CN117172256B - Laboratory management method and system based on modularized setting - Google Patents

Abstract

A laboratory management method and system based on modular setup is disclosed. The method comprises the following steps: configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a symbol comparison module and an ending mode module; and verifying the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects. In this way, the module combination may be semantically understood to determine whether the module combination has a logical defect.

Description

Laboratory management method and system based on modularized setting

Technical Field

The present application relates to the field of laboratory management, and more particularly, to a laboratory management method and system based on a modular setup.

Background

Traditional laboratory management is usually based on manual management, and experimental data is usually recorded by adopting manual recording or electronic forms, so that the problems of experimental errors and inaccurate data are easily caused due to lack of standardized experimental procedures and quality control standards. Moreover, conventional laboratory management cannot perform physical experiments and tests to optimize laboratory management process flows until no experimental physical production line is constructed.

Thus, an optimized laboratory management solution is desired.

Disclosure of Invention

The present application has been made in order to solve the above technical problems. The embodiment of the application provides a laboratory management method and system based on modular arrangement. The method comprises the following steps: configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a symbol comparison module and an ending mode module; and verifying the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects. In this way, the module combination may be semantically understood to determine whether the module combination has a logical defect.

According to one aspect of the present application, there is provided a laboratory management method based on a modular setup, comprising:

configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a symbol comparison module and an ending mode module; and

and verifying the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects.

In the above laboratory management method based on modular arrangement, verifying all module schemes required by the production link to determine whether a plurality of modules required by the production link have logic defects, including:

acquiring text descriptions of each module in the plurality of modules;

the text description of each module is passed through a Word2Vec model to obtain a sequence of module description embedded vectors;

passing the sequence of module description embedded vectors through a module sequence semantic understand based on a converter module to obtain a plurality of context module description semantic understand feature vectors;

cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector;

the context module description semantic understanding feature vectors pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector;

fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and

and passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the module combination has a logic defect or not.

In the above-mentioned laboratory management method based on modular setup, passing the sequence of module description embedded vectors through a module sequence semantic understand device based on a converter module to obtain a plurality of context module description semantic understand feature vectors, including:

one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector;

calculating the product between the global module description feature vector and the transpose vector of each module description embedded vector in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices;

respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;

obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and

and weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the context module description semantic understanding feature vectors.

In the above laboratory management method based on modular setup, cascading the plurality of context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector includes:

cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector;

wherein, the cascade formula is:

V ₁ ＝Concat[V _a1 ，V _a2 ，V _a3 ……V _an ]

wherein V is _a1 ，V _a2 ，V _a3 ……V _an Representing the plurality of context modules describing a semantic understanding feature vector, concat [ &]Representing a cascade function, V ₁ Representing the first scale module combined semantic understanding feature vector.

In the above laboratory management method based on modular arrangement, fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector includes:

fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector by the following fusion formula to obtain the module combination semantic understanding feature vector;

wherein, the fusion formula is:

Wherein V is ₁ Is the first scale module combined semantic understanding feature vector, V ₂ Is the second scale module combined semantic understanding feature vector,<<s sum of>>s represents shifting the feature vector left by s bits and right by s bits, respectively, round is a rounding functionThe number of the product is the number,is the average value of all feature values of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and II is II ₁ Represents a norm, d (V) ₁ ,V ₂ ) Is the distance between the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and log is a logarithmic function based on 2, < + >>For vector subtraction, α and β are weight super parameters.

In the above laboratory management method based on modular arrangement, the module combination semantic understanding feature vector is passed through a classifier to obtain a classification result, where the classification result is used to indicate whether a module combination has a logic defect, and the method includes:

performing full-connection coding on the module combination semantic understanding feature vector by using a full-connection layer of the classifier to obtain a coding classification feature vector; and

and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

According to another aspect of the present application, there is provided a laboratory management system based on a modular setup, comprising:

the module configuration unit is used for configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a comparison symbol module and an ending mode module; and

and the verification judging unit is used for verifying the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects or not.

In the above laboratory management system based on modular arrangement, the verification judging unit includes:

a text description obtaining subunit, configured to obtain text descriptions of each of the plurality of modules;

the embedded coding subunit is used for enabling the text description of each module to pass through a Word2Vec model to obtain a sequence of module description embedded vectors;

a module sequence semantic understanding subunit, configured to pass the sequence of module description embedded vectors through a module sequence semantic understanding device based on a converter module to obtain a plurality of context module description semantic understanding feature vectors;

the cascading subunit is used for cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector;

The convolution coding subunit is used for enabling the context module description semantic understanding feature vectors to pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector;

the fusion subunit is used for fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and

and the classification subunit is used for passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the module combination has a logic defect or not.

In the above laboratory management system based on modular arrangement, the module sequence semantic understanding subunit is configured to:

one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector;

calculating the product between the global module description feature vector and the transpose vector of each module description embedded vector in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices;

respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;

Obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and

and weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the context module description semantic understanding feature vectors.

In the above laboratory management system based on a modular arrangement, the cascade subunit is configured to:

cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector;

wherein, the cascade formula is:

V ₁ ＝Concat[V _a1 ，V _a2 ，V _a3 ……V _an ]

wherein V is _a1 ，V _a2 ，V _a3 ……V _an Representing the plurality of context modules describing a semantic understanding feature vector, concat [ &]Representing a cascade function, V ₁ Representing the first scale module combined semantic understanding feature vector.

Compared with the prior art, the laboratory management method and system based on the modularized arrangement, which are provided by the application, comprise the following steps: configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a symbol comparison module and an ending mode module; and verifying the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects. In this way, the module combination may be semantically understood to determine whether the module combination has a logical defect.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly introduced below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. The following drawings are not intended to be drawn to scale, with emphasis instead being placed upon illustrating the principles of the present application.

FIG. 1 is a flow chart of a method of laboratory management based on modular setup according to an embodiment of the present application.

Fig. 2 is an application scenario diagram of substep S120 of a laboratory management method based on modular settings according to an embodiment of the present application.

Fig. 3 is a flowchart of substep S120 of the laboratory management method based on the modular setup according to an embodiment of the present application.

Fig. 4 is an architectural diagram of substep S120 of a laboratory management method based on modular setup according to an embodiment of the present application.

Fig. 5 is a flowchart of substep S123 of the laboratory management method based on the modular setup according to an embodiment of the present application.

Fig. 6 is a flowchart of substep S127 of the laboratory management method based on the modular setup according to an embodiment of the present application.

Fig. 7 is a block diagram of a laboratory management system based on a modular setup according to an embodiment of the present application.

Fig. 8 is a schematic flow chart of experimental setup and experimental implementation according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present application without making any inventive effort, are also within the scope of the present application.

As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.

Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.

Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.

In view of the above technical problems, the technical idea of the present application is to provide a laboratory management solution based on a modular arrangement. Specifically, the laboratory management system based on the modular arrangement is a laboratory management system based on a process flow. The pre-configuration of the laboratory management system based on the modularized arrangement comprises enterprise factory management, data sources, modules, module schemes, unit schemes and action libraries. In the experimental stage, a user can select templates, schemes and experiments and submit the templates, schemes and experiments to enter a 'do' process. The process comprises bill of materials, steps, stages, processes, modules and parameters, and a user can watch the online edition experiment dynamics. The experimental report may generate an offline version of the report including process node status, equipment bill of materials, production records, assay reports, video playback, and data playback. In the experimental stage, the user needs to perform distribution, auditing, starting, verification requesting and other operations.

In particular, in the technical solution of the present application, the laboratory management system based on modular setup supports configuring all module types for a specific production link for data simulation experiments and optimization. That is, in the early configuration stage, the laboratory management system based on the modular arrangement can configure all model types of the production link, that is, configure a combination of functional modules required for a specific production link, where the functional modules include a parameter control action module, a start selection module, an end action module, a comparison symbol module, and an end mode module, and the reference control action module includes a rotation speed control module, a temperature control module, an angle control module, and the like.

Accordingly, as shown in fig. 1, a laboratory management method based on modular setup provided in the present application includes: s110, configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a comparison symbol module and an ending mode module; and S120, checking a plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects.

It should be appreciated that in configuring all model types of a production link for data simulation experiments and module combination optimization, all model types configured may have logic deviations (e.g., end-then-start actions) to produce unnecessary data simulation and test procedures. Therefore, in the technical solution of the present application, the laboratory management solution based on the modular arrangement further supports semantic understanding of the module combination to determine whether the module combination has a logic defect.

Specifically, a textual description of each of the plurality of modules is first obtained. As described above, in the laboratory management method based on the modular arrangement, each module has its specific functions and roles. Therefore, in order to perform semantic understanding on the modules and determine whether the module combination has a logic defect, a text description of each module needs to be acquired first, where the text description of each module may include information such as a name, a function, input and output of the module.

The text description of each module is then passed through the Word2Vec model to obtain a sequence of module description embedded vectors. It should be appreciated that the textual descriptions of the individual modules are textual data, which is unstructured data, and therefore, the textual descriptions of the individual modules need to be first converted into a structured language that can be recognized and processed by a computer.

Specifically, in the technical scheme of the application, text description of each module can be converted into vector representation through a Word2Vec model, so that subsequent semantic understanding and calculation are facilitated. One of ordinary skill in the art will appreciate that the Word2Vec model is a neural network-based language model that maps each Word or phrase to a vector in a high-dimensional vector space that captures semantic relationships between the words or phrases.

The sequence of module description embedded vectors is then passed through a module sequence semantic understand based on the converter module to obtain a plurality of context module description semantic understand feature vectors. In the technical solution of the present application, semantic understanding needs to be performed on the module combination of the plurality of functional modules to determine whether the module combination has a semantic logic defect. Thus, after obtaining the sequence of module description embedded vectors, passing the sequence of module description embedded vectors through a converter module-based module sequence semantic interpreter to obtain the plurality of context module description semantic understanding feature vectors, wherein the converter module-based module sequence semantic interpreter utilizes a transducer mechanism to perform global context semantic encoding of the sequence of module description embedded vectors based on a self-attention mechanism to capture semantic association information, such as, for example, dependencies and interactions between text descriptions of individual modules in the module combination.

Furthermore, the plurality of context module description semantic understanding feature vectors are concatenated to obtain a first scale module combination semantic understanding feature vector. That is, the respective local feature distributions of the plurality of context module descriptive semantic understanding feature vectors are aggregated to obtain a global semantic feature representation representing the module combination, i.e. the first scale module combination semantic understanding feature vector. That is, by concatenating the plurality of context module description semantic understanding feature vectors, semantic information of the plurality of modules may be integrated, thereby obtaining a more comprehensive and accurate module combination semantic understanding feature vector.

While the module sequence semantic understander based on the converter module can capture global context dependent information (i.e., long-range dependent information) of each module description embedded vector in the sequence of module description embedded vectors by using a transducer mechanism, it is insufficient in terms of extraction capability of short-range and medium-range semantic information dependent information. It should be understood that in the technical solution of the present application, the association strength between modules with different distances among the plurality of modules is different, and generally, the stronger the association strength between modules with similar distances is. In order to make up for semantic information, after the semantic understanding feature vectors described by the context modules are obtained, the semantic understanding feature vectors described by the context modules are passed through a two-way long-short-term memory neural network model to obtain a second-scale module combined semantic understanding feature vector. Accordingly, short-range semantic information and medium-range semantic information among various modules in the module combination can be better captured by using the two-way long-short-term memory neural network model. In particular, the two-way long and short term memory neural network model may consider both forward and backward context information to better understand the dependencies and interactions between modules.

And after the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are obtained, fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector. It should be appreciated that the first-scale module-combination semantic understanding feature vector is obtained by concatenating a sequence of a plurality of module description embedding vectors, which can capture long Cheng Yuyi information between the respective modules in the module combination, and the second-scale module-combination semantic understanding feature vector is obtained by passing a plurality of context module description semantic understanding feature vectors through a two-way long-short-term memory neural network model, which can better capture short-and medium-term dependent semantic information between the respective modules in the module combination. Therefore, by fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector, the module combination semantic understanding feature vector can be obtained more comprehensively and accurately, and accordingly the accuracy of subsequent classification and judgment is improved.

The module combination semantic understanding feature vector is then passed through a classifier to obtain a classification result that is used to represent whether the module combination has a logical defect. That is, the classifier is used to determine a class probability tag to which the module combination semantic understanding feature vector belongs, where the class probability tag is used to indicate whether a logic defect exists in the module combination. In this way, the module combination is semantically understood to determine whether the module combination has a logic defect to generate unnecessary data simulation and test procedures on the surface.

In the technical scheme of the application, the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector are obtained by conducting text semantic coding based on different semantic association scales through different semantic coding models respectively, so that the distribution of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector is not aligned. Therefore, when the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are fused, information loss is generated when the respective misaligned text semantic coding features of the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are transmitted in a model, and the expression effect of the module combination semantic understanding feature vector on the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector is affected.

Based on this, the applicant of the present application combines semantically understood feature vectors, e.g., denoted as V, for the first scale module ₁ Combining semantic understanding feature vectors, e.g., denoted as V, with the second scale module ₂ Performing forward propagation information retention fusion to obtain a module combination semantic understanding feature vector V ', wherein V' is expressed as:

<<s sum of>>s represents shifting the feature vector left by s bits and right by s bits, respectively, round is a rounding function,is the feature vector V ₁ And V ₂ Average of all eigenvalues of (ii) - (ii) ₁ Represents a norm, d (V) ₁ ,V ₂ ) Is the feature vector V ₁ And V ₂ The distance between them, and log is the base 2 logarithm.

Here, quantization errors and information loss in the forward propagation process are balanced and standardized by introducing a bitwise displacement operation of vectors from the viewpoint of uniformizing information for the forward propagation process of features in the network model, and distribution diversity is introduced by remodelling the distribution of feature parameters before fusion, thereby performing information retention in such a manner as to expand information entropy. In this way, the expression effect of the module combination semantic understanding feature vector on the text semantic features of the text description of the module combination of the different-scale semantic association information is improved, and therefore the accuracy of the classification result obtained by the module combination semantic understanding feature vector through the classifier is improved.

Fig. 2 is an application scenario diagram of substep S120 of a laboratory management method based on modular settings according to an embodiment of the present application. As shown in fig. 2, in this application scenario, first, a text description of each module of the plurality of modules (for example, D illustrated in fig. 2) is acquired, and then, the text description of each module is input into a server (for example, S illustrated in fig. 2) in which a module-based laboratory management algorithm is deployed, where the server can process the text description of each module using the module-based laboratory management algorithm to obtain a classification result for indicating whether a module combination has a logic defect.

Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.

Fig. 3 is a flowchart of substep S120 of the laboratory management method based on the modular setup according to an embodiment of the present application. As shown in fig. 3, according to the laboratory management method based on the modular arrangement in the embodiment of the present application, all module schemes required by the production link are checked to determine whether a plurality of modules required by the production link have logic defects, including: s121, acquiring text descriptions of each module in the plurality of modules; s122, the text description of each module passes through a Word2Vec model to obtain a sequence of module description embedded vectors; s123, enabling the sequence of the module description embedded vectors to pass through a module sequence semantic comprehension device based on a converter module to obtain a plurality of context module description semantic comprehension feature vectors; s124, cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector; s125, the semantic understanding feature vectors are described by the context modules and pass through a two-way long-short-term memory neural network model to obtain a second scale module combined semantic understanding feature vector; s126, fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and S127, passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the module combination has a logic defect.

Fig. 4 is an architectural diagram of substep S120 of a laboratory management method based on modular setup according to an embodiment of the present application. In this network architecture, as shown in fig. 4, first, a text description of each of the plurality of modules is acquired; then, the text description of each module is passed through a Word2Vec model to obtain a sequence of module description embedded vectors; then, passing the sequence of module description embedded vectors through a module sequence semantic understand based on a converter module to obtain a plurality of context module description semantic understand feature vectors; then, cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector; then, the context module description semantic understanding feature vectors pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector; then, fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and finally, passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the module combination has a logic defect.

More specifically, in step S121, a text description of each of the plurality of modules is acquired. In the laboratory management method based on the modular arrangement, each module has its specific functions and roles. Therefore, in order to perform semantic understanding on the modules and determine whether the module combination has a logic defect, a text description of each module needs to be acquired first, where the text description of each module may include information such as a name, a function, input and output of the module.

More specifically, in step S122, the text descriptions of the respective modules are passed through a Word2Vec model to obtain a sequence of module description embedded vectors. The textual descriptions of the individual modules are textual data, which is unstructured data, and therefore, the textual descriptions of the individual modules need to be first converted into a structured language that can be recognized and processed by a computer. The text is converted into the vector representation, so that subsequent semantic understanding and calculation can be conveniently performed.

More specifically, in step S123, the sequence of module description embedded vectors is passed through a module sequence semantic understand based on the converter module to obtain a plurality of context module description semantic understand feature vectors. In this way, semantic association information between textual descriptions of individual modules in the module combination, such as, for example, dependencies and interactions between modules, may be captured.

Accordingly, in one specific example, as shown in fig. 5, passing the sequence of module description embedded vectors through a module sequence semantic understand based on a converter module to obtain a plurality of context module description semantic understand feature vectors includes: s1231, one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector; s1232, calculating products between the global module description feature vector and transpose vectors of the module description embedded vectors in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices; s1233, respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; s1234, each normalized self-attention correlation matrix in the normalized self-attention correlation matrices is processed by a Softmax classification function to obtain a plurality of probability values; and S1235, weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the plurality of context module description semantic understanding feature vectors.

More specifically, in step S124, the plurality of context module description semantic understanding feature vectors are concatenated to obtain a first scale module combination semantic understanding feature vector. That is, the respective local feature distributions of the plurality of context module descriptive semantic understanding feature vectors are aggregated to obtain a global semantic feature representation representing the module combination, i.e. the first scale module combination semantic understanding feature vector. That is, by concatenating the plurality of context module description semantic understanding feature vectors, semantic information of the plurality of modules may be integrated, thereby obtaining a more comprehensive and accurate module combination semantic understanding feature vector.

Accordingly, in one specific example, concatenating the plurality of context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector includes: cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector; wherein, the cascade formula is:

V ₁ ＝Concat[V _a1 ，V _a2 ，V _a3 ……V _an ]

wherein V is _a1 ，V _a2 ，V _a3 ……V _an Representing the plurality of context modules describing a semantic understanding feature vector, concat [ &]Representing a cascade function, V ₁ Representing the first scale module combined semantic understanding feature vector.

More specifically, in step S125, the plurality of context module description semantic understanding feature vectors are passed through a two-way long-short term memory neural network model to obtain a second scale module combination semantic understanding feature vector. By using the two-way long-short term memory neural network model, short-range semantic information and medium-range semantic information among various modules in the module combination can be better captured. In particular, the two-way long and short term memory neural network model may consider both forward and backward context information to better understand the dependencies and interactions between modules.

More specifically, in step S126, the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are fused to obtain a module combination semantic understanding feature vector. In the technical scheme of the application, the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector are obtained by conducting text semantic coding based on different semantic association scales through different semantic coding models respectively, so that the distribution of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector is not aligned. Therefore, when the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are fused, information loss is generated when the respective misaligned text semantic coding features of the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector are transmitted in a model, and the expression effect of the module combination semantic understanding feature vector on the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector is affected. Based on the information, the applicant of the application performs forward propagation information retention fusion on the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector.

Accordingly, in a specific example, fusing the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector to obtain a module combined semantic understanding feature vector includes: fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector by the following fusion formula to obtain the module combination semantic understanding feature vector; wherein, the fusion formula is:

wherein V is ₁ Is the first scale module combined semantic understanding feature vector, V ₂ Is the second scale module combined semantic understanding feature vector,<<s sum of>>s represents shifting the feature vector left by s bits and right by s bits, respectively, round is a rounding function,is the average value of all feature values of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and II is II ₁ Represents a norm, d (V) ₁ ,V ₂ ) Is the distance between the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and log is a logarithmic function based on 2, < + >>For vector subtraction, α and β are weight super parameters.

Here, quantization errors and information loss in the forward propagation process are balanced and standardized by introducing a bitwise displacement operation of vectors from the viewpoint of uniformizing information in the forward propagation process of features in the network model, and distribution diversity is introduced by remolding the distribution of feature parameters before fusion, thereby performing information retention in a manner of expanding information entropy. In this way, the expression effect of the module combination semantic understanding feature vector on the text semantic features of the text description of the module combination of the different-scale semantic association information is improved, and therefore the accuracy of the classification result obtained by the module combination semantic understanding feature vector through the classifier is improved.

More specifically, in step S127, the module combination semantic understanding feature vector is passed through a classifier to obtain a classification result indicating whether or not the module combination has a logical defect. In this way, the module combination is semantically understood to determine whether the module combination has a logic defect to generate unnecessary data simulation and test procedures on the surface.

That is, in the technical solution of the present application, the labels of the classifier include that a module combination has a logic defect (first label) and that a module combination does not have a logic defect (second label), wherein the classifier determines, through a soft maximum function, to which classification label the module combination semantic understanding feature vector belongs. It should be noted that the first tag p1 and the second tag p2 do not include a manually set concept, and in fact, during the training process, the computer model does not have a concept of "whether a module combination has a logic defect", which is only two kinds of classification tags, and the probability that the output feature is under the two classification tags, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the module combination has the logic defect is actually converted into the class probability distribution conforming to the two classes of the natural law through classifying the labels, and the physical meaning of the natural probability distribution of the labels is essentially used instead of the language text meaning of whether the module combination has the logic defect.

It should be appreciated that the role of the classifier is to learn the classification rules and classifier using a given class, known training data, and then classify (or predict) the unknown data. Logistic regression (logistics), SVM, etc. are commonly used to solve the classification problem, and for multi-classification problems (multi-class classification), logistic regression or SVM can be used as well, but multiple bi-classifications are required to compose multiple classifications, but this is error-prone and inefficient, and the commonly used multi-classification method is the Softmax classification function.

Accordingly, in one specific example, as shown in fig. 6, the module combination semantic understanding feature vector is passed through a classifier to obtain a classification result, where the classification result is used to indicate whether a logic defect exists in the module combination, and the method includes: s1271, performing full-connection coding on the module combination semantic understanding feature vector by using a full-connection layer of the classifier to obtain a coding classification feature vector; and S1272, inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

In summary, according to the laboratory management method based on the modular arrangement of the embodiment of the application, firstly, text descriptions of each module in the plurality of modules are obtained, then the text descriptions of each module are passed through a Word2Vec model to obtain a sequence of module description embedded vectors, then the sequence of module description embedded vectors is passed through a module sequence semantic comprehension device based on a converter module to obtain a plurality of context module description semantic comprehension feature vectors, then the plurality of context module description semantic comprehension feature vectors are cascaded to obtain a first-scale module combination semantic comprehension feature vector, then the plurality of context module description semantic comprehension feature vectors are passed through a two-way long-short-term memory neural network model to obtain a second-scale module combination semantic comprehension feature vector, then the first-scale module combination semantic comprehension feature vector and the second-scale module combination semantic comprehension feature vector are fused to obtain a module combination semantic comprehension feature vector, and finally the module combination semantic comprehension feature vector is passed through a classifier to obtain a classification result for indicating whether a logic defect exists in the module combination. In this way, a semantic understanding of the module combination may be performed to determine whether the module combination has a logical defect.

Fig. 7 is a block diagram of a modular setup-based laboratory management system 100, according to an embodiment of the present application. As shown in fig. 7, a laboratory management system 100 based on a modular setup according to an embodiment of the present application includes: a module configuration unit 110, configured to configure a plurality of modules required by a production link, where the modules include a parameter control action module, a start selection module, an end action module, a comparison symbol module, and an end mode module; and a verification judging unit 120, configured to verify the plurality of modules required by the production link to judge whether the plurality of modules required by the production link have a logic defect.

In one example, in the above-described laboratory management system 100 based on a modular arrangement, the verification judging unit 120 includes; a text description obtaining subunit, configured to obtain text descriptions of each of the plurality of modules; the embedded coding subunit is used for enabling the text description of each module to pass through a Word2Vec model to obtain a sequence of module description embedded vectors; a module sequence semantic understanding subunit, configured to pass the sequence of module description embedded vectors through a module sequence semantic understanding device based on a converter module to obtain a plurality of context module description semantic understanding feature vectors; the cascading subunit is used for cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector; the convolution coding subunit is used for enabling the context module description semantic understanding feature vectors to pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector; the fusion subunit is used for fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and the classifying subunit is used for enabling the module combination semantic understanding feature vector to pass through a classifier to obtain a classifying result, and the classifying result is used for indicating whether the module combination has a logic defect or not.

In one example, in the above-described modular setup-based laboratory management system 100, the module sequence semantic understanding subunit is configured to: one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector; calculating the product between the global module description feature vector and the transpose vector of each module description embedded vector in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain the context module description semantic understanding feature vectors.

In one example, in the above-described laboratory management system 100 based on a modular arrangement, the cascade subunit is configured to: cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector; wherein, the cascade formula is:

V ₁ ＝Concat[V _a1 ，V _a2 ，V _a3 ……V _an ]

Wherein V is _a1 ，V _a2 ，V _a3 ……V _an Representing the plurality of context modules describing a semantic understanding feature vector, concat [ &]Representing a cascade function, V ₁ Representing the first scale module combined semantic understanding feature vector.

In one example, in the above-described laboratory management system 100 based on a modular setup, the fusion subunit is configured to: fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector by the following fusion formula to obtain the module combination semantic understanding feature vector; wherein, the fusion formula is:

wherein V is ₁ Is the first scale module combined semantic understanding feature vector, V ₂ Is the second scale module combined semantic understanding feature vector,<<s sum of>>s represents shifting the feature vector left by s bits and right by s bits, respectively, round is a rounding function,is the average value of all feature values of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and II is II ₁ Represents a norm, d (V) ₁ ,V ₂ ) Is the first scale mouldThe distance between the block combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector, and log is a logarithmic function based on 2, +. >For vector subtraction, α and β are weight super parameters.

In one example, in the above-described laboratory management system 100 based on a modular setup, the sorting sub-unit is configured to: performing full-connection coding on the module combination semantic understanding feature vector by using a full-connection layer of the classifier to obtain a coding classification feature vector; and inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

Here, it will be understood by those skilled in the art that the specific functions and operations of the respective modules in the above-described module-based laboratory management system 100 have been described in detail in the above description of the module-based laboratory management method with reference to fig. 1 to 6, and thus, repetitive descriptions thereof will be omitted.

As described above, the laboratory management system 100 based on the modular setup according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server or the like having a laboratory management algorithm based on the modular setup. In one example, the modular setup-based laboratory management system 100 according to embodiments of the present application may be integrated into a wireless terminal as one software module and/or hardware module. For example, the modularly-configured laboratory management system 100 may be a software module in the operating system of the wireless terminal, or may be an application developed for the wireless terminal; of course, the modular setup-based laboratory management system 100 may also be one of a number of hardware modules of the wireless terminal.

Alternatively, in another example, the modularly-configured laboratory management system 100 and the wireless terminal may be separate devices, and the modularly-configured laboratory management system 100 may be connected to the wireless terminal via a wired and/or wireless network and communicate the interaction information in a agreed data format.

Further, the present application also provides schemes of how experiments can be established and conducted. As shown in FIG. 8, wherein, with respect to the pre-configuration section, enterprise factory floor management: enterprise-factory-area. And the material factory level, the building experiment is not classified, and the experimental area level is made. Data source: configuration process-selection of pull-down parameters. And (3) a module: and configuring all module types of the production link. The module scheme is as follows: based on the module, constructing a function list and a parameter list of the function of the module; a list of devices needed to build the module. The unit scheme is as follows: different module schemes can build a set of virtual units. A unit: and a set of entity units built based on the unit scheme. Action library: process fragments. And (3) collocating a Nogan module scheme, and collocating module setting of a stage/process.

Regarding the creation of the experimental section, templates: a process, a project-level process. The scheme is as follows: project-under-path process. Experiment: project-path/project-path-process under scheme. The commit enters the "do" flow. The process comprises the following steps: bill of materials + step (unit scheme) -stage-process-module-parameters. The process in production can view online edition experimental dynamics (process node status/equipment bill of materials/production record/assay report/live video/live data). Experimental report: the experimental conclusion is filled out and an offline version report (process node status/equipment bill of materials/production record/assay report/video playback/data playback) is generated, wherein the report format may be HTML/WORD/PDF, etc.

Regarding the experimental part of implementation, the allocation: assigned to which factory to do and provides a material application. Auditing: auditing process/dispensing unit/completing material dispensing. Starting: the management experiment enters a generation queue. Please check: the sample application is assayed and the result is fed back.

According to another aspect of the present application, there is also provided a non-volatile computer-readable storage medium having stored thereon computer-readable instructions which, when executed by a computer, can perform a method as described above.

Program portions of the technology may be considered to be "products" or "articles of manufacture" in the form of executable code and/or associated data, embodied or carried out by a computer readable medium. A tangible, persistent storage medium may include any memory or storage used by a computer, processor, or similar device or related module. Such as various semiconductor memories, tape drives, disk drives, or the like, capable of providing storage functionality for software.

All or a portion of the software may sometimes communicate over a network, such as the internet or other communication network. Such communication may load software from one computer device or processor to another. For example: a hardware platform loaded from a server or host computer of the video object detection device to a computer environment, or other computer environment implementing the system, or similar functioning system related to providing information needed for object detection. Thus, another medium capable of carrying software elements may also be used as a physical connection between local devices, such as optical, electrical, electromagnetic, etc., propagating through cable, optical cable, air, etc. Physical media used for carrier waves, such as electrical, wireless, or optical, may also be considered to be software-bearing media. Unless limited to a tangible "storage" medium, other terms used herein to refer to a computer or machine "readable medium" mean any medium that participates in the execution of any instructions by a processor.

This application uses specific words to describe embodiments of the application. Reference to "a first/second embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the present application. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present application may be combined as suitable.

Furthermore, those skilled in the art will appreciate that the various aspects of the invention are illustrated and described in the context of a number of patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof.

Accordingly, aspects of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.) or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the present application may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

The foregoing is illustrative of the present invention and is not to be construed as limiting thereof. Although a few exemplary embodiments of this invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without materially departing from the novel teachings and advantages of this invention. Accordingly, all such modifications are intended to be included within the scope of this invention as defined in the following claims. It is to be understood that the foregoing is illustrative of the present invention and is not to be construed as limited to the specific embodiments disclosed, and that modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The invention is defined by the claims and their equivalents.

Claims (4)

1. A laboratory management method based on a modular setup, comprising:

configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a symbol comparison module and an ending mode module; and

checking a plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects or not;

and verifying all module schemes required by the production link to judge whether a plurality of modules required by the production link have logic defects, wherein the method comprises the following steps of:

acquiring text descriptions of each module in the plurality of modules;

the text description of each module is passed through a Word2Vec model to obtain a sequence of module description embedded vectors;

passing the sequence of module description embedded vectors through a module sequence semantic understand based on a converter module to obtain a plurality of context module description semantic understand feature vectors;

cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector;

the context module description semantic understanding feature vectors pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector;

Fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and

passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether a module combination has a logic defect or not;

wherein passing the sequence of module description embedded vectors through a module sequence semantic understand based on a converter module to obtain a plurality of context module description semantic understand feature vectors, comprising:

one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector;

calculating the product between the global module description feature vector and the transpose vector of each module description embedded vector in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices;

respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;

obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and

Weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain a plurality of context module description semantic understanding feature vectors;

the cascade connection of the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector comprises the following steps:

cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector;

wherein, the cascade formula is:

wherein,representing the plurality of context modules describing a semantic understanding feature vector,/a plurality of context modules describing a semantic understanding feature vector>Representing a cascade function->Representing the first scale module combined semantic understanding feature vector.

2. The modular setup-based laboratory management method of claim 1, wherein fusing the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector to obtain a module combined semantic understanding feature vector, comprises:

fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector by the following fusion formula to obtain the module combination semantic understanding feature vector;

Wherein, the fusion formula is:

wherein,is the first scale module combined semantic understanding feature vector,>is the second scale module combined semantic understanding feature vector,>and->Respectively represent the left shift of the feature vector +.>Bit and right shift->Bit (s)/(s)>In order to be a function of the rounding-off,is the mean value of all feature values of the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, +.>Representing a norm of the feature vector, +.>Is the distance between the first scale module combined semantic understanding feature vector and the second scale module combined semantic understanding feature vector, and +.>As a logarithmic function with base 2 +.>For vector subtraction, ++>Is a weight super parameter.

3. The modular setup-based laboratory management method of claim 2, wherein passing the module combination semantic understanding feature vector through a classifier to obtain a classification result, the classification result being used to represent whether a module combination has a logical defect, comprises:

performing full-connection coding on the module combination semantic understanding feature vector by using a full-connection layer of the classifier to obtain a coding classification feature vector; and

And inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

4. A laboratory management system based on a modular setup, comprising:

the module configuration unit is used for configuring a plurality of modules required by a production link, wherein the modules comprise a parameter control action module, a starting selection module, an ending action module, a comparison symbol module and an ending mode module; and

the verification judging unit is used for verifying a plurality of modules required by the production link to judge whether the plurality of modules required by the production link have logic defects or not;

wherein, the check judgment unit includes:

a text description obtaining subunit, configured to obtain text descriptions of each of the plurality of modules;

the embedded coding subunit is used for enabling the text description of each module to pass through a Word2Vec model to obtain a sequence of module description embedded vectors;

a module sequence semantic understanding subunit, configured to pass the sequence of module description embedded vectors through a module sequence semantic understanding device based on a converter module to obtain a plurality of context module description semantic understanding feature vectors;

The cascading subunit is used for cascading the context module description semantic understanding feature vectors to obtain a first scale module combination semantic understanding feature vector;

the convolution coding subunit is used for enabling the context module description semantic understanding feature vectors to pass through a two-way long-short-term memory neural network model to obtain a second scale module combination semantic understanding feature vector;

the fusion subunit is used for fusing the first scale module combination semantic understanding feature vector and the second scale module combination semantic understanding feature vector to obtain a module combination semantic understanding feature vector; and

the classifying subunit is used for passing the module combination semantic understanding feature vector through a classifier to obtain a classifying result, wherein the classifying result is used for indicating whether the module combination has a logic defect or not;

wherein the module sequence semantic understanding subunit is configured to:

one-dimensional arrangement is carried out on the sequence of the module description embedded vector so as to obtain a global module description feature vector;

calculating the product between the global module description feature vector and the transpose vector of each module description embedded vector in the sequence of module description embedded vectors to obtain a plurality of self-attention association matrices;

Respectively carrying out standardization processing on each self-attention correlation matrix in the plurality of self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices;

obtaining a plurality of probability values by using a Softmax classification function through each normalized self-attention correlation matrix in the normalized self-attention correlation matrices; and

weighting each module description embedded vector in the sequence of module description embedded vectors by taking each probability value in the plurality of probability values as a weight to obtain a plurality of context module description semantic understanding feature vectors;

wherein the cascade subunit is configured to:

cascading the plurality of context module description semantic understanding feature vectors with the following cascading formula to obtain the first scale module combination semantic understanding feature vector;

wherein, the cascade formula is:

wherein,representing the plurality of context modules describes a semantic understanding feature vector,representing cascade functionsCount (n)/(l)>Representing the first scale module combined semantic understanding feature vector.