CN113743584A - Integral gradient model prediction method, system, electronic device and storage medium - Google Patents
Integral gradient model prediction method, system, electronic device and storage medium Download PDFInfo
- Publication number
- CN113743584A CN113743584A CN202110937966.8A CN202110937966A CN113743584A CN 113743584 A CN113743584 A CN 113743584A CN 202110937966 A CN202110937966 A CN 202110937966A CN 113743584 A CN113743584 A CN 113743584A
- Authority
- CN
- China
- Prior art keywords
- model
- multitask
- training
- sample
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 238000012549 training Methods 0.000 claims abstract description 69
- 238000003062 neural network model Methods 0.000 claims abstract description 29
- 238000010276 construction Methods 0.000 claims abstract description 13
- 230000001364 causal effect Effects 0.000 claims description 18
- 238000004590 computer program Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 241000282414 Homo sapiens Species 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000013145 classification model Methods 0.000 description 4
- 239000013598 vector Substances 0.000 description 4
- 230000008451 emotion Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a method, a system, electronic equipment and a storage medium for predicting an integral gradient model, wherein the technical scheme of the method comprises the steps of training an original model, defining a training set, constructing a neural network model, and training the neural network model on the training set; a multitask training set construction step, namely initializing an empty multitask data set, selecting a sample from the training set which is not put back, and constructing the multitask data set according to the sample; a multitask model construction step, namely adding an attribution structure in the trained neural network model to obtain a multitask model, and training the multitask model on the multitask data set; and a multi-task model prediction step, namely predicting in an application scene by using the trained multi-task model. The method solves the problems that the existing method is low in attribution speed and cannot give attribution of a prediction result while the model predicts.
Description
Technical Field
The invention belongs to the technical field of machine learning, and particularly relates to an integral gradient model prediction method, an integral gradient model prediction system, electronic equipment and a storage medium.
Background
In recent years, deep learning plays an important role in the fields of computer vision, natural language processing, personalized recommendation and the like. However, as the effect of the model on various tasks is improved, the structure of the model becomes more and more complex, so that the behavior of the model cannot be well explained. Thus, the deep learning model is also often referred to as a "black box" model. Recently, researchers have begun to turn their attention to model interpretability, expecting to be able to uncover the "black box" model to some extent.
In addition, the explanation is provided for the prediction behavior of the model, so that the trust of people on the model can be enhanced, the deployment rate of the model is improved, errors in the model can be found, and the model is further improved. For example, for a text emotion classification model with an accuracy of more than 95%, one would typically choose to trust the model and deploy it. However, if in the training and test samples, all negative samples contain some special characters, but none of the positive samples. Then, the model may not learn the rule of judging the positive and negative samples correctly, and is classified only by special symbols. Clearly, such a model is problematic, and interpretability can assist people in quickly locating such errors.
The integral gradient method is a gradient-based model interpretation method proposed by Sundararajan et al, and theoretically proves the superiority of the method. However, this method requires several tens of gradient calculations to be performed for each sample to be interpreted, which consumes a lot of time, preventing the method from being applied in practice. At present, the integral gradient method has two main disadvantages: the attribution of a single sample requires more than 20 times of gradient calculation, and the attribution speed is greatly slowed down; the model cannot predict and attribute the prediction result.
Disclosure of Invention
The embodiment of the application provides an integral gradient model prediction method, an integral gradient model prediction system, electronic equipment and a storage medium, and at least solves the problems that the existing method is low in attribution speed and cannot give attribution of a prediction result while the model predicts.
In a first aspect, an embodiment of the present application provides an integral gradient model prediction method, including: an original model training step, defining a training set, constructing a neural network model, and training the neural network model on the training set; a multitask training set construction step, namely initializing an empty multitask data set, selecting a sample from the training set which is not put back, and constructing the multitask data set according to the sample; a multitask model construction step, namely adding an attribution structure in the trained neural network model to obtain a multitask model, and training the multitask model on the multitask data set; and a multi-task model prediction step, namely predicting in an application scene by using the trained multi-task model.
Preferably, the multitask training set constructing step further comprises: predicting a signature of the sample using the neural network model and generating a causal baseline sample based on the sample, calculating a causal, and adding the reduced cost, the signature, and the causal as new samples to the multitask dataset.
Preferably, the attribution is calculated using an integral gradient method.
Preferably, the multitask model predicting step further comprises: when model prediction is carried out, a prediction result and an attribution result are simultaneously output.
In a second aspect, an integral gradient model prediction system is provided in an embodiment of the present application, and is applicable to the integral gradient model prediction method, including: the original model training module defines a training set, constructs a neural network model and trains the neural network model on the training set; the multitask training set construction module is used for initializing an empty multitask data set, selecting a sample from the training set which is not put back, and constructing the multitask data set according to the sample; the multitask model construction module is used for adding an attribution structure into the trained neural network model to obtain a multitask model and training the multitask model on the multitask data set; and the multi-task model prediction module is used for predicting in an application scene by using the trained multi-task model.
In some embodiments, the multitask training set constructing module further comprises: predicting a signature of the sample using the neural network model and generating a causal baseline sample based on the sample, calculating a causal, and adding the reduced cost, the signature, and the causal as new samples to the multitask dataset.
In some of these embodiments, the attribution is calculated using an integral gradient method.
In some of these embodiments, the multitask model prediction module further comprises: when model prediction is carried out, a prediction result and an attribution result are simultaneously output.
In a third aspect, an embodiment of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor, when executing the computer program, implements an integral gradient model prediction method as described in the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements an integral gradient model prediction method as described in the first aspect above.
The method and the device can be applied to the technical field of deep learning. Compared with the related art, the integral gradient model prediction method provided by the embodiment of the application constructs a multitask data set based on an integral gradient method, and model prediction and attribution are converted into a plurality of tasks on the same model, so that prediction attribution is given while prediction is achieved, and attribution speed is greatly improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a flow chart of an integrated gradient model prediction method of the present invention;
FIG. 2 is a block diagram of an integrated gradient model prediction system of the present invention;
FIG. 3 is a block diagram of an electronic device of the present invention;
in the above figures:
1. an original model training module; 2. a multitask training set construction module; 3. a multitask model construction module; 4. a multitask model prediction module; 60. a bus; 61. a processor; 62. a memory; 63. a communication interface.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.
It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that in the development of any such actual implementation, as in any engineering or design project, numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints, which may vary from one implementation to another.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.
Unless defined otherwise, technical or scientific terms referred to herein shall have the ordinary meaning as understood by those of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The present application is directed to the use of the terms "including," "comprising," "having," and any variations thereof, which are intended to cover non-exclusive inclusions; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
First a formal definition of the cause is given:
let a neural network be a function F:
the input of the network is
The prediction value of sample x is f (x). Then the attribute to input x is represented by the predicted value F (x)
Wherein, aiCan be regarded as xiFor the contribution of the predicted value F (x), when ai>When 0, it represents xiHas a positive contribution to the prediction, ai<0 then represents xiContributing negatively to the prediction.
Intuitively, attribution is the basis for finding a model for prediction. Take a text emotion classification model as an example. If the model considers the text "this product is really junk! "has negative emotion. Then attribution requires finding the contribution of each word (or word) to the prediction. Obviously, the word "garbage" should contribute more to the prediction of the model, and its corresponding attribution score should also be larger.
The integral gradient method is a gradient-based model interpretation method proposed by sundaraajan et al. The principle and implementation details of the method are presented below.
1. How human beings attribute to
Human attribution generally relies on the counterintuitive idea that when a human attributing some responsibility to a cause, implicitly compares the absence of that cause as a baseline. For example, if a human determines that "hungry" is not the reason for "wanting to eat", then it is compared to "not hungry" if the human still wants to eat.
2. Attribution of baselines
Based on the principle of human attribution, a sample called "attribution baseline" is used in the integral gradient method as a control in the absence of the original sample. If the input sample for the model is x, then the "attributed baseline" sample is denoted as x'. X 'in the two scenes of image classification and text classification are described below to better understand the meaning of x'.
In the image classification model, the input sample x is a normal image to be classified, and x' may be a pure black image. Thus x' can be used as a control in the absence of x.
In a text classification model, words (or words) are typically converted to embedded vectors. Thus, if x represents an embedded vector of the input text, x' is an all-0 embedded vector of a length equal to x.
3. Integral gradient
Let the function F:
a deep neural network is represented that is,
which represents the input samples of the sample to be tested,
indicating an "attributed baseline" sample. x is the number ofiRepresents the first of the input samples xi components, aiThen is xiThe corresponding attribution score. Integral gradient method for calculating aiIs composed of
Wherein the content of the first and second substances,
represents the gradient of the function f (x) in the ith dimension of the input sample x.
A more intuitive explanation of this approach is given below. Both x and x' are n-dimensional vectors, so these two samples can be viewed as two points in n-dimensional space. There must be a line l connecting the two points in space, and the function F can be microminiature everywhere on the line l. The integral gradient method is to integrate all gradients of the function f (x) on the line l.
4. Calculating integral gradients
It is clear that the integral calculation is difficult for a computer. Thus, to facilitate the calculation of the gradient integral, the integral gradient can be approximated by selecting m points at equal intervals from the line l and adding the integrals of these points, i.e. the integral gradient
The larger m is, the more approximate it is, but the larger the amount of calculation is. In practice, m may be between 20 and 300.
Embodiments of the invention are described in detail below with reference to the accompanying drawings:
the core idea of the invention is to construct a multi-task training set and a multi-task model, so that the model can be attributed while predicting. Specifically, the method can be divided into four stages, which are respectively: original model training, constructing a multi-task training set, constructing a multi-task model, and predicting by using the multi-task model.
Fig. 1 is a flowchart of an integral gradient model prediction method of the present invention, please refer to fig. 1, the integral gradient model prediction method of the present invention includes the following steps:
s1: defining a training set, constructing a neural network model, and training the neural network model on the training set.
In the implementation, the original model training is performed first, and a training set D ═ (X, Y) is set, where n samples are included, that is, | D | ═ n; and constructing a neural network model, and training on a training set D, wherein the trained model is called F.
S2: initializing an empty multitask data set, selecting a sample from the training set which is not replaced, and constructing the multitask data set according to the sample.
Optionally, the neural network model is used to predict a signature of the sample, and a causal baseline sample is generated based on the sample, a causal is calculated, and the sample, the signature, and the causal are added as a new sample to the multitask dataset. Alternatively, the attribution is calculated using an integral gradient method.
In the specific implementation, an empty multitask training set is initialized, namely D' { }; selecting a sample X from the samples X without being replaced, namely X ═ X/{ X }; using model F to predict the label of x, i.e., y' ═ F (x); generating an "attributed baseline" sample x' based on sample x; integral gradient method calculation attribute, namely a ═ IG (x, x ', y', F), IG (·) represents integral gradient method; adding x, y ' and a as new samples to D ', i.e. D ' ═ D ' @ goute (x, y ', a); if X is not empty, continuously selecting a sample which is not put back from X to be repeatedly executed, otherwise, ending the stage.
S3: and adding an attribution structure in the trained neural network model to obtain a multi-task model, and training the multi-task model on the multi-task data set.
In the specific implementation, F is taken as a pre-training model, and a structure for attribution is added on the basis of F, so that a multitask model F' is finally constructed; let model F 'be trained on training set D'.
S4: and predicting in an application scene by using the trained multitask model.
Optionally, when performing model prediction, the prediction result and the attribution result are output at the same time.
In a specific implementation, when the model F ' performs prediction, the prediction result and the attribution result, i.e., y ', a ═ F ' (x), are output simultaneously.
It should be noted that the steps illustrated in the above-described flow diagrams or in the flow diagrams of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow diagrams, in some cases, the steps illustrated or described may be performed in an order different than here.
The embodiment of the application provides an integral gradient model prediction system, which is suitable for the integral gradient model prediction method. As used below, the terms "unit," "module," and the like may implement a combination of software and/or hardware of predetermined functions. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware or a combination of software and hardware is also possible and contemplated.
FIG. 2 is a block diagram of an integral gradient model prediction system according to the present invention, please refer to FIG. 2, which includes:
original model training module 1: defining a training set, constructing a neural network model, and training the neural network model on the training set.
In the implementation, the original model training is performed first, and a training set D ═ (X, Y) is set, where n samples are included, that is, | D | ═ n; and constructing a neural network model, and training on a training set D, wherein the trained model is called F.
The multi-task training set constructing module 2: initializing an empty multitask data set, selecting a sample from the training set which is not replaced, and constructing the multitask data set according to the sample.
Optionally, the neural network model is used to predict a signature of the sample, and a causal baseline sample is generated based on the sample, a causal is calculated, and the reduced cost, the signature, and the causal are added as new samples to the multitask dataset. Alternatively, the attribution is calculated using an integral gradient method.
In the specific implementation, an empty multitask training set is initialized, namely D' { }; selecting a sample X from the samples X without being replaced, namely X ═ X/{ X }; using model F to predict the label of x, i.e., y' ═ F (x); generating an "attributed baseline" sample x' based on sample x; integral gradient method calculation attribute, namely a ═ IG (x, x ', y', F), IG (·) represents integral gradient method; adding x, y ' and a as new samples to D ', i.e. D ' ═ D ' @ goute (x, y ', a); if X is not empty, continuously selecting a sample which is not put back from X to be repeatedly executed, otherwise, ending the stage.
Multitask model construction module 3: and adding an attribution structure in the trained neural network model to obtain a multi-task model, and training the multi-task model on the multi-task data set.
In the specific implementation, F is taken as a pre-training model, and a structure for attribution is added on the basis of F, so that a multitask model F' is finally constructed; let model F 'be trained on training set D'.
The multitask model prediction module 4: and predicting in an application scene by using the trained multitask model.
Optionally, when performing model prediction, the prediction result and the attribution result are output at the same time.
In a specific implementation, when the model F ' performs prediction, the prediction result and the attribution result, i.e., y ', a ═ F ' (x), are output simultaneously.
Additionally, one method of integral gradient model prediction described in connection with FIG. 1 may be implemented by an electronic device. Fig. 3 is a block diagram of an electronic device of the present invention.
The electronic device may comprise a processor 61 and a memory 62 in which computer program instructions are stored.
Specifically, the processor 61 may include a Central Processing Unit (CPU), or A Specific Integrated Circuit (ASIC), or may be configured to implement one or more Integrated circuits of the embodiments of the present Application.
The memory 62 may be used to store or cache various data files that need to be processed and/or used for communication, as well as possible computer program instructions executed by the processor 61.
The processor 61 implements any one of the methods of integral gradient model prediction in the above embodiments by reading and executing computer program instructions stored in the memory 62.
In some of these embodiments, the electronic device may also include a communication interface 63 and a bus 60. As shown in fig. 3, the processor 61, the memory 62, and the communication interface 63 are connected via a bus 60 to complete communication therebetween.
The communication port 63 may be implemented with other components such as: the data communication is carried out among external equipment, image/data acquisition equipment, a database, external storage, an image/data processing workstation and the like.
The bus 60 includes hardware, software, or both to couple the components of the electronic device to one another. Bus 60 includes, but is not limited to, at least one of the following: data Bus (Data Bus), Address Bus (Address Bus), Control Bus (Control Bus), Expansion Bus (Expansion Bus), and Local Bus (Local Bus). By way of example, and not limitation, Bus 60 may include an Accelerated Graphics Port (AGP) or other Graphics Bus, an Enhanced Industry Standard Architecture (EISA) Bus, a Front-Side Bus (FSB), a Hyper Transport (HT) Interconnect, an ISA (ISA) Bus, an InfiniBand (InfiniBand) Interconnect, a Low Pin Count (LPC) Bus, a memory Bus, a microchannel Architecture (MCA) Bus, a PCI (Peripheral Component Interconnect) Bus, a PCI-Express (PCI-X) Bus, a Serial Advanced Technology Attachment (SATA) Bus, a Video Electronics Bus (audio Electronics Association), abbreviated VLB) bus or other suitable bus or a combination of two or more of these. Bus 60 may include one or more buses, where appropriate. Although specific buses are described and shown in the embodiments of the application, any suitable buses or interconnects are contemplated by the application.
The electronic device may perform an integral gradient model prediction method in an embodiment of the present application.
In addition, in combination with the method for predicting the integral gradient model in the above embodiments, the embodiments of the present application may be implemented by providing a computer readable storage medium. The computer readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement any one of the methods of integral gradient model prediction in the above embodiments.
And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for predicting an integrated gradient model, comprising:
an original model training step, defining a training set, constructing a neural network model, and training the neural network model on the training set;
a multitask training set construction step, namely initializing an empty multitask data set, selecting a sample from the training set which is not put back, and constructing the multitask data set according to the sample;
a multitask model construction step, namely adding an attribution structure in the trained neural network model to obtain a multitask model, and training the multitask model on the multitask data set;
and a multi-task model prediction step, namely predicting in an application scene by using the trained multi-task model.
2. The method of predicting an integrated gradient model of claim 1, wherein the step of constructing the multi-tasking training set further comprises: predicting a signature of the sample using the neural network model and generating a causal baseline sample based on the sample, calculating a causal, and adding the reduced cost, the signature, and the causal as new samples to the multitask dataset.
3. The integrated gradient model prediction method of claim 2, wherein the attribution is calculated using an integrated gradient method.
4. The integrated gradient model prediction method of claim 1, wherein the multitask model prediction step further comprises: when model prediction is carried out, a prediction result and an attribution result are simultaneously output.
5. An integrated gradient model prediction system, comprising:
the original model training module defines a training set, constructs a neural network model and trains the neural network model on the training set;
the multitask training set construction module is used for initializing an empty multitask data set, selecting a sample from the training set which is not put back, and constructing the multitask data set according to the sample;
the multitask model construction module is used for adding an attribution structure into the trained neural network model to obtain a multitask model and training the multitask model on the multitask data set;
and the multi-task model prediction module is used for predicting in an application scene by using the trained multi-task model.
6. The integrated gradient model prediction system of claim 5, wherein the multitask training set constructing module further comprises: predicting a signature of the sample using the neural network model and generating a causal baseline sample based on the sample, calculating a causal, and adding the reduced cost, the signature, and the causal as new samples to the multitask dataset.
7. The integrated gradient model prediction system of claim 6, wherein the attribution is calculated using an integrated gradient method.
8. The integrated gradient model prediction system of claim 5, wherein the multi-tasking model prediction module further comprises: when model prediction is carried out, a prediction result and an attribution result are simultaneously output.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the integrated gradient model prediction method of any one of claims 1 to 4 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the integrated gradient model prediction method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110937966.8A CN113743584A (en) | 2021-08-16 | 2021-08-16 | Integral gradient model prediction method, system, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110937966.8A CN113743584A (en) | 2021-08-16 | 2021-08-16 | Integral gradient model prediction method, system, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113743584A true CN113743584A (en) | 2021-12-03 |
Family
ID=78731236
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110937966.8A Pending CN113743584A (en) | 2021-08-16 | 2021-08-16 | Integral gradient model prediction method, system, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113743584A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332558A (en) * | 2021-12-15 | 2022-04-12 | 厦门市美亚柏科信息股份有限公司 | Training method and device for multitask neural network, computing equipment and storage medium |
| CN117910367A (en) * | 2024-03-20 | 2024-04-19 | 浙江大学 | Power system disturbed trajectory prediction method based on physical information neural network |
-
2021
- 2021-08-16 CN CN202110937966.8A patent/CN113743584A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| GUOYUN TU等: "A Multi-Task Neural Approach for Emotion Attribution, Classification and Summarization", 《IEEE TRANSACTIONS ON MULTIMEDIA》, vol. 22, no. 01, pages 1 - 5 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114332558A (en) * | 2021-12-15 | 2022-04-12 | 厦门市美亚柏科信息股份有限公司 | Training method and device for multitask neural network, computing equipment and storage medium |
| CN117910367A (en) * | 2024-03-20 | 2024-04-19 | 浙江大学 | Power system disturbed trajectory prediction method based on physical information neural network |
| CN117910367B (en) * | 2024-03-20 | 2024-05-28 | 浙江大学 | Power system disturbed trajectory prediction method based on physical information neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9866218B2 (en) | Boolean logic in a state machine lattice | |
| EP3227836B1 (en) | Active machine learning | |
| CN111985229B (en) | Sequence labeling method and device and computer equipment | |
| CN109344406B (en) | Part-of-speech tagging method and device and electronic equipment | |
| CN110673748A (en) | Method and device for providing candidate long sentences in input method | |
| KR102264233B1 (en) | An explanation-added document classification method by an artificial neural network that learns the correlation between sentences, document feature values, and sentence weights | |
| CN113743584A (en) | Integral gradient model prediction method, system, electronic device and storage medium | |
| CN112199473A (en) | Multi-turn dialogue method and device in knowledge question-answering system | |
| US20180018392A1 (en) | Topic identification based on functional summarization | |
| CN116227467A (en) | Model training method, text processing method and device | |
| CN114048288A (en) | Fine-grained emotion analysis method and system, computer equipment and storage medium | |
| CN115035890B (en) | Training method and device of voice recognition model, electronic equipment and storage medium | |
| CN110781292A (en) | Text data multi-level classification method and device, electronic equipment and storage medium | |
| US11604931B2 (en) | Electronic device and controlling method of electronic device | |
| CN113536773A (en) | Commodity comment sentiment analysis method and system, electronic equipment and storage medium | |
| CN114065768B (en) | Feature fusion model training and text processing method and device | |
| CN116226382B (en) | Text classification method and device for given keywords, electronic equipment and medium | |
| US12032614B2 (en) | Multiclass classification with diversified precision and recall weightings | |
| AU2021221148B2 (en) | Multiclass classification with diversified precision and recall weightings | |
| CN113919335A (en) | Pre-training word vector generation method and system, electronic device and storage medium | |
| US20240143927A1 (en) | Method for generating summary and system therefor | |
| US20240152760A1 (en) | Method, apparatus, device and medium for training and applying a contrastive learning model | |
| US20240111955A1 (en) | Named Entity Disambiguation Using Capsule Networks | |
| CN113987173A (en) | Short text classification method, system, electronic device and medium | |
| CN113919350A (en) | Entity identification method, system, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |