarity and the yield.
[0102] In yet another embodiment, when the generating module 403 selects the target vector from the candidate vectors based on the similarity and the yield, the generating module 403 is configured to:
[0103] Determining a confidence value for each of the candidate chemical reactions based on the similarity and the yield;
[0104] From the plurality of candidate chemical reactions, candidate vectors corresponding to confidence values greater than a threshold are selected as the target vectors.
[0105] In yet another embodiment, the device for generating similar chemical reactions further includes a construction module 404 configured to:
[0106] Build an initial model;
[0107] Training the initial model using training samples so that the initial model obtains chemical reaction rules, wherein the training samples include multiple chemical reaction samples;
[0108] Based on the chemical reaction rules, the model parameters of the initial model are adjusted to obtain the generation model.
[0109] The present application generates a target chemical reaction similar to a reference chemical reaction by generating a model. Compared with the prior art which relies on manual and empirical determination of chemical reactions, this application does not require human involvement, saves time and effort, and improves generation efficiency. At the same time, generating the target chemical reaction based on similarity and yield not only reduces the impact of the chemical reaction characteristics on the generation results, ensuring the success rate of the generated target chemical reaction, but also ensures a high degree of similarity between the target chemical reaction and the reference chemical reaction, thereby improving the confidence level of the target chemical reaction.
[0110] A third aspect of the present application further provides a storage medium, which is a computer-readable medium and stores a computer program. When the computer program is executed by a processor, the method provided in any embodiment of the present application is implemented, including the following steps:
[0111] S11, obtain reference chemical reaction;
[0112] S12, converting the reference chemical reaction into a reference vector;
[0113] S13, calculating the reference vector through a generation model to generate a target chemical reaction; wherein the generation model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0114] The present application generates a target chemical reaction similar to a reference chemical reaction by generating a model. Compared with the prior art which relies on manual and empirical determination of chemical reactions, this application does not require human involvement, saves time and effort, and improves generation efficiency. At the same time, generating the target chemical reaction based on similarity and yield not only reduces the impact of the chemical reaction characteristics on the generation results, ensuring the success rate of the generated target chemical reaction, but also ensures a high degree of similarity between the target chemical reaction and the reference chemical reaction, thereby improving the confidence level of the target chemical reaction.
[0115] It should be noted that the storage medium mentioned above in this application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any storage medium other than a computer-readable storage medium that can transmit, propagate, or convey a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the storage medium may be conveyed using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), or any suitable combination thereof.
[0116] The fourth aspect of the present application also provides an electronic device, such as Figure 5 As shown, the electronic device includes at least a memory 501 and a processor 502. The memory 501 stores a computer program. The processor 502 implements the method provided by any embodiment of the present application when executing the computer program on the memory 501. Exemplarily, the method executed by the electronic device computer program is as follows:
[0117] S21, obtain reference chemical reaction;
[0118] S22, converting the reference chemical reaction into a reference vector;
[0119] S23, calculating the reference vector through a generation model to generate a target chemical reaction; wherein the generation model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
[0120] The present application generates a target chemical reaction similar to a reference chemical reaction by generating a model. Compared with the prior art which relies on manual and empirical determination of chemical reactions, this application does not require human involvement, saves time and effort, and improves generation efficiency. At the same time, generating the target chemical reaction based on similarity and yield not only reduces the impact of the chemical reaction characteristics on the generation results, ensuring the success rate of the generated target chemical reaction, but also ensures a high degree of similarity between the target chemical reaction and the reference chemical reaction, thereby improving the confidence level of the target chemical reaction.
[0121] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.
[0122] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.
[0123] In addition, although adopting specific order to describe each operation, this should not be interpreted as requiring these operations to be executed in the specific order shown or in sequential order.Under certain environment, multitasking and parallel processing may be advantageous.Similarly, although comprising some specific implementation details in the above discussion, these should not be interpreted as limiting the scope of the application.Some features described in the context of separate embodiment can also be implemented in a single embodiment in combination.On the contrary, the various features described in the context of a single embodiment also can be implemented in multiple embodiments individually or in the mode of any suitable subcombination.
[0124] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims.
[0125] The above describes in detail several embodiments of the present application, but the present application is not limited to these specific embodiments. Based on the concept of the present application, technicians in this field can make various variations and modifications to the embodiments, and these variations and modifications should fall within the scope of protection required by the present application.
Claims
1. A method for generating a similar chemical reaction, characterized in that: include: Obtain reference chemical reactions; converting the reference chemical reaction into a reference vector; The reference vector is calculated by a generative model to generate a target chemical reaction; wherein the generative model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
2. The generation method according to claim 1, characterized in that The step of calculating the reference vector by generating a model to generate a target chemical reaction includes: Generate universal vectors based on the reference vectors using the generative model, each universal vector corresponding to a universal chemical reaction; The target chemical reaction is determined from the general chemical reactions based on the scoring function.
3. The generation method according to claim 2, characterized in that Determining the target chemical reaction from the general chemical reactions based on the scoring function includes: Taking the reference vector as an anchor point, selecting the candidate vector from the space vector corresponding to the chemical space formed by the universal chemical reaction; Based on the scoring function, selecting a target vector from the candidate vectors; The target vector is converted into a target chemical reaction.
4. The generation method according to claim 3, characterized in that The selecting the candidate vector from the space vectors formed by the general vector using the reference vector as an anchor point includes: The reference vector and the spatial vector are input into a selection model so that the selection model calculates the reference vector and the spatial vector to obtain the candidate vector, and the selection model includes an autoencoder or a diffusion model.
5. The generation method according to claim 3, characterized in that The step of selecting a target vector from the candidate vectors based on the scoring function includes: calculating, for each candidate chemical reaction, a similarity between the candidate chemical reaction and the reference chemical reaction using the scoring function, and predicting a yield of each candidate chemical reaction, wherein the yield includes at least one or more of an existence probability, a reaction degree, and a reaction duration of the candidate chemical reaction; A target vector is selected from the candidate vectors based on the similarity and the yield.
6. The generation method according to claim 5, characterized in that The selecting a target vector from the candidate vectors based on the similarity and the yield includes: Determining a confidence value for each of the candidate chemical reactions based on the similarity and the yield; From the plurality of candidate chemical reactions, candidate vectors corresponding to confidence values greater than a threshold are selected as the target vectors.
7. The generation method according to any one of claims 1 to 6, characterized in that: The generative model is trained by the following steps: Build an initial model; Training the initial model using training samples so that the initial model obtains chemical reaction rules, wherein the training samples include multiple chemical reaction samples; Based on the chemical reaction rules, the model parameters of the initial model are adjusted to obtain the generation model.
8. A device for generating similar chemical reactions, characterized in that: include: an acquisition module configured to acquire a reference chemical reaction; a conversion module configured to convert the reference chemical reaction into a reference vector; A generation module is configured to calculate the reference vector through a generation model to generate a target chemical reaction; wherein the generation model includes a scoring function, and the scoring function can obtain the similarity between the reference chemical reaction and the target chemical reaction and the yield of the target chemical reaction.
9. A storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the method for generating a similar chemical reaction as claimed in any one of claims 1 to 7.
10. An electronic device, characterized in that: include: A processor and a memory, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate via a bus. When the machine-readable instructions are executed by the processor, the steps of the method for generating a similar chemical reaction as described in any one of claims 1 to 7 are performed.