CN109993358B

CN109993358B - Method and device for training yield prediction model

Info

Publication number: CN109993358B
Application number: CN201910231165.2A
Authority: CN
Inventors: 杨帆; 金宝宝; 金继民; 余健伟; 张成松
Original assignee: Lenovo Beijing Ltd
Current assignee: Lenovo Beijing Ltd
Priority date: 2019-03-25
Filing date: 2019-03-25
Publication date: 2021-07-16
Anticipated expiration: 2039-03-25
Also published as: CN109993358A

Abstract

The present disclosure provides a method for training a yield prediction model, including obtaining training data, where the training data includes data of a plurality of first indexes and data of a yield, where the plurality of first indexes include controllable indexes and uncontrollable indexes, determining, based on the data of the plurality of first indexes, an uncontrollable index having a correlation with any of the controllable indexes greater than a first threshold value from among the uncontrollable indexes as a first type of uncontrollable index, and determining other uncontrollable indexes as a second type of uncontrollable indexes, taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as inputs, and taking the data of the yield as an output, and training the yield prediction model. The present disclosure also provides a device for training a yield prediction model.

Description

Method and device for training yield prediction model

Technical Field

The present disclosure relates to a method and apparatus for training a yield prediction model.

Background

For the petrochemical industry and other process manufacturing industries, the process indexes (such as temperature, pressure and the like) of the device directly determine the yield of the product. The process index of the device is effectively adjusted, and the method is very critical for improving the product yield and further realizing the maximization of the economic benefit of the device. Based on the massive historical data accumulated by the device, a machine learning technology is adopted to mine a process index value which enables the yield of the target product of the device to reach the maximum. However, the present inventors have found that the effect of the existing models on the prediction of product yield is not ideal.

Disclosure of Invention

One aspect of the present disclosure provides a method for training a yield prediction model, including obtaining training data, where the training data includes data of a plurality of first indicators and data of a yield, where the plurality of first indicators include controllable indicators and uncontrollable indicators, determining, based on the data of the plurality of first indicators, an uncontrollable indicator having a correlation with any of the controllable indicators greater than a first threshold value from among the uncontrollable indicators as a first type of uncontrollable indicator, and determining other uncontrollable indicators as a second type of uncontrollable indicator, taking the data of the controllable indicators and the data of the second type of uncontrollable indicator as inputs, and taking the data of the yield as an output, and training the yield prediction model.

Optionally, the training of the yield prediction model includes training the yield prediction model by taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as inputs and the data of the yield as outputs.

Optionally, the obtaining training data includes obtaining raw data, the raw data including data of a plurality of second indexes and data of yield, determining, as a first index, a second index having a correlation with the yield higher than a second threshold value among the plurality of second indexes, and generating training data including data of the plurality of first indexes and the data of the yield.

Optionally, the yield prediction model comprises a neural network model for predicting the yield, the neural network model comprising a plurality of hidden layers, the hidden layers comprising at least one of fully-connected layers, RNN layers, or LSTM layers.

Optionally, the method further includes obtaining data of the controllable indexes at the time point to be adjusted and data of the second type of the controllable indexes, obtaining prediction data of the yield at the time point to be adjusted through the trained yield prediction model, and adjusting the input data of the controllable indexes so that the prediction data of the yield reaches a maximum value, and determining the data of the controllable indexes, which make the prediction data of the yield reach the maximum value, as an optimization result.

Another aspect of the disclosure provides an apparatus for training a yield prediction model, comprising a first obtaining module, a classification module, and a training module. The system comprises a first obtaining module, a second obtaining module and a third obtaining module, wherein the first obtaining module is used for obtaining training data, the training data comprise data of a plurality of first indexes and data of yield, and the plurality of first indexes comprise controllable indexes and uncontrollable indexes. And the classification module is used for determining the uncontrollable indexes with the correlation larger than a first threshold value with any controllable indexes from the uncontrollable indexes as a first-class uncontrollable index and determining other uncontrollable indexes as a second-class uncontrollable index based on the data of the first indexes. And the training module is used for taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as input, taking the data of the yield as output and training the yield prediction model.

Optionally, the training module is configured to train the yield prediction model by taking the data of the controllable indicator and the data of the second type of uncontrollable indicator as inputs and taking the data of the first type of uncontrollable indicator and the data of the yield as outputs.

Optionally, the first obtaining module includes an obtaining submodule, a determining submodule, and a generating submodule. And the obtaining submodule is used for obtaining the original data, and the original data comprises data of a plurality of second indexes and data of yield. A determination submodule for determining, as the first indicator, a second indicator of the plurality of second indicators having a correlation with the yield higher than a second threshold. And the generation submodule is used for generating training data comprising data of a plurality of first indexes and data of yield.

Optionally, the apparatus further comprises a second obtaining module, a predicting module, and an adjusting module. And the second obtaining module is used for obtaining the data of the controllable indexes and the data of the second type of uncontrollable indexes at the time point to be optimized. And the prediction module is used for obtaining the prediction data of the yield of the time point to be adjusted and optimized through the trained yield prediction model. And the adjusting and optimizing module is used for adjusting the input data of the controllable indexes to enable the predicted data of the yield to reach the maximum value, and determining the data of the controllable indexes which enable the predicted data of the yield to reach the maximum value as an optimization result.

Another aspect of the disclosure provides an electronic device comprising at least one processor and at least one memory storing one or more computer-readable instructions, wherein the one or more computer-readable instructions, when executed by the at least one processor, cause the processor to perform the method as described above.

Another aspect of the disclosure provides a non-volatile storage medium storing computer-executable instructions for implementing the method as described above when executed.

Another aspect of the disclosure provides a computer program comprising computer executable instructions for implementing the method as described above when executed.

Drawings

For a more complete understanding of the present disclosure and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

fig. 1 schematically illustrates an application scenario of yield prediction according to an embodiment of the present disclosure;

FIG. 2A schematically illustrates a flow diagram of a method of training a yield prediction model, in accordance with an embodiment of the present disclosure;

fig. 2B schematically illustrates a schematic of a yield prediction model according to an embodiment of the disclosure.

FIG. 3 schematically shows a flow chart for obtaining training data according to an embodiment of the present disclosure;

FIG. 4 schematically illustrates a flow diagram for yield optimization according to an embodiment of the disclosure;

FIG. 5 schematically illustrates a block diagram of an apparatus for training a yield prediction model, in accordance with an embodiment of the present disclosure;

FIG. 6 schematically shows a block diagram of a first obtaining module according to an embodiment of the disclosure;

FIG. 7 schematically illustrates a block diagram of an apparatus for training a yield prediction model according to another embodiment of the present disclosure; and

FIG. 8 schematically shows a block diagram of an electronic device according to an embodiment of the disclosure.

Detailed Description

Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. It should be understood that the description is illustrative only and is not intended to limit the scope of the present disclosure. In the following detailed description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the disclosure. It may be evident, however, that one or more embodiments may be practiced without these specific details. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present disclosure.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The terms "comprises," "comprising," and the like, as used herein, specify the presence of stated features, steps, operations, and/or components, but do not preclude the presence or addition of one or more other features, steps, operations, or components.

All terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. It is noted that the terms used herein should be interpreted as having a meaning that is consistent with the context of this specification and should not be interpreted in an idealized or overly formal sense.

Where a convention analogous to "at least one of A, B and C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B and C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). Where a convention analogous to "A, B or at least one of C, etc." is used, in general such a construction is intended in the sense one having skill in the art would understand the convention (e.g., "a system having at least one of A, B or C" would include but not be limited to systems that have a alone, B alone, C alone, a and B together, a and C together, B and C together, and/or A, B, C together, etc.). It will be further understood by those within the art that virtually any disjunctive word and/or phrase presenting two or more alternative terms, whether in the description, claims, or drawings, should be understood to contemplate the possibilities of including one of the terms, either of the terms, or both terms. For example, the phrase "a or B" should be understood to include the possibility of "a" or "B", or "a and B".

Some block diagrams and/or flow diagrams are shown in the figures. It will be understood that some blocks of the block diagrams and/or flowchart illustrations, or combinations thereof, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the instructions, which execute via the processor, create means for implementing the functions/acts specified in the block diagrams and/or flowchart block or blocks.

Accordingly, the techniques of this disclosure may be implemented in hardware and/or software (including firmware, microcode, etc.). In addition, the techniques of this disclosure may take the form of a computer program product on a computer-readable medium having instructions stored thereon for use by or in connection with an instruction execution system. In the context of this disclosure, a computer-readable medium may be any medium that can contain, store, communicate, propagate, or transport the instructions. For example, the computer readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium. Specific examples of the computer readable medium include: magnetic storage devices, such as magnetic tape or Hard Disk Drives (HDDs); optical storage devices, such as compact disks (CD-ROMs); a memory, such as a Random Access Memory (RAM) or a flash memory; and/or wired/wireless communication links.

An embodiment of the present disclosure provides a method for training a yield prediction model, including obtaining training data, where the training data includes data of a plurality of first indicators and data of a yield, where the plurality of first indicators include controllable indicators and uncontrollable indicators, determining, based on the data of the plurality of first indicators, an uncontrollable indicator having a correlation with any of the controllable indicators greater than a first threshold value from the uncontrollable indicators as a first type of uncontrollable indicator, and determining other uncontrollable indicators as a second type of uncontrollable indicator, taking the data of the controllable indicators and the data of the second type of uncontrollable indicators as inputs, and taking the data of the yield as an output, and training the yield prediction model.

Fig. 1 schematically illustrates an application scenario of yield prediction according to an embodiment of the present disclosure. It should be noted that fig. 1 is only an example of a scenario in which the embodiments of the present disclosure may be applied to help those skilled in the art understand the technical content of the present disclosure, but does not mean that the embodiments of the present disclosure may not be applied to other devices, systems, environments or scenarios.

As shown in fig. 1, the yield of chemical synthesis is usually determined by many factors, including but not limited to the ratio of raw materials, the feeding sequence and timing of raw materials, various conditions of reaction, environmental conditions, etc., and the relationship between these factors and the yield is usually complicated, especially in some special types of reactions, such as catalytic cracking, it is difficult to explain the relationship between each factor and the yield of product.

For the petrochemical industry and other process manufacturing industries, the process indexes (such as temperature, pressure and the like) of the device directly determine the yield of high-value products. The process index of the device is effectively adjusted, and the method is very critical for improving the yield of high-value products and further realizing the maximization of the economic benefit of the device. However, as the processes in the process manufacturing industry are various, the processes are complex, and the processes have complex mutual influence, the process indexes of the device are adjusted by simply depending on the personal experience of workshop workers, so that the device has great limitation, and the economic benefit of the device is difficult to be maximized.

Based on the massive historical data accumulated by the device, a machine learning technology is adopted, and a process index value which enables the yield of the target product of the device to reach the maximum is mined, so that the method is an effective mode for realizing the maximization of the economic benefit of the device. The method comprises the steps of firstly establishing a relation Y ═ F (X) between a device process index X and a target product yield Y based on historical data, enabling Y calculated based on Y ═ F (X) to be maximum by changing the value of a controllable target in the device process index X in the actual production process, and selecting X enabling Y to be maximum as a regulation value of the device process index.

One of the main problems of this method is that the controllable index and the uncontrollable index are strongly dependent, and when the value of X that maximizes the product yield of the device is found in the actual production process, only the value of the controllable index is changed, and the value of the uncontrollable index is not adjusted (assuming that the value of the uncontrollable index is not changed), which is inconsistent with the actual situation that the actual value of the uncontrollable index changes after the value of the controllable index is changed. Such inconsistency can greatly affect the effectiveness of the given index control value.

According to the method provided by the embodiment of the invention, the uncontrollable indexes are divided into a first type of uncontrollable indexes and a second type of uncontrollable indexes, the data of the controllable indexes and the data of the second type of uncontrollable indexes are used as input, the data of the yield is used as output, and the yield prediction model is trained, so that the constructed yield prediction model can capture the uncontrollable index change caused by the controllable indexes, and further the information of the yield is influenced. The regulation and control value of the device process index given based on the method is more effective. A method for constructing a yield prediction model according to an embodiment of the present disclosure will be described below with reference to fig. 2A.

Fig. 2A schematically illustrates a flow diagram of a method of training a yield prediction model according to an embodiment of the present disclosure.

As shown in fig. 2A, the method includes operations S210 to S230.

In operation S210, training data is obtained, where the training data includes data of a plurality of first indicators including a controllable indicator and an uncontrollable indicator, and data of a yield. The controllable indicators may for example comprise flow, pressure, temperature, etc. of the gas used for heating, and the uncontrollable indicators may for example comprise catalyst activity, etc.

For example, the collected index (expressed by: X1, X2, and Xn, etc., such as the temperature of the device) representing the operation condition of the device and the yield (expressed by Y, such as the gasoline yield of the petrochemical industry) of the product to be predicted can be obtained from the database (such as the Info Plus 21 real-time database of the petrochemical industry) of the target process manufacturing enterprise, so as to obtain the data table shown in Table 1.

Table 1: example of plant operating data and product yield data

Time	X1	X2	...	Xn	Y
						T1	X1_1	X2_1	...	Xn_1	Y_1
T2	X1_2	X2_2	...	Xn_2	Y_2
						T3	X1_3	X2_3	...	Xn_3	Y_3
T4	X1_4	X2_4	...	Xn_4	Y_4
						...	...	...	...	...	...
TM	X1_m	X2_m	...	Xn_m	Y_m

Where "time" is the time at which the index value is collected, the yield Y of the high value product may be obtained directly from a meter on the device (if there is a corresponding measurement point on the device), or indirectly calculated from the values at other measurement points.

According to the embodiments of the present disclosure, not all indexes can affect the yield, and the embodiments of the present disclosure may select an index related to the yield as the first index to analyze. This is explained below with reference to fig. 3.

Fig. 3 schematically shows a flow chart for obtaining training data according to another embodiment of the present disclosure.

As shown in fig. 3, the method includes operations S310 to S330.

In operation S310, raw data including data of a plurality of second indexes and data of yield is obtained.

In operation S320, a second index having a correlation with the yield higher than a second threshold value among the plurality of second indexes is determined as the first index. For example, the correlation between the second index and the yield can be determined by using parameters such as a pearson correlation coefficient or a transfer entropy as a characteristic value of the correlation between the second index and the yield.

In operation S330, training data including data of a plurality of first indexes and data of a yield is generated.

The method screens out the first index related to the yield from all the second indexes which can be obtained, and avoids the influence of the index unrelated to the yield on the yield prediction.

Reference is made back to fig. 2A. In operation S220, based on the data of the plurality of first indicators, an uncontrollable indicator having a correlation with any of the controllable indicators greater than a first threshold is determined from the uncontrollable indicators as a first type of uncontrollable indicator, and other uncontrollable indicators are determined as a second type of uncontrollable indicator.

For example, when one controllable index XFC-1 changes, another uncontrollable index XFH-1 obviously changes along with the change of the controllable index XFC-1, and the correlation is greater than a first threshold value, then the uncontrollable index XFH-1 is determined as a first type of uncontrollable index. For another example, if there is no significant correlation between a certain uncontrollable index XFL-1 and each controllable index XFC, the uncontrollable index XFL-1 is determined as the second type of uncontrollable index.

According to the embodiment of the disclosure, for example, parameters such as a pearson correlation coefficient or a transfer entropy can be used as a characteristic value of the correlation between the uncontrollable index and the controllable index, and the correlation between the uncontrollable index and the controllable index can be determined.

In operation S230, the data of the controllable indicator and the data of the second type of uncontrollable indicator are used as input, the data of the yield is used as output, and the yield prediction model is trained.

Because the first type of controllable indexes change along with the controllable indexes, the controllable indexes and the second type of uncontrollable indexes are only used as input for prediction, and the first type of uncontrollable indexes are prevented from interfering with yield prediction.

The yield prediction model may include, for example, a neural network model for predicting the yield. The neural network model may, for example, include a plurality of hidden layers. The hidden layers may include, for example, a full connection layer, an RNN (Recurrent Neural Network) layer, or an LSTM (Long Short-Term Memory) layer.

Optionally, the embodiment of the present disclosure may further use the first type of uncontrollable indicator and the data of the yield as two outputs of a model, and train the yield prediction model. The output of the first-class uncontrollable indexes is used for supervising the training of the model, so that the prediction effect of the yield prediction model can be further improved.

The yield prediction model of the disclosed embodiment is described below with reference to fig. 2B.

As shown in fig. 2B, the yield prediction model includes two input layers, a plurality of hidden layers, and two output layers. Wherein, the circle represents the input layer of the neural network, and comprises a controllable index XFC and a second type of uncontrollable index XFL; the rectangle box represents a hidden layer of the neural network; the ellipse represents an output layer of the neural network, wherein XF 'represents an output layer with a first type of uncontrollable indexes XF as fitting supervision quantity, and Y' represents an output layer with a fitting yield Y; the arrows represent the flowing direction of data in the neural network, two out-arrows on one hidden layer represent two identical copies of the hidden layer output, and two in-arrows on one neuron represent the concatenation of the two inputs.

It should be noted that the specific implementation form of the hidden layer in the figure is not limited, and may be a full connection layer, an RNN layer, an LSTM layer, and the like. In the case of the recurrent neural network structure such as RNN, LSTM, etc., the upper diagram is only one time sequence, and can be extended to a plurality of time sequences.

According to an embodiment of the present disclosure, the method may further include adjusting a format of the input data based on the model structure of fig. 2B, so that the data conforms to a requirement of the model for the input data. Such as the need for a recurrent neural network to adjust the input data to time series data.

In the model training process, parameters of the neural network are updated, so that the index and yield Y 'of the first type of uncontrollable index XFH' predicted on the basis of the two types of indexes, namely the controllable index XFC and the second type of uncontrollable index XFL, are more accurate, namely, the deviation between the predicted XFH 'and the actual XFH and the deviation between the predicted Y' and the actual Y are minimized. The deviation can be characterized by, for example, an average absolute error, a root mean square error, and the like.

According to the method, the controllable indexes and the second type of uncontrollable indexes which are not related to other controllable indexes are used as the input of the model together, so that the yield prediction can be influenced by the trained yield prediction model based on the change of the second type of uncontrollable indexes, and the prediction result is more effective.

After the yield prediction model is trained by the method described above, the model can be used to optimize yield. The following description is made with reference to the embodiment illustrated in fig. 4.

Fig. 4 schematically illustrates a flow diagram for yield optimization according to an embodiment of the disclosure.

As shown in fig. 4, the method includes operations S410 to S430.

In operation S410, data of the controllable indicator and data of the second type of the uncontrollable indicator at a time point to be tuned are obtained.

In operation S420, prediction data of the yield at the time point to be tuned is obtained through the trained yield prediction model.

In operation S430, the input data of the controllable index is adjusted so that the predicted data of the yield reaches a maximum value, and the data of the controllable index, which reaches the maximum value of the predicted data of the yield, is determined as an optimization result.

According to the method, the trained model is adopted, when the data of the controllable indexes are adjusted, the model can predict the yield by integrating the factors of the first type of uncontrollable indexes changing along with the controllable indexes, the prediction result is more reliable, and the yield can reach a higher level by determining the data of the controllable indexes which enable the prediction data of the yield to reach the maximum value on the basis.

Fig. 5 schematically illustrates a block diagram of an apparatus 500 for training a yield prediction model according to an embodiment of the present disclosure.

As shown in fig. 5, the apparatus 500 for training a yield prediction model includes a first obtaining module 510, a classifying module 520, and a training module 530. The apparatus 500 may perform the method described above with reference to fig. 2A to train a yield prediction model.

The first obtaining module 510, for example performing operation S210 described with reference to the above, is configured to obtain training data, the training data including data of a plurality of first indicators including a controllable indicator and an uncontrollable indicator, and data of a yield.

The classification module 520, for example, performs operation S220 described above with reference to the plurality of first indicators, and is configured to determine, as a first class of uncontrollable indicators, uncontrollable indicators having a correlation with any of the controllable indicators greater than a first threshold value from among the uncontrollable indicators, and determine other uncontrollable indicators as a second class of uncontrollable indicators, based on the data of the plurality of first indicators.

The training module 530, for example, performs operation S230 described above with reference to, for example, training the yield prediction model with the data of the controllable indicator and the data of the second type of uncontrollable indicator as inputs and the data of the yield as an output.

According to the embodiment of the disclosure, the training module is configured to train the yield prediction model by taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as input and taking the data of the first type of uncontrollable indexes and the data of the yield as output.

Fig. 6 schematically illustrates a block diagram of a first obtaining module 600 according to an embodiment of the disclosure.

As shown in fig. 6, the first obtaining module 600 includes an obtaining sub-module 610, a determining sub-module 620, and a generating sub-module 630.

The obtaining sub-module 610, for example, performs operation S310 described with reference to the above, for obtaining raw data including data of a plurality of second indexes and data of yield.

The determining sub-module 620, for example, performs operation S320 described with reference to the above, for determining, as the first index, a second index having a correlation with the yield higher than a second threshold value among the plurality of second indexes.

The generation sub-module 630, for example, performs operation S330 described above with reference to, for example, generating training data including data of a plurality of first indexes and data of yield.

According to an embodiment of the present disclosure, the yield prediction model includes a neural network model for predicting the yield, the neural network model including a plurality of hidden layers, the hidden layers including at least one of a fully-connected layer, an RNN layer, or an LSTM layer.

Fig. 7 schematically illustrates a block diagram of an apparatus for training a yield prediction model 700 according to another embodiment of the present disclosure.

As shown in fig. 700, the apparatus may further include a second obtaining module 710, a predicting module 720 and an adjusting module 730 based on the foregoing embodiments.

The second obtaining module 710, for example, performs operation S410 described above with reference to obtain data of the controllable indicator and data of the second type of the uncontrollable indicator at a time point to be tuned.

The prediction module 720, for example performing operation S420 described above with reference, is configured to obtain prediction data of the yield at the time point to be tuned by the trained yield prediction model.

The tuning module 730, for example, performs the operation S430 described above with reference to adjust the input data of the controllable index so that the predicted data of the yield reaches the maximum value, and determines the data of the controllable index that makes the predicted data of the yield reach the maximum value as the optimized result.

Any number of modules, sub-modules, units, sub-units, or at least part of the functionality of any number thereof according to embodiments of the present disclosure may be implemented in one module. Any one or more of the modules, sub-modules, units, and sub-units according to the embodiments of the present disclosure may be implemented by being split into a plurality of modules. Any one or more of the modules, sub-modules, units, sub-units according to embodiments of the present disclosure may be implemented at least in part as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented in any other reasonable manner of hardware or firmware by integrating or packaging a circuit, or in any one of or a suitable combination of software, hardware, and firmware implementations. Alternatively, one or more of the modules, sub-modules, units, sub-units according to embodiments of the disclosure may be at least partially implemented as a computer program module, which when executed may perform the corresponding functions.

For example, any of the first obtaining module 510, the classifying module 520, the training module 530, the obtaining sub-module 610, the determining sub-module 620, the generating sub-module 630, the second obtaining module 710, the predicting module 720, and the tuning module 730 may be combined into one module to be implemented, or any one of them may be split into a plurality of modules. Alternatively, at least part of the functionality of one or more of these modules may be combined with at least part of the functionality of the other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the first obtaining module 510, the classifying module 520, the training module 530, the obtaining sub-module 610, the determining sub-module 620, the generating sub-module 630, the second obtaining module 710, the predicting module 720, and the tuning module 730 may be implemented at least in part as a hardware circuit, such as a Field Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a system on a chip, a system on a substrate, a system on a package, an Application Specific Integrated Circuit (ASIC), or may be implemented by hardware or firmware in any other reasonable manner of integrating or packaging a circuit, or implemented by any one of three implementations of software, hardware, and firmware, or a suitable combination of any of them. Alternatively, at least one of the first obtaining module 510, the classifying module 520, the training module 530, the obtaining sub-module 610, the determining sub-module 620, the generating sub-module 630, the second obtaining module 710, the predicting module 720 and the tuning module 730 may be at least partially implemented as a computer program module which, when executed, may perform a corresponding function.

Fig. 8 schematically shows a block diagram of an electronic device 800 according to an embodiment of the disclosure. The computer system illustrated in FIG. 8 is only one example and should not impose any limitations on the scope of use or functionality of embodiments of the disclosure.

As shown in fig. 8, electronic device 800 includes a processor 810 and a computer-readable storage medium 820. The electronic device 800 may perform a method according to an embodiment of the disclosure.

In particular, processor 810 may include, for example, a general purpose microprocessor, an instruction set processor and/or related chip set and/or a special purpose microprocessor (e.g., an Application Specific Integrated Circuit (ASIC)), and/or the like. The processor 810 may also include on-board memory for caching purposes. Processor 810 may be a single processing unit or a plurality of processing units for performing different actions of a method flow according to embodiments of the disclosure.

Computer-readable storage medium 820 may be, for example, any medium that can contain, store, communicate, propagate, or transport the instructions. For example, a readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium. Specific examples of the readable storage medium include: magnetic storage devices, such as magnetic tape or Hard Disk Drives (HDDs); optical storage devices, such as compact disks (CD-ROMs); a memory, such as a Random Access Memory (RAM) or a flash memory; and/or wired/wireless communication links.

The computer-readable storage medium 820 may include a computer program 821, which computer program 821 may include code/computer-executable instructions that, when executed by the processor 810, cause the processor 810 to perform a method according to an embodiment of the present disclosure, or any variation thereof.

The computer program 821 may be configured with, for example, computer program code comprising computer program modules. For example, in an example embodiment, code in computer program 821 may include one or more program modules, including for example 821A, modules 821B, … …. It should be noted that the division and number of modules are not fixed, and those skilled in the art may use suitable program modules or program module combinations according to actual situations, and when the program modules are executed by the processor 810, the processor 810 may execute the method according to the embodiment of the present disclosure or any variation thereof.

According to an embodiment of the present disclosure, at least one of the first obtaining module 510, the classifying module 520, the training module 530, the obtaining sub-module 610, the determining sub-module 620, the generating sub-module 630, the second obtaining module 710, the predicting module 720, and the tuning module 730 may be implemented as a computer program module described with reference to fig. 8, which, when executed by the processor 810, may implement the respective operations described above.

The present disclosure also provides a computer-readable medium, which may be embodied in the apparatus/device/system described in the above embodiments; or may exist separately and not be assembled into the device/apparatus/system. The computer readable medium carries one or more programs which, when executed, implement the method according to the embodiments of the present disclosure, or any variations thereof.

According to embodiments of the present disclosure, a computer readable medium may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, 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 disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In contrast, in the present disclosure, a computer-readable signal medium may include a propagated data signal with computer-readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, optical fiber cable, radio frequency signals, etc., or any suitable combination of the foregoing.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

Those skilled in the art will appreciate that various combinations and/or combinations of features recited in the various embodiments and/or claims of the present disclosure can be made, even if such combinations or combinations are not expressly recited in the present disclosure. In particular, various combinations and/or combinations of the features recited in the various embodiments and/or claims of the present disclosure may be made without departing from the spirit or teaching of the present disclosure. All such combinations and/or associations are within the scope of the present disclosure.

While the disclosure has been shown and described with reference to certain exemplary embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the disclosure as defined by the appended claims and their equivalents. Accordingly, the scope of the present disclosure should not be limited to the above-described embodiments, but should be defined not only by the appended claims, but also by equivalents thereof.

Claims

1. A method of training a yield prediction model, comprising:

obtaining training data, wherein the training data comprises data of a plurality of first indexes and data of yield, and the plurality of first indexes comprise controllable indexes and uncontrollable indexes;

determining, from the uncontrollable indicators, uncontrollable indicators having a correlation with any of the controllable indicators greater than a first threshold as a first type of uncontrollable indicator based on the data of the first plurality of indicators, and determining other uncontrollable indicators as a second type of uncontrollable indicator, the first type of uncontrollable indicator varying with the controllable indicators;

taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as input, taking the data of the first type of uncontrollable indexes and the data of the yield as output, and training the yield prediction model;

obtaining data of the controllable indexes and data of the second type of uncontrollable indexes at a time point to be optimized;

obtaining the prediction data of the yield of the time point to be adjusted and optimized through the trained yield prediction model; and

and adjusting the input data of the controllable indexes to enable the prediction data of the yield to reach the maximum value, and determining the data of the controllable indexes enabling the prediction data of the yield to reach the maximum value as an optimization result.

2. The method of claim 1, wherein the obtaining training data comprises:

obtaining raw data, wherein the raw data comprises data of a plurality of second indexes and data of yield;

determining a second index having a correlation with the yield higher than a second threshold value among the plurality of second indexes as a first index;

training data including data for a plurality of first indicators and data for a yield is generated.

3. The method of claim 1, wherein the yield prediction model comprises a neural network model for predicting yield, the neural network model comprising a plurality of hidden layers, the hidden layers comprising at least one of fully-connected layers, RNN layers, or LSTM layers.

4. An apparatus for training a yield prediction model, comprising:

the system comprises a first obtaining module, a second obtaining module and a third obtaining module, wherein the first obtaining module is used for obtaining training data, the training data comprises data of a plurality of first indexes and data of yield, and the plurality of first indexes comprise controllable indexes and uncontrollable indexes;

the classification module is used for determining an uncontrollable index with the correlation larger than a first threshold value with any controllable index from the uncontrollable indexes as a first-class uncontrollable index and determining other uncontrollable indexes as a second-class uncontrollable index based on the data of the first indexes, wherein the first-class uncontrollable index can change along with the controllable indexes;

the training module is used for taking the data of the controllable indexes and the data of the second type of uncontrollable indexes as input, taking the data of the first type of uncontrollable indexes and the data of the yield as output, and training the yield prediction model;

the second obtaining module is used for obtaining the data of the controllable indexes and the data of the second type of uncontrollable indexes at the time point to be adjusted and optimized;

the prediction module is used for obtaining the prediction data of the yield of the time point to be adjusted and optimized through the trained yield prediction model; and

and the adjusting and optimizing module is used for adjusting the input data of the controllable indexes to enable the predicted data of the yield to reach the maximum value, and determining the data of the controllable indexes which enable the predicted data of the yield to reach the maximum value as an optimization result.

5. The apparatus of claim 4, wherein the first obtaining means comprises:

an obtaining submodule for obtaining raw data, the raw data including data of a plurality of second indicators and data of a yield;

a determination submodule for determining, as the first index, a second index having a correlation with the yield higher than a second threshold value among the plurality of second indexes;

and the generation submodule is used for generating training data comprising data of a plurality of first indexes and data of yield.

6. The apparatus of claim 4, wherein the yield prediction model comprises a neural network model for predicting yield, the neural network model comprising a plurality of hidden layers, the hidden layers comprising at least one of fully-connected layers, RNN layers, or LSTM layers.