CN116027874A

CN116027874A - Notebook computer power consumption control method and system thereof

Info

Publication number: CN116027874A
Application number: CN202211296239.9A
Authority: CN
Inventors: 曾蓉蓉; 彭斌
Original assignee: Shenzhen Chuang Ruixin Technology Co ltd
Current assignee: Shenzhen Chuang Ruixin Technology Co ltd
Priority date: 2022-10-21
Filing date: 2022-10-21
Publication date: 2023-04-28

Abstract

The method comprises the steps of extracting dynamic change characteristics of power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including a current time point in a time sequence dimension through a time sequence encoder comprising a one-dimensional convolution layer, and encoding text descriptions of each application through a context encoder comprising an embedded layer to extract global text semantic association characteristics of each application; and then, carrying out graph structure data coding on dynamic change characteristics of power consumption data and global text semantic association characteristics of each application by using a graph neural network through the learnable neural network parameters so as to comprehensively carry out closing control of each application of the notebook computer, thereby realizing power consumption control of the notebook computer. Therefore, whether the application to be evaluated of the notebook computer is closed or not can be accurately evaluated and judged, and further the power consumption of the notebook computer is reduced at the control end, so that the performance of the notebook computer is improved.

Description

Notebook computer power consumption control method and system thereof

Technical Field

The application relates to the technical field of intelligent control, and more particularly relates to a method and a system for controlling power consumption of a notebook computer.

Background

When the notebook computer is running, a user can start a plurality of applications to meet the use requirements of the notebook computer, but after the notebook computer is used, the user often forgets to thoroughly turn off the corresponding applications, and the application hung in the background is still in the process, so that a certain amount of energy is consumed, and unnecessary power consumption is caused. If the power consumption of the notebook computer is too high, the heating is serious, and the performance of the notebook computer is seriously affected by the too high power consumption due to the limited heat dissipation capacity of the notebook computer.

Some existing solutions judge whether to automatically close based on the time when the application is not started, and in this way, although the power consumption can be saved, the time threshold is difficult to set due to different use habits of different users; second, when a user uses an application, there is often a relationship between applications, and although the frequency of use of a certain application by the user is low, the importance of the application in the whole application group is not necessarily low.

Therefore, a more optimal power consumption control scheme for notebook computers 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 power consumption control method and a system thereof for a notebook computer, wherein dynamic change characteristics of power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including a current time point in a time sequence dimension are extracted through a time sequence encoder comprising a one-dimensional convolution layer, and text descriptions of each application are encoded through a context encoder comprising an embedded layer so as to extract global text semantic association characteristics of each application; and then, carrying out graph structure data coding on dynamic change characteristics of power consumption data and global text semantic association characteristics of each application by using a graph neural network through the learnable neural network parameters so as to comprehensively carry out closing control of each application of the notebook computer, thereby realizing power consumption control of the notebook computer. Therefore, whether the application to be evaluated of the notebook computer is closed or not can be accurately evaluated and judged, and further the power consumption of the notebook computer is reduced at the control end, so that the performance of the notebook computer is improved.

According to one aspect of the present application, there is provided a method for controlling power consumption of a notebook computer, including:

acquiring power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including a current time point and text descriptions of the applications;

the power consumption data applied to a plurality of preset time points including the current time point are respectively arranged into power consumption input vectors and then pass through a time sequence encoder comprising a one-dimensional convolution layer to obtain a plurality of power consumption characteristic vectors;

performing two-dimensional matrixing on the power consumption feature vectors to obtain a power consumption feature matrix;

the text description of each application is respectively passed through a context encoder comprising an embedded layer to obtain a plurality of application text description semantic feature vectors;

calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors respectively to obtain an application association degree feature matrix;

based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix to obtain an optimized power consumption feature matrix;

the application association degree characteristic matrix and the optimized power consumption characteristic matrix pass through a graph neural network to obtain a semantic topological power consumption characteristic matrix;

Extracting row vectors corresponding to the application to be evaluated from the semantic topological power consumption feature matrix to serve as classification feature vectors; and

and the classification feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to close the application to be evaluated.

In the above method for controlling power consumption of a notebook computer, the arranging the power consumption data applied at a plurality of predetermined time points including a current time point into power consumption input vectors, respectively, and then obtaining a plurality of power consumption feature vectors by a time sequence encoder including a one-dimensional convolution layer, includes: respectively arranging the power consumption data of each application at a plurality of preset time points including the current time point into power consumption input vectors according to the time dimension; using the timing codesThe full-connection layer of the device respectively carries out full-connection coding on the power consumption input vector according to the following formula to respectively extract high-dimensional implicit characteristics of characteristic values of all positions in the power consumption input vector, wherein the formula is as follows:

wherein X is the power consumption input vector, Y is the power consumption output vector, W is the weight matrix, B is the bias vector,>

representing a matrix multiplication; and performing one-dimensional convolution coding on the power consumption input vector by using a one-dimensional convolution layer of the time sequence encoder to extract high-dimensional implicit correlation features among feature values of each position in the power consumption input vector, wherein the formula is as follows:

Wherein a is the width of the convolution kernel in the X direction, F (a) is a convolution kernel parameter vector, G (X-a) is a local vector matrix operated with a convolution kernel function, w is the size of the convolution kernel, and X represents the power consumption input vector.

In the above method for controlling power consumption of a notebook computer, the text description of each application is respectively passed through a context encoder including an embedded layer to obtain a plurality of application text description semantic feature vectors, including: word segmentation processing is carried out on the text description of each application so as to convert the text description of each application into a word sequence composed of a plurality of words; mapping each word in the word sequence to a word vector using an embedding layer of the context encoder including the embedding layer to obtain a sequence of word vectors; and performing global-based context semantic coding on the sequence of word vectors using the context encoder including the embedded layer to obtain the plurality of application text description semantic feature vectors.

In the above method for controlling power consumption of a notebook computer, the calculating the relevance between each two application text description semantic feature vectors in the plurality of application text description semantic feature vectors to obtain an application relevance feature matrix includes: calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors respectively according to the following formula to obtain a plurality of association degrees; wherein, the formula is:

Wherein V is _i And V _j Each two application text description semantic feature vectors of the plurality of application text description semantic feature vectors are represented separately,

and->

Respectively representing feature values of respective positions of each two application text description semantic feature vectors of the plurality of application text description semantic feature vectors, d (V _i ,V _j ) Representing a degree of association between each two of the plurality of application text description semantic feature vectors; and performing two-dimensional matrixing arrangement on the plurality of relevancy to obtain the application relevancy feature matrix.

In the above method for controlling power consumption of a notebook computer, the performing simultaneous optimization on the power consumption feature matrix based on the application relevance feature matrix to obtain an optimized power consumption feature matrix includes: based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix according to the following formula to obtain the optimized power consumption feature matrix; wherein, the formula is:

wherein M is ₁ And M ₂ Respectively representing the application association degree characteristic matrix and the power consumption characteristic matrix, f _i Representation ofThe applying of the eigenvalues of each position of the relevance matrix,

Is the global mean of the feature value sets of all positions in the application relevance feature matrix, N is the scale of the application relevance feature matrix, and alpha is a weighted hyper-parameter,/A>

Indicates addition by position, by position point multiplication, exp indicates an exponential operation of a value, which indicates calculation of a natural exponential function value by which the value is a power.

In the above method for controlling power consumption of a notebook computer, the step of passing the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to close an application to be evaluated, includes: performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; 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 notebook computer power consumption control system, including:

the data acquisition module is used for acquiring power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including the current time point;

the power consumption data time sequence feature extraction module is used for respectively arranging the power consumption data applied to a plurality of preset time points including the current time point into power consumption input vectors and then obtaining a plurality of power consumption feature vectors through a time sequence encoder comprising a one-dimensional convolution layer;

The two-dimensional matrixing module is used for two-dimensionally matrixing the power consumption feature vectors to obtain a power consumption feature matrix;

the context semantic coding module is used for enabling the text description of each application to pass through a context encoder comprising an embedded layer respectively so as to obtain a plurality of application text description semantic feature vectors;

the relevance calculating module is used for calculating relevance between every two application text description semantic feature vectors in the application text description semantic feature vectors to obtain an application relevance feature matrix;

the simultaneous optimization module is used for performing simultaneous optimization on the power consumption characteristic matrix based on the application association degree characteristic matrix to obtain an optimized power consumption characteristic matrix;

the graph structure data coding module is used for enabling the application association degree characteristic matrix and the optimized power consumption characteristic matrix to pass through a graph neural network so as to obtain a semantic topological power consumption characteristic matrix;

the extraction module is used for extracting row vectors corresponding to the application to be evaluated from the semantic topological power consumption feature matrix to serve as classification feature vectors; and

and the control result generation module is used for enabling the classification feature vector to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the application to be evaluated is closed or not.

Compared with the prior art, the method and the system for controlling the power consumption of the notebook computer extract dynamic change characteristics of power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including the current time point in a time sequence dimension through a time sequence encoder comprising a one-dimensional convolution layer, and code text descriptions of each application through a context encoder comprising an embedded layer so as to extract global text semantic association characteristics of each application; and then, carrying out graph structure data coding on dynamic change characteristics of power consumption data and global text semantic association characteristics of each application by using a graph neural network through the learnable neural network parameters so as to comprehensively carry out closing control of each application of the notebook computer, thereby realizing power consumption control of the notebook computer. Therefore, whether the application to be evaluated of the notebook computer is closed or not can be accurately evaluated and judged, and further the power consumption of the notebook computer is reduced at the control end, so that the performance of the notebook computer is improved.

Drawings

The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.

Fig. 1 illustrates an application scenario diagram of a notebook computer power consumption control method according to an embodiment of the present application.

Fig. 2 illustrates a flowchart of a method for controlling power consumption of a notebook computer according to an embodiment of the present application.

Fig. 3 illustrates an architecture diagram of a notebook computer power consumption control method according to an embodiment of the present application.

Fig. 4 illustrates a flowchart of a text description of each application to obtain semantic feature vectors of a plurality of application text descriptions through a context encoder including an embedded layer in a method for controlling power consumption of a notebook computer according to an embodiment of the present application.

Fig. 5 illustrates a flowchart of a method for controlling power consumption of a notebook computer according to an embodiment of the present application, in which the classification feature vector is passed through a classifier to obtain a classification result, where the classification result is used to represent whether to close an application to be evaluated.

Fig. 6 illustrates a block diagram of a notebook computer power consumption control system according to an embodiment of the present application.

Detailed Description

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.

Scene overview

As described above, when the notebook computer is running, the user can open many applications to meet the use requirements, but after the notebook computer is used, the user often forgets to turn off the corresponding applications thoroughly, and the application suspended in the background is still in the process, which consumes a certain amount of energy and causes unnecessary power consumption. Some solutions judge whether to automatically close based on the time when the application is not started, and in this way, although the power consumption can be saved, the time threshold is difficult to set due to different use habits of different users; second, when a user uses an application, there is often a relationship between applications, and although the frequency of use of a certain application by the user is low, the importance of the application in the whole application group is not necessarily low. Therefore, a more optimal power consumption control scheme for notebook computers is desired.

Specifically, in the technical scheme of the application, in order to periodically extract the power consumption implicit characteristics of each application of the notebook computer, an artificial intelligent control technology based on deep learning is adopted to deeply mine the power consumption time sequence dynamic characteristics of each application of the notebook computer. In order to evaluate the relevance features of each application to identify the importance degree of each application in the whole assembly, the global semantic features of each application need to be further extracted, and the high-dimensional implicit relevance features of each application are represented by the relevance degree of the global semantic features of each application. And then, closing control of each application of the notebook computer is comprehensively performed based on the power consumption time sequence characteristic and the application association characteristic, so that the power consumption control of the notebook computer is realized, the power consumption is reduced, and the performance of the notebook computer is improved.

Specifically, in the technical scheme of the application, first, power consumption data of a plurality of applications of a notebook computer at a plurality of predetermined time points including a current time point is acquired. Then, considering that the power consumption data of a plurality of applications of the notebook computer have regular features of dynamic property in time dimension, in order to fully mine the dynamic feature, the power consumption data of each application at a plurality of preset time points including the current time point are further arranged into power consumption input vectors respectively and then encoded in a time sequence encoder comprising a one-dimensional convolution layer, so that the dynamic change feature of the power consumption data of each application in the time sequence dimension is extracted, and a plurality of power consumption feature vectors are obtained. In a specific example of the present application, the timing encoder is composed of a fully-connected layer and a one-dimensional convolution layer which are alternately arranged, and extracts the association of the power consumption data of each application in the timing dimension through one-dimensional convolution coding and extracts the high-dimensional implicit characteristic of the power consumption data of each application through fully-connected coding. Further, the power consumption feature vectors are subjected to two-dimensional matrixing to integrate the power consumption dynamic change feature information of each application so as to obtain a power consumption feature matrix.

Then, in order to determine the importance of each application in the whole system component by judging the relevance between each application of the notebook computer so as to improve the accuracy of judging whether to close the application, text semantic information of each application needs to be extracted, and then implicit relevance characteristic information between each application is evaluated. Specifically, the text description of each application is encoded in a context encoder comprising an embedded layer, so as to extract global text semantic association feature information of each application, and a plurality of application text description semantic feature vectors are obtained. And then, respectively calculating the association degree between every two application text description semantic feature vectors in the plurality of application text description semantic feature vectors, for example, calculating cosine distances to obtain high-dimensional implicit association feature representations among the applications, so as to obtain an application association degree feature matrix.

Further, the power consumption feature vector of each application is used as the feature representation of the node, the application association degree feature matrix is used as the feature representation of the edge between the nodes, and the power consumption feature matrix and the application association degree feature matrix which are obtained by two-dimensional arrangement of the plurality of power consumption feature vectors pass through a graph neural network to obtain a semantic topological power consumption feature matrix. Specifically, the graph neural network performs graph structure data coding on the application association degree feature matrix and the power consumption feature matrix through a learnable neural network parameter to obtain the semantic topological power consumption feature matrix containing implicit association features among all applications and power consumption dynamic feature information of all applications.

When closing judgment is carried out on the application to be evaluated, row vectors corresponding to the application to be evaluated can be extracted from the semantic topological power consumption feature matrix to be classified as classification feature vectors, and whether the application to be evaluated is closed or not can be accurately evaluated and judged.

In particular, in the technical solution of the present application, before passing the application relevance feature matrix and the power consumption feature matrix through the graph neural network, it is considered that the application relevance feature matrix is obtained by a context encoder including an embedded layer and a convolutional neural network as a feature extractor for the text description of each application, and the power consumption feature matrix is obtained by a time sequence encoder for the power consumption data of each application at a plurality of predetermined time points including a current time point, so that there is a depth difference with respect to the application relevance feature matrix, thereby affecting the consistency of the feature depth expression thereof.

Thus, preferably, the application relevance feature matrix and the power consumption feature matrix are first subjected to an attention-directed hierarchical depth simultaneous optimization:

M ₁ and M ₂ The application association degree feature matrix and the power consumption feature matrix are respectively f _i Is each eigenvalue of the application relevance eigenvalue matrix,

is the global average of all feature values of the application relevance feature matrix, N is the scale of the application relevance feature matrix, i.e. width times height, and α is a weighted hyper-parameter.

Here, the application association degree feature matrix M as a deep feature ₁ As attention-directed weights, forThe power consumption characteristic matrix M as a shallow characteristic ₂ Applying a uniform attention mechanism of sub-dimension distribution to perform volume matching between high-dimensional manifolds with depth differences, so that the power consumption characteristic matrix M ₂ Can be related to the application relevance feature matrix M ₁ And carrying out simultaneous distribution with high consistency on each sub-dimension, thereby improving the consistency of the characteristic depth expression and further improving the classification accuracy. Therefore, the evaluation and judgment on whether the application to be evaluated of the notebook computer is closed can be further improved, the power consumption of the notebook computer is further reduced at the control end, and the performance of the notebook computer is improved.

Based on this, the application proposes a notebook computer power consumption control method, which includes: acquiring power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including a current time point; the power consumption data applied to a plurality of preset time points including the current time point are respectively arranged into power consumption input vectors and then pass through a time sequence encoder comprising a one-dimensional convolution layer to obtain a plurality of power consumption characteristic vectors; performing two-dimensional matrixing on the power consumption feature vectors to obtain a power consumption feature matrix; the text description of each application is respectively passed through a context encoder comprising an embedded layer to obtain a plurality of application text description semantic feature vectors; calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors respectively to obtain an application association degree feature matrix; based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix to obtain an optimized power consumption feature matrix; the application association degree characteristic matrix and the optimized power consumption characteristic matrix pass through a graph neural network to obtain a semantic topological power consumption characteristic matrix; extracting row vectors corresponding to the application to be evaluated from the semantic topological power consumption feature matrix to serve as classification feature vectors; and passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to close the application to be evaluated.

Fig. 1 illustrates an application scenario diagram of a notebook computer power consumption control method according to an embodiment of the present application. As shown in fig. 1, in the application scenario, first, power consumption data (e.g., M as illustrated in fig. 1) of a plurality of applications (e.g., F as illustrated in fig. 1) of a notebook computer at a plurality of predetermined time points including a current time point and text descriptions (e.g., C as illustrated in fig. 1) of the respective applications are acquired; then, the acquired power consumption data and text description are input into a server (e.g., S as illustrated in fig. 1) in which a notebook power consumption control system is deployed, wherein the server processes the power consumption data and the text description with a notebook power consumption control algorithm to output a classification result indicating whether to close an application to be evaluated.

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.

Exemplary method

Fig. 2 illustrates a flowchart of a method for controlling power consumption of a notebook computer according to an embodiment of the present application. As shown in fig. 2, a method for controlling power consumption of a notebook computer according to an embodiment of the present application includes: s110, acquiring power consumption data of a plurality of applications of the notebook computer at a plurality of preset time points including a current time point and text description of each application; s120, respectively arranging the power consumption data applied to a plurality of preset time points including the current time point into power consumption input vectors, and then obtaining a plurality of power consumption characteristic vectors through a time sequence encoder comprising a one-dimensional convolution layer; s130, carrying out two-dimensional matrixing on the power consumption feature vectors to obtain a power consumption feature matrix; s140, text descriptions of the applications are respectively passed through a context encoder comprising an embedded layer to obtain a plurality of application text description semantic feature vectors; s150, respectively calculating the association degree between every two application text description semantic feature vectors in the plurality of application text description semantic feature vectors to obtain an application association degree feature matrix; s160, based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix to obtain an optimized power consumption feature matrix; s170, passing the application association degree feature matrix and the optimized power consumption feature matrix through a graph neural network to obtain a semantic topology power consumption feature matrix; s180, extracting row vectors corresponding to the application to be evaluated from the semantic topological power consumption feature matrix to serve as classification feature vectors; and S190, passing the classification feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to close the application to be evaluated.

Fig. 3 illustrates an architecture diagram of a notebook computer power consumption control method according to an embodiment of the present application. As shown in fig. 3, in the network architecture of the method for controlling power consumption of a notebook computer, first, power consumption data of a plurality of applications of the notebook computer at a plurality of predetermined time points including a current time point and text descriptions of the respective applications are acquired; then, respectively arranging the power consumption data applied to a plurality of preset time points including the current time point into power consumption input vectors, and then obtaining a plurality of power consumption characteristic vectors through a time sequence encoder comprising a one-dimensional convolution layer; then, carrying out two-dimensional matrixing on the power consumption feature vectors to obtain a power consumption feature matrix; then, the text description of each application is respectively passed through a context encoder comprising an embedded layer to obtain a plurality of application text description semantic feature vectors; then, calculating the association degree between every two application text description semantic feature vectors in the plurality of application text description semantic feature vectors respectively to obtain an application association degree feature matrix; then, based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix to obtain an optimized power consumption feature matrix; then, the application association degree characteristic matrix and the optimized power consumption characteristic matrix pass through a graph neural network to obtain a semantic topological power consumption characteristic matrix; then, extracting a row vector corresponding to the application to be evaluated from the semantic topological power consumption feature matrix to serve as a classification feature vector; and finally, the classification feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to close the application to be evaluated.

In step S110, power consumption data of a plurality of applications of the notebook computer at a plurality of predetermined time points including a current time point and text descriptions of the respective applications are acquired. As described above, when the notebook computer is running, the user can open many applications to meet the use requirement, but after the notebook computer is used, the user often forgets to turn off the corresponding applications thoroughly, and the application suspended in the background is still in the process, which consumes a certain amount of energy and causes unnecessary power consumption. Some solutions judge whether to automatically close based on the time when the application is not started, and in this way, although the power consumption can be saved, the time threshold is difficult to set due to different use habits of different users; second, when a user uses an application, there is often a relationship between applications, and although the frequency of use of a certain application by the user is low, the importance of the application in the whole application group is not necessarily low. Therefore, a more optimal power consumption control scheme for notebook computers is desired.

Specifically, in the technical scheme of the application, first, power consumption data of a plurality of applications of a notebook computer at a plurality of preset time points including a current time point and text descriptions of the applications are obtained.

In step S120 and step S130, the power consumption data applied at a plurality of predetermined time points including the current time point are respectively arranged into power consumption input vectors, and then a plurality of power consumption feature vectors are obtained through a time sequence encoder including a one-dimensional convolution layer, and then two-dimensional matrixing is performed on the plurality of power consumption feature vectors to obtain a power consumption feature matrix. It should be understood that, after obtaining the power consumption data of the applications of the notebook computer at a plurality of predetermined time points including the current time point, the power consumption data of the applications of the notebook computer are considered to have a regular characteristic of dynamic property in the time dimension.

Therefore, in order to fully mine out the dynamic characteristics, the power consumption data of each application at a plurality of preset time points including the current time point are further arranged into power consumption input vectors respectively and then are encoded in a time sequence encoder comprising a one-dimensional convolution layer, so that the dynamic change characteristics of the power consumption data of each application in the time sequence dimension are extracted, and a plurality of power consumption characteristic vectors are obtained. In a specific example of the present application, the timing encoder is composed of a fully-connected layer and a one-dimensional convolution layer which are alternately arranged, and extracts the association of the power consumption data of each application in the timing dimension through one-dimensional convolution coding and extracts the high-dimensional implicit characteristic of the power consumption data of each application through fully-connected coding.

Specifically, in the embodiment of the present application, after the power consumption data applied to the respective predetermined time points including the current time point are respectively arranged into the power consumption input vectors, the power consumption input vectors are obtained by a time sequence encoder including a one-dimensional convolution layer, which includes: respectively arranging the power consumption data of each application at a plurality of preset time points including the current time point into power consumption input vectors according to the time dimension; and respectively performing full-connection coding on the power consumption input vectors by using a full-connection layer of the time sequence coder according to the following formula to respectively extract high-dimensional implicit characteristics of characteristic values of all positions in the power consumption input vectors, wherein the formula is as follows:

wherein X is the power consumption input vector and Y isPower consumption output vector, W is weight matrix, B is bias vector, +.>

Further, the power consumption feature vectors are subjected to two-dimensional matrixing to integrate the power consumption dynamic change feature information of each application so as to obtain a power consumption feature matrix.

In step S140, the text descriptions of the respective applications are passed through a context encoder including an embedded layer, respectively, to obtain a plurality of application text description semantic feature vectors. It should be understood that, in order to determine the importance of each application in the whole system component by judging the relevance between each application of the notebook computer, so as to improve the accuracy of judging whether to close the application, text semantic information of each application needs to be extracted, so as to evaluate the implicit relevance characteristic information between each application.

Specifically, the text description of each application is encoded in a context encoder comprising an embedded layer, so as to extract global text semantic association feature information of each application, and a plurality of application text description semantic feature vectors are obtained. In this embodiment of the present application, fig. 4 illustrates a flowchart of respectively passing text descriptions of the respective applications through a context encoder including an embedded layer to obtain a plurality of application text description semantic feature vectors in a method for controlling power consumption of a notebook computer according to an embodiment of the present application, and as shown in fig. 4, respectively passing text descriptions of the respective applications through a context encoder including an embedded layer to obtain a plurality of application text description semantic feature vectors, including: s210, performing word segmentation processing on the text description of each application to convert the text description of each application into a word sequence composed of a plurality of words; s220, mapping each word in the word sequence to a word vector by using an embedding layer of the context encoder comprising the embedding layer to obtain a sequence of word vectors; and S230, performing global-based context semantic coding on the sequence of word vectors by using the context encoder comprising the embedded layer to obtain the plurality of application text description semantic feature vectors.

In step S150, the relevance between each two application text description semantic feature vectors in the plurality of application text description semantic feature vectors is calculated to obtain an application relevance feature matrix. That is, the relevance between every two application text description semantic feature vectors in the application text description semantic feature vectors is calculated respectively, for example, a cosine distance is calculated to obtain a high-dimensional implicit relevance feature representation between the applications, so as to obtain an application relevance feature matrix.

Further, in an embodiment of the present application, the calculating the relevance between each two application text description semantic feature vectors in the plurality of application text description semantic feature vectors to obtain an application relevance feature matrix includes: calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors respectively according to the following formula to obtain a plurality of association degrees; wherein, the formula is:

wherein V is _i And V _j Each two application text description semantic feature vectors in the plurality of application text description semantic feature vectors are respectively represented，

And->

In step S160, based on the application association degree feature matrix, the power consumption feature matrix is simultaneously optimized to obtain an optimized power consumption feature matrix. In particular, in the technical solution of the present application, before passing the application relevance feature matrix and the power consumption feature matrix through the graph neural network, it is considered that the application relevance feature matrix is obtained by a context encoder including an embedded layer and a convolutional neural network as a feature extractor for the text description of each application, and the power consumption feature matrix is obtained by a time sequence encoder for the power consumption data of each application at a plurality of predetermined time points including a current time point, so that there is a depth difference with respect to the application relevance feature matrix, thereby affecting the consistency of the feature depth expression thereof.

Therefore, preferably, the application relevance feature matrix and the power consumption feature matrix are subjected to attention-oriented hierarchical depth simultaneous optimization, that is, based on the application relevance feature matrix, the power consumption feature matrix is subjected to simultaneous optimization according to the following formula to obtain the optimized power consumption feature matrix; wherein, the formula is:

Wherein M is ₁ And M ₂ Respectively represent the said shouldUsing the association degree characteristic matrix and the power consumption characteristic matrix, f _i A feature value representing each position of the application association feature matrix,

Here, the application association degree feature matrix M as a deep feature ₁ As a attention-directing weight, for said power consumption feature matrix M as shallow features ₂ Applying a uniform attention mechanism of sub-dimension distribution to perform volume matching between high-dimensional manifolds with depth differences, so that the power consumption characteristic matrix M ₂ Can be related to the application relevance feature matrix M ₁ And carrying out simultaneous distribution with high consistency on each sub-dimension, thereby improving the consistency of the characteristic depth expression and further improving the classification accuracy. Therefore, the evaluation and judgment on whether the application to be evaluated of the notebook computer is closed can be further improved, the power consumption of the notebook computer is further reduced at the control end, and the performance of the notebook computer is improved.

In step S170 and step S180, the application association degree feature matrix and the optimized power consumption feature matrix are passed through a graph neural network to obtain a semantic topological power consumption feature matrix, and then row vectors corresponding to the application to be evaluated are extracted from the semantic topological power consumption feature matrix as classification feature vectors. Further, the power consumption feature vector of each application is used as the feature representation of the node, the application association degree feature matrix is used as the feature representation of the edge between the nodes, and the power consumption feature matrix and the application association degree feature matrix which are obtained by two-dimensional arrangement of the plurality of power consumption feature vectors pass through a graph neural network to obtain a semantic topological power consumption feature matrix.

Specifically, the graph neural network performs graph structure data coding on the application association degree feature matrix and the power consumption feature matrix through a learnable neural network parameter to obtain the semantic topological power consumption feature matrix containing implicit association features among all applications and power consumption dynamic feature information of all applications. When closing judgment is carried out on the application to be evaluated, row vectors corresponding to the application to be evaluated can be extracted from the semantic topological power consumption feature matrix to be classified as classification feature vectors, and whether the application to be evaluated is closed or not can be accurately evaluated and judged.

In step S190, the classification feature vector is passed through a classifier to obtain a classification result, where the classification result is used to indicate whether to close the application to be evaluated.

Specifically, in the embodiment of the present application, fig. 5 illustrates a flowchart for passing the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to close an application to be evaluated, and as shown in fig. 5, the classifying feature vector is passed through the classifier to obtain a classification result, where the classification result is used to indicate whether to close the application to be evaluated, where the classifying result includes: s310, performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and S320, inputting the coding classification feature vector into a Softmax classification function of the classifier to obtain the classification result.

In a specific example of the present application, the classifier is used to process the classification feature vector to generate a classification result according to the following formula:

O＝softmax{(W _n ,B _n ):…:(W ₁ ,B ₁ ) X, wherein X represents the classification feature vector, W ₁ To W _n Weight matrix for all the connection layers of each layer, B ₁ To B _n Representing the bias vector for each fully connected layer.

In summary, a method and a system for controlling power consumption of a notebook computer according to embodiments of the present application are illustrated, in which dynamic variation features of power consumption data of a plurality of applications of the notebook computer at a plurality of predetermined time points including a current time point in a time sequence dimension are extracted by a time sequence encoder including a one-dimensional convolution layer, and text descriptions of each application are encoded by a context encoder including an embedding layer to extract global text semantic association features of each application; and then, carrying out graph structure data coding on dynamic change characteristics of power consumption data and global text semantic association characteristics of each application by using a graph neural network through the learnable neural network parameters so as to comprehensively carry out closing control of each application of the notebook computer, thereby realizing power consumption control of the notebook computer. Therefore, whether the application to be evaluated of the notebook computer is closed or not can be accurately evaluated and judged, and further the power consumption of the notebook computer is reduced at the control end, so that the performance of the notebook computer is improved.

Exemplary System

Fig. 6 illustrates a block diagram of a notebook computer power consumption control system according to an embodiment of the present application. As shown in fig. 6, a notebook computer power consumption control system 100 according to an embodiment of the present application includes: a data acquisition module 110, configured to acquire power consumption data of a plurality of applications of the notebook computer at a plurality of predetermined time points including a current time point; a power consumption data timing characteristic extraction module 120, configured to arrange the power consumption data applied at a plurality of predetermined time points including a current time point into power consumption input vectors respectively, and then obtain a plurality of power consumption characteristic vectors through a timing encoder including a one-dimensional convolution layer; a two-dimensional matrixing module 130, configured to two-dimensionally matrize the multiple power consumption feature vectors to obtain a power consumption feature matrix; a context semantic coding module 140, configured to pass the text descriptions of the respective applications through a context encoder including an embedded layer to obtain a plurality of application text description semantic feature vectors; the relevance calculating module 150 is configured to calculate relevance between each two application text description semantic feature vectors in the plurality of application text description semantic feature vectors to obtain an application relevance feature matrix; the simultaneous optimization module 160 is configured to perform simultaneous optimization on the power consumption feature matrix based on the application association degree feature matrix to obtain an optimized power consumption feature matrix; the graph structure data encoding module 170 is configured to pass the application association degree feature matrix and the optimized power consumption feature matrix through a graph neural network to obtain a semantic topology power consumption feature matrix; an extracting module 180, configured to extract, from the semantic topological power consumption feature matrix, a row vector corresponding to an application to be evaluated as a classification feature vector; and a control result generating module 190, configured to pass the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to close the application to be evaluated.

In one example, in the above-mentioned notebook computer power consumption control system 100, the power consumption data timing feature extraction module includes: a power consumption input vector construction unit for arranging the power consumption data of each application at a plurality of preset time points including the current time point into power consumption input vectors according to the time dimension; the full-connection coding unit is used for respectively carrying out full-connection coding on the power consumption input vectors by using the full-connection layer of the time sequence coder according to the following formula to respectively extract high-dimensional implicit characteristics of characteristic values of all positions in the power consumption input vectors, wherein the formula is as follows:

representing a matrix multiplication; and a one-dimensional convolution encoding unit, configured to perform one-dimensional convolution encoding on the power consumption input vectors by using a one-dimensional convolution layer of the timing encoder to extract high-dimensional implicit correlation features between feature values of each position in the power consumption input vectors, where the formula is:

In one example, in the above-mentioned notebook computer power consumption control system 100, the context semantic coding module includes: a word segmentation unit, configured to perform word segmentation processing on the text descriptions of the applications to convert the text descriptions of the applications into word sequences composed of a plurality of words; an embedded encoding unit, configured to map each word in the word sequence to a word vector using an embedded layer of the context encoder including the embedded layer to obtain a sequence of word vectors; and a context coding unit, configured to perform global-based context semantic coding on the sequence of word vectors using the context encoder including the embedded layer to obtain the plurality of application text description semantic feature vectors.

In one example, in the above-mentioned notebook computer power consumption control system 100, the association calculation module includes: correlation degree calculation unit: the method comprises the steps of respectively calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors to obtain a plurality of association degrees according to the following formula; wherein, the formula is:

And->

Representing the plurality of applications respectivelyEvery two application texts in the text description semantic feature vector describe feature values of the respective positions of the text description semantic feature vector, d (V _i ,V _j ) Representing a degree of association between each two of the plurality of application text description semantic feature vectors; and a two-dimensional matrixing unit for two-dimensionally matrixing the plurality of relevancy to obtain the application relevancy feature matrix.

In one example, in the above-mentioned notebook computer power consumption control system 100, the simultaneous optimization module includes: based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix according to the following formula to obtain the optimized power consumption feature matrix; wherein, the formula is:

wherein M is ₁ And M ₂ Respectively representing the application association degree characteristic matrix and the power consumption characteristic matrix, f _i A feature value representing each position of the application association feature matrix,

In one example, in the above-mentioned notebook computer power consumption control system 100, the control result generating module includes: the full-connection coding unit is used for carrying out full-connection coding on the classification feature vectors by using a full-connection layer of the classifier so as to obtain coded classification feature vectors; and a classification result generating unit for 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 units and modules in the above-described notebook computer power consumption control system 100 have been described in detail in the above description of the notebook computer power consumption control method with reference to fig. 1 to 5, and thus, repetitive descriptions thereof will be omitted.

As described above, the notebook computer power consumption control system 100 according to the embodiment of the present application may be implemented in various terminal devices, such as a server or the like for notebook computer power consumption control. In one example, the notebook computer power consumption control system 100 according to embodiments of the present application may be integrated into the terminal device as one software module and/or hardware module. For example, the notebook computer power consumption control system 100 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the notebook computer power consumption control system 100 can also be one of a plurality of hardware modules of the terminal device.

Alternatively, in another example, the notebook power consumption control system 100 and the terminal device may be separate devices, and the notebook power consumption control system 100 may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information in a contracted data format.

The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.

The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.

It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.

Claims

1. The power consumption control method of the notebook computer is characterized by comprising the following steps:

2. The method for controlling power consumption of a notebook computer according to claim 1, wherein the step of arranging the power consumption data of each application at a plurality of predetermined time points including a current time point into power consumption input vectors and obtaining a plurality of power consumption feature vectors by a time sequence encoder including a one-dimensional convolution layer, respectively, comprises:

respectively arranging the power consumption data of each application at a plurality of preset time points including the current time point into power consumption input vectors according to the time dimension;

the full-connection layer of the time sequence encoder is used for respectively carrying out full-connection encoding on the power consumption input vector in the following formula so as to respectively extract the high values of the characteristic values of all the positions in the power consumption input vectorDimensional implicit features, wherein the formula is:

representing a matrix multiplication; and

and respectively carrying out one-dimensional convolution coding on the power consumption input vectors by using a one-dimensional convolution layer of the time sequence encoder to respectively extract high-dimensional implicit correlation features among feature values of each position in the power consumption input vectors, wherein the formula is as follows:

3. The method for controlling power consumption of a notebook computer according to claim 2, wherein said passing the text descriptions of the respective applications through a context encoder including an embedded layer to obtain a plurality of application text description semantic feature vectors, respectively, comprises:

word segmentation processing is carried out on the text description of each application so as to convert the text description of each application into a word sequence composed of a plurality of words;

mapping each word in the word sequence to a word vector using an embedding layer of the context encoder including the embedding layer to obtain a sequence of word vectors; and

global-based context semantic coding of the sequence of word vectors using the context encoder including an embedded layer to obtain the plurality of application text description semantic feature vectors.

4. The method for controlling power consumption of a notebook computer according to claim 3, wherein calculating the degree of association between each two application text description semantic feature vectors of the plurality of application text description semantic feature vectors to obtain the application degree of association feature matrix includes:

Calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors respectively according to the following formula to obtain a plurality of association degrees;

wherein, the formula is:

and->

Respectively representing feature values of respective positions of each two application text description semantic feature vectors of the plurality of application text description semantic feature vectors, d (V _i ,V _j ) Representing a degree of association between each two of the plurality of application text description semantic feature vectors; and

and carrying out two-dimensional matrixing arrangement on the plurality of relevancy to obtain the application relevancy feature matrix.

5. The method for controlling power consumption of a notebook computer according to claim 4, wherein the performing simultaneous optimization on the power consumption feature matrix based on the application relevance feature matrix to obtain an optimized power consumption feature matrix includes:

based on the application association degree feature matrix, performing simultaneous optimization on the power consumption feature matrix according to the following formula to obtain the optimized power consumption feature matrix;

Wherein, the formula is:

6. The method for controlling power consumption of a notebook computer according to claim 5, wherein the step of passing the classification feature vector through a classifier to obtain a classification result, the classification result being used for indicating whether to close an application to be evaluated, includes:

performing full-connection coding on the classification feature vectors by using a full-connection layer of the classifier to obtain coded classification feature vectors; and

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

7. A notebook computer power consumption control system, comprising:

8. The notebook computer power consumption control system of claim 7, wherein the power consumption data timing feature extraction module comprises:

a power consumption input vector construction unit for arranging the power consumption data of each application at a plurality of preset time points including the current time point into power consumption input vectors according to the time dimension;

the full-connection coding unit is used for respectively carrying out full-connection coding on the power consumption input vectors by using the full-connection layer of the time sequence coder according to the following formula to respectively extract high-dimensional implicit characteristics of characteristic values of all positions in the power consumption input vectors, wherein the formula is as follows:

representing a matrix multiplication; and

the one-dimensional convolution coding unit is used for respectively carrying out one-dimensional convolution coding on the power consumption input vector by using a one-dimensional convolution layer of the time sequence coder to respectively extract high-dimensional implicit correlation features among feature values of all positions in the power consumption input vector, wherein the formula is as follows:

9. The notebook computer power consumption control system of claim 8, wherein the context semantic coding module comprises:

a word segmentation unit, configured to perform word segmentation processing on the text descriptions of the applications to convert the text descriptions of the applications into word sequences composed of a plurality of words;

an embedded encoding unit, configured to map each word in the word sequence to a word vector using an embedded layer of the context encoder including the embedded layer to obtain a sequence of word vectors; and

And the context coding unit is used for carrying out global-based context semantic coding on the sequence of the word vectors by using the context coder comprising the embedded layer so as to obtain the plurality of application text description semantic feature vectors.

10. The notebook computer power consumption control system of claim 9, wherein the association calculation module comprises:

correlation degree calculation unit: the method comprises the steps of respectively calculating the association degree between every two application text description semantic feature vectors in the application text description semantic feature vectors to obtain a plurality of association degrees according to the following formula;

wherein, the formula is:

and->

and the two-dimensional matrixing unit is used for carrying out two-dimensional matrixing on the plurality of relevancy to obtain the application relevancy feature matrix.