CN112070233B - Model joint training method, device, electronic equipment and storage medium - Google Patents
Model joint training method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112070233B CN112070233B CN202010866349.9A CN202010866349A CN112070233B CN 112070233 B CN112070233 B CN 112070233B CN 202010866349 A CN202010866349 A CN 202010866349A CN 112070233 B CN112070233 B CN 112070233B
- Authority
- CN
- China
- Prior art keywords
- model
- user
- similar
- target
- target task
- 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.)
- Active
Links
- 238000012549 training Methods 0.000 title claims abstract description 132
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000012512 characterization method Methods 0.000 claims abstract description 95
- 238000004821 distillation Methods 0.000 claims abstract description 64
- 238000013140 knowledge distillation Methods 0.000 claims abstract description 16
- 230000015654 memory Effects 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 12
- 230000003993 interaction Effects 0.000 claims description 10
- 238000013136 deep learning model Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 238000013473 artificial intelligence Methods 0.000 abstract description 3
- 238000013135 deep learning Methods 0.000 description 9
- 230000007547 defect Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application discloses a model joint training method, a device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence and cloud computing. The specific implementation scheme is as follows: carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model; and constructing a target task model according to the user target characteristics under the target task and the user target labels acquired from the label provider based on the similar characterization model, and performing target task prediction on the user. The method and the device can improve the prediction accuracy of the target task model.
Description
Technical Field
The application relates to the technical field of artificial intelligence and cloud computing, in particular to the technical field of deep learning, and specifically relates to a model joint training method, a device, electronic equipment and a storage medium.
Background
Machine learning is the core of artificial intelligence, and based on multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and the like, the existing knowledge structure is reorganized by a computer to continuously improve the performance of the machine.
Joint machine learning (Federated machine learning) is used to assist multiple institutions in data usage and machine learning modeling in meeting the requirements of user privacy protection, data security, and government regulations.
Disclosure of Invention
The present disclosure provides a method, apparatus, electronic device, and storage medium for model joint training.
According to an aspect of the present disclosure, there is provided a model joint training method, including:
carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model;
and constructing a target task model according to the user target characteristics under the target task and the user target labels acquired from the label provider based on the similar characterization model, and performing target task prediction on the user.
According to yet another aspect of the present disclosure, there is provided a model joint training method including:
determining a user target label under a target task;
and sending a model training request carrying the user target label to a model training party, wherein the model training request is used for indicating the model training party to execute the following steps: carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model; and constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model.
According to yet another aspect of the present disclosure, there is provided a model joint training apparatus including:
the characterization model module is used for obtaining a similar distillation model by carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks, and obtaining a similar characterization model according to the similar distillation model;
and the target task model module is used for constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider, and is used for predicting the target task of the user.
According to yet another aspect of the present disclosure, there is provided a model joint training apparatus including:
the target label determining module is used for determining a user target label under a target task;
the training request sending module is used for sending a model training request carrying the user target label to a model training party and indicating the model training party to execute the following steps: carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model; and constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model.
According to a fifth aspect, there is provided an electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the model joint training method as in any one of the embodiments of the present application.
According to a sixth aspect, there is provided a non-transitory computer readable storage medium storing computer instructions for causing a computer to perform a model joint training method as in any of the embodiments of the present application.
According to the technology, the prediction accuracy of the target task model can be improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.
Drawings
The drawings are for better understanding of the present solution and do not constitute a limitation of the present application. Wherein:
FIG. 1 is a flow chart of a model joint training method provided according to an embodiment of the present application;
FIG. 2 is a flow chart of another model joint training method provided in accordance with an embodiment of the present application;
FIG. 3 is a flow chart of yet another model joint training method provided in accordance with an embodiment of the present application;
FIG. 4 is a flow chart of yet another model joint training method provided in accordance with an embodiment of the present application;
FIG. 5 is a flow chart of yet another model joint training method provided in accordance with an embodiment of the present application;
FIG. 6 is a schematic diagram of a model joint training apparatus according to an embodiment of the present application;
FIG. 7 is a schematic structural diagram of yet another model joint training apparatus provided in accordance with an embodiment of the present application;
FIG. 8 is a block diagram of an electronic device for implementing a model joint training method of an embodiment of the present application.
Detailed Description
Exemplary embodiments of the present application are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present application to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 is a schematic flow chart of a model joint training method according to an embodiment of the present application. The embodiment can be suitable for the condition that the label provider and the model trainer jointly train the target task model. The model joint training method disclosed in the embodiment may be executed by an electronic device, and in particular, may be executed by a model joint training apparatus, where the apparatus may be implemented by software and/or hardware, and configured in an electronic device of a model training party. Referring to fig. 1, the model joint training method provided in this embodiment includes:
s110, carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model.
The scenes to which the similar task and the target task belong may be similar, and the participants may be different, that is, the predicted targets between the similar task and the target task are similar. Participants to the target task may be tag providers and model trainers, while participants to the similar task may be model trainers.
The label provider is used for providing user target labels under target tasks, and the model training party is used for providing user similar labels under similar tasks and user characteristics under the target tasks and the similar tasks. Additionally, tag extraction may also be used to provide part of the user features. Taking different wind control scenes as an example, if the first party has a small amount of wind control branch label values of the A wind control scene, the second party provides wind control branch prediction service of the B wind control scene, and the prediction service is determined based on a large amount of wind control branch label values of the second party, the first party can be used as a label provider, the wind control branch label values of the first party are used as user target labels, the second party is used as a model training party, the wind control branch label values of the second party are used as user similar labels, and user characteristics associated with the user similar labels are used as user similar characteristics.
Specifically, the user similar characteristics are used as input of a distillation model, and the user similar labels are used as distillation targets to carry out distillation model training, so that a similar distillation model is obtained. And, the model before the middle layer of the similar distillation model can be used as a similar characterization model, namely, the middle layer of the similar distillation model is output as a similar characterization feature. For example, the output layer of the similar distillation model may be removed to obtain a similar characterization model. Moreover, the dimension of the similar characteristic feature can be adjusted according to the sparse condition of the characteristic feature of the similar characteristic feature of the user, and if sparse, the dimension of the similar characteristic feature can be reduced; if dense, the dimensions of similar characterization features may be increased. For example, the dimension of similar characterization features may be set to 50. And distilling the similar task model under the similar task, so that the similar distillation model and the similar characterization model can keep information of the user similar labels.
And S120, constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider, and performing target task prediction on the user.
In the joint modeling process, the modeling effect is greatly affected by the training sample size, the modeling is insufficient due to the small training sample size, and the modeling effect is limited due to the limited sample tag data size generally provided by a tag provider. In the method, the user similar label information in the similar scene is reserved in the similar characterization model, the target task model is built based on the similar characterization model, namely, the target task model is built by indirectly using the user similar label, so that the target task model not only covers the user target label of the label provider, but also comprises the user similar label used in the similar task in the similar scene, the defect of insufficient number of the user target labels can be overcome, and the prediction accuracy of the target task model can be improved.
The user target characteristics under the target tasks can be determined by the model training party or can be determined by the model training party and the label provider together, that is, the model training party and the label provider can both provide partial user target characteristics.
The target task model is a combined training result. After the target task model is built, the characteristics of the user to be predicted are input into the target task model, and the prediction result of the user to be predicted under the target task can be obtained.
According to the technical scheme, the similar scene and the prediction model of the similar task are distilled, the similar characterization model is extracted from the similar distillation model, and joint modeling is performed based on the similar characterization model, namely, the target task model is built by indirectly using the user similar label, so that the defect of insufficient target labels of users can be overcome, and the prediction accuracy of the target task model can be improved.
Fig. 2 is a flow chart of a model joint training method according to an embodiment of the present application. This embodiment is an alternative to the embodiments described above. Referring to fig. 2, the model joint training method provided in this embodiment includes:
S210, training the deep learning model according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model.
In the related technology of joint learning, under the condition of ensuring data security through security policies such as homomorphic encryption, the data of each participant cannot be out of the local area, and joint training is carried out. Since homomorphic encryption only supports addition and multiplication operations, the related art can only use a joint modeling model such as XGBoost tree model including only addition and multiplication operations, and model training using homomorphic encryption algorithm is inefficient.
In the application, the user similar features and the user similar labels are provided by a model training party, so that the model training party does not need to adopt a homomorphic encryption algorithm to construct a similar distillation model, and therefore the model is not limited by operators and training efficiency supported by the homomorphic encryption algorithm, the similar distillation model can adopt a deep learning network structure, for example, an MLP (Multilayer Perceptron, multi-layer perceptron) _deep model can be adopted. Compared with a tree model, the method has better generalization and characteristic characterization capability by constructing the similar distillation model based on the deep learning network structure, so that the similar characterization model obtained according to the similar distillation model has better expression on similar scenes.
S220, constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider, and performing target task prediction on the user.
In an alternative embodiment, S220 includes: the similar representation model is connected with a prediction output layer of a target task to obtain a joint prediction model of joint modeling; and training the joint prediction model according to the user target characteristics under the target task and the user target labels acquired from the label provider to obtain the target task model.
Specifically, a joint prediction model is built according to the similar identification model, user target characteristics are used as input of the joint prediction model, user target labels are used as output of the joint prediction model, the joint prediction model is trained, and a training result is used as a target task model. And in the transfer learning, the simulated characterization model is used as a pre-training model, and the supervised transfer learning is performed based on the user target labels, so that the over-fitting problem caused by the limited user target labels in the supervised learning can be relieved.
In an alternative embodiment, S220 includes: taking the user target characteristics under the target task as the input of the similar characterization model to obtain the similar characterization characteristics of the user; fusing the user target characteristics under the target task and the similar characterization characteristics to obtain user joint characteristics; and constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider.
Specifically, the user joint characteristics can be obtained by splicing the user target characteristics and the similar characterization characteristics, the user joint characteristics are used as input, and the user target labels are used as output to perform model training. The model may be a tree model. The similar characteristic features are spliced to the user target features as newly added features, a more flexible application mode is provided, the joint modeling model is more selected, and the similar identification features also provide additional characteristic information to assist modeling.
According to the technical scheme, the deep learning distillation model is adopted, so that the method has better generalization and characteristic characterization capability; moreover, by providing two joint modeling modes, the problem of overfitting caused by limited label data in supervised learning can be relieved or the flexibility of model selection is provided.
Fig. 3 is a flow chart of a model joint training method according to an embodiment of the present application. This embodiment is an alternative to the embodiments described above. Referring to fig. 3, the model joint training method provided in this embodiment includes:
s310, sampling training samples of similar tasks according to user tag value distribution of the similar tasks.
Specifically, training samples of similar tasks can be sampled in proportion according to the distribution of different user label intervals. Taking the wind control scene as an example, the wind control scene is distributed according to different fractional segments of the wind control division, and the wind control scene is sampled proportionally. Sample extraction is carried out according to the user tag value distribution in proportion, so that the sample balance of the subsequent similar distillation model can be improved, and the accuracy of the similar distillation model is improved.
S320, determining user similar characteristics and user similar labels under similar tasks according to the extracted training samples.
Specifically, the features in the sampling training sample are used as user similar features, and the labels in the sampling training sample are used as user similar labels.
S330, determining the user identification text according to the user identification ciphertext obtained from the tag provider.
Specifically, according to the user identification ciphertext, mapping to obtain the user identification text of the model training party. Taking the user identification ciphertext as the user mobile phone number ciphertext as an example, the model training party obtains the user mobile phone number text through mapping.
S340, determining the user target characteristics according to the user identification original text.
Specifically, according to the user identification text, the user target characteristics of the model training party are obtained. The tag provider transmits the user identification ciphertext instead of the user identification text to the model training party, so that the user identification is prevented from being leaked in the transmission process, and the data security can be further improved.
S350, carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model.
S360, constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider, and performing target task prediction on the user.
In an alternative embodiment, a homomorphic encryption algorithm is adopted between the model trainer and the tag provider for data interaction. Specifically, homomorphic encryption algorithm is adopted to conduct user identification data interaction or partial user target feature interaction. After the similar characterization model is constructed, the homomorphic encryption algorithm is introduced in the construction process of the target task model, so that the safety of the interactive data can be further improved.
According to the technical scheme, the target task model is built indirectly through the user similarity labels, the prediction accuracy of the target task model can be improved, and the prediction accuracy of the target task model and the data safety of each participant are further improved through the determination of the user similarity characteristics and the user target characteristics.
Fig. 4 is a flow chart of a model joint training method according to an embodiment of the present application. The embodiment can be suitable for the condition that the label provider and the model trainer jointly train the target task model. The model joint training method disclosed in the embodiment may be executed by an electronic device, and in particular, may be executed by a model joint training apparatus, where the apparatus may be implemented by software and/or hardware, and configured in an electronic device of a tag provider. Referring to fig. 4, the model joint training method provided in this embodiment includes:
s410, determining a user target label under the target task.
S420, sending a model training request carrying the user target label to a model training party, wherein the model training request is used for indicating the model training party to execute the following steps: carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model; and constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model.
The scenes to which the similar task and the target task belong may be similar, and the participants may be different, that is, the predicted targets between the similar task and the target task are similar. User similar tags and user target tags refer to user tags under similar tasks and target tasks, respectively. It should be noted that, samples used in the process of constructing the similar distillation model and the target task model may be the same user, or may belong to different users.
Because the similar characterization model keeps the user similar label information in the similar scene, the target task model is built based on the similar characterization model, namely, the target task model is built indirectly by using the user similar label, the defect of insufficient target labels of users can be overcome, and the prediction accuracy of the target task model can be improved.
In an alternative embodiment, the similar distillation model is a deep learning model. Compared with a tree model, the method has better generalization and characteristic characterization capability by constructing the similar distillation model based on the deep learning network structure, so that the similar characterization model obtained according to the similar distillation model has better expression on similar scenes.
According to the technical scheme, the target task model is built by indirectly using the user similar labels, so that the defect of insufficient target labels of users can be overcome, and the prediction accuracy of the target task model can be improved. Moreover, by adopting the deep learning distillation model, the method has better generalization and characteristic characterization capability, and further improves the prediction accuracy of the target task model.
Fig. 5 is a flow chart of a model joint training method according to an embodiment of the present application. The embodiment is a specific implementation scheme provided on the basis of the embodiment. Referring to fig. 5, the model joint training method provided in this embodiment includes:
s510, the label provider determines a user target label under the target task.
S520, the label providing direction model training party sends the user target label.
And S530, the model training party carries out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model.
S540, the model training party obtains a similar characterization model according to the similar distillation model.
S550, the model training party builds a target task model based on the similar representation model according to the user target characteristics and the user target labels under the target task.
In an alternative embodiment, S550 may include: obtaining user id texts of the model training party according to the user id ciphertext mapping obtained from the label provider, and obtaining user target characteristics of the model training party according to the user id texts; constructing a joint prediction model based on the similarity characterization model; and training the combined prediction model by adopting the user target characteristics and the user target labels to obtain a target task model.
In an alternative embodiment, S550 may include: obtaining user id texts of the model training party according to the user id ciphertext mapping obtained from the label provider, and obtaining user target characteristics of the model training party according to the user id texts; taking the user target characteristics as the input of a similar characterization model to obtain similar characterization characteristics of the user; splicing the user target characteristics and the similar characterization characteristics of the user to obtain user joint characteristics; and constructing a target task model according to the user joint characteristics and the user target labels.
According to the technical scheme, the method and the device support secure fusion and modeling calculation of data of any two parties, and provide full-flow service from feature analysis processing, model training and effect evaluation to model application deployment. The method comprises the steps of obtaining the characterization information of similar tasks under similar scenes through representation learning, and applying the characterization information of the similar tasks to a joint modeling scene with insufficient label information, so that the defect that the model is trained to fit or is inadequately trained due to the insufficient label is overcome to a certain extent; moreover, the obtained characterization information is dense, and the feature expression capability is stronger; the characterization model can adopt a deep learning model, and compared with the traditional tree-type combined modeling model, the generalization and feature expression capability are stronger; the obtained characterization features can participate in joint modeling in various modes, so that a target task model not only covers user target labels of a label provider, but also comprises user similar labels used in similar tasks in similar scenes; the expansibility is strong, and the method can be expanded to different service scenes.
Fig. 6 is a schematic structural diagram of a model joint training apparatus according to an embodiment of the present application, where the apparatus may be configured in an electronic device of a model training party. Referring to fig. 6, a model joint training apparatus 600 provided in an embodiment of the present application may include:
the characterization model module 601 is configured to perform knowledge distillation according to user similar features and user similar labels under similar tasks to obtain a similar distillation model, and obtain a similar characterization model according to the similar distillation model;
the target task model module 602 is configured to construct a target task model for performing target task prediction on the user according to the user target characteristics under the target task and the user target labels acquired from the label provider, based on the similar characterization model.
Optionally, the characterization model module 601 is specifically configured to:
training the deep learning model according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model.
Optionally, the apparatus 600 further includes a similar data module, where the similar data module includes:
the sample extraction unit is used for sampling training samples of similar tasks according to the user tag value distribution of the similar tasks;
And the similar data unit is used for determining the user similar characteristics and the user similar labels according to the extracted training samples.
Optionally, the apparatus 600 further includes a target feature module, where the target feature module includes:
the identification original text unit is used for determining the user identification original text according to the user identification ciphertext acquired from the tag provider;
and the target feature unit is used for determining the target feature of the user according to the user identification original text.
Optionally, the target task model module 602 includes:
the joint model unit is used for connecting the similar representation model with a prediction output layer of a target task to obtain a joint prediction model of joint modeling;
and the target task model unit is used for training the joint prediction model according to the user target characteristics under the target task and the user target labels acquired from the label provider to obtain the target task model.
Optionally, the target task model module 602 includes:
the characteristic feature unit is used for taking the user target feature under the target task as the input of the similar characteristic model to obtain the similar characteristic feature of the user;
the combined feature unit is used for fusing the user target feature under the target task and the similar characterization feature to obtain a user combined feature;
And the target task model unit is used for constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider.
Optionally, a homomorphic encryption algorithm is adopted between the model training party and the label provider for data interaction.
According to the technical scheme, the target task model is built by indirectly using the user similar labels through the model training party, so that the defect of insufficient user target labels provided by the label provider can be overcome, and the prediction accuracy of the target task model can be improved. By adopting the deep learning distillation model, the method has better generalization and characteristic characterization capability, and further improves the prediction accuracy of the target task model. In addition, by providing two joint learning modes based on the similar characterization model, the prediction accuracy can be further improved.
Fig. 7 is a schematic structural diagram of a model joint training device according to an embodiment of the present application, where the device may be configured in an electronic device of a tag provider. Referring to fig. 7, a model joint training apparatus 700 provided in an embodiment of the present application may include:
a target tag determination module 701, configured to determine a target tag of a user under a target task;
The training request sending module 702 is configured to send a model training request carrying the user target tag to a model trainer, and is configured to instruct the model trainer to perform the following steps: carrying out knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and obtaining a similar characterization model according to the similar distillation model; and constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model.
According to the technical scheme, the target task model is built by indirectly using the user similar labels through the model training party, so that the defect of insufficient user target labels provided by the label provider can be overcome, and the prediction accuracy of the target task model can be improved. By adopting the deep learning distillation model, the method has better generalization and characteristic characterization capability, and further improves the prediction accuracy of the target task model. In addition, by providing two joint learning modes based on the similar characterization model, the prediction accuracy can be further improved.
According to embodiments of the present application, an electronic device and a readable storage medium are also provided.
As shown in fig. 8, is a block diagram of an electronic device of a method of model joint training according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the application described and/or claimed herein.
As shown in fig. 8, the electronic device includes: one or more processors 801, memory 802, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions executing within the electronic device, including instructions stored in or on memory to display graphical information of the GUI on an external input/output device, such as a display device coupled to the interface. In other embodiments, multiple processors and/or multiple buses may be used, if desired, along with multiple memories and multiple memories. Also, multiple electronic devices may be connected, each providing a portion of the necessary operations (e.g., as a server array, a set of blade servers, or a multiprocessor system). One processor 801 is illustrated in fig. 8.
Memory 802 is a non-transitory computer-readable storage medium provided herein. Wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of model joint training provided herein. The non-transitory computer readable storage medium of the present application stores computer instructions for causing a computer to perform the method of model joint training provided herein.
The memory 802 is used as a non-transitory computer readable storage medium, and may be used to store non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules corresponding to the method of model joint training in the embodiments of the present application (e.g., the characterization model module 601 and the target task model module 602 shown in fig. 6, and the target tag determination module 701 and the training request sending module 702 shown in fig. 7). The processor 801 executes various functional applications of the server and model co-training, i.e., a method of implementing model co-training in the above-described method embodiments, by running non-transitory software programs, instructions, and modules stored in the memory 802.
Memory 802 may include a storage program area that may store an operating system, at least one application program required for functionality, and a storage data area; the storage data area may store data created from the use of the model-trained electronic device, and the like. In addition, memory 802 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 802 may optionally include memory remotely located with respect to processor 801, which may be connected to the model joint training electronics via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The electronic device of the method of model joint training may further include: an input device 803 and an output device 804. The processor 801, memory 802, input devices 803, and output devices 804 may be connected by a bus or other means, for example in fig. 8.
The input device 803 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the model-trained electronic device, such as a touch screen, keypad, mouse, trackpad, touchpad, pointer stick, one or more mouse buttons, trackball, joystick, and the like. The output device 804 may include a display apparatus, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device may be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASIC (application specific integrated circuit), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
These computing programs (also referred to as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
According to the technical scheme, the target task model is built by indirectly using the user similar labels through the model training party, so that the defect of insufficient user target labels provided by the label provider can be overcome, and the prediction accuracy of the target task model can be improved. By adopting the deep learning distillation model, the method has better generalization and characteristic characterization capability, and further improves the prediction accuracy of the target task model. In addition, by providing two joint learning modes based on the similar characterization model, the prediction accuracy can be further improved.
The above embodiments do not limit the scope of the application. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application are intended to be included within the scope of the present application.
Claims (14)
1. A model joint training method, comprising:
performing knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and outputting a model in front of a middle layer of the similar distillation model as a similar characterization model by using the model in front of the middle layer of the similar distillation model as a similar characterization characteristic;
based on the similar characterization model, constructing a target task model according to user target characteristics under a target task and user target labels acquired from a label provider, and performing target task prediction on a user; the model training party and the label provider carry out data interaction by adopting homomorphic encryption algorithm;
the constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider comprises the following steps:
taking the user target characteristics under the target task as the input of the similar characterization model to obtain the similar characterization characteristics of the user;
fusing the user target characteristics under the target task and the similar characterization characteristics to obtain user joint characteristics;
and constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider.
2. The method of claim 1, wherein the performing knowledge distillation based on the user-similar features and user-similar tags under similar tasks to obtain a similar distillation model comprises:
training the deep learning model according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model.
3. The method of claim 1, further comprising:
sampling training samples of the similar tasks according to the user tag value distribution of the similar tasks;
and determining the user similar characteristics and the user similar labels according to the extracted training samples.
4. The method of claim 1, further comprising:
determining a user identification text according to the user identification ciphertext obtained from the tag provider;
and determining the user target characteristics according to the user identification original text.
5. The method of claim 1, wherein the constructing a target task model based on the similar characterization model from user target features under a target task and user target tags obtained from a tag provider further comprises:
the similar representation model is connected with a prediction output layer of a target task to obtain a joint prediction model of joint modeling;
And training the joint prediction model according to the user target characteristics under the target task and the user target labels acquired from the label provider to obtain the target task model.
6. A model joint training method, comprising:
determining a user target label under a target task;
and sending a model training request carrying the user target label to a model training party, wherein the model training request is used for indicating the model training party to execute the following steps: performing knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and outputting a model in front of a middle layer of the similar distillation model as a similar characterization model by using the model in front of the middle layer of the similar distillation model as a similar characterization characteristic; constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model; based on the similar characterization model, constructing a target task model according to the user target characteristics and the user target labels under the target task comprises the following steps: taking the user target characteristics under the target task as the input of the similar characterization model to obtain the similar characterization characteristics of the user; fusing the user target characteristics under the target task and the similar characterization characteristics to obtain user joint characteristics; constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider;
And the model training party and the label provider adopt homomorphic encryption algorithm to conduct data interaction.
7. A model joint training apparatus comprising:
the characterization model module is used for obtaining a similar distillation model according to the user similar characteristics and the user similar labels under the similar tasks through knowledge distillation, and a model before the middle layer of the similar distillation model is used as a similar characterization model, and the middle layer of the similar distillation model is output as the similar characterization characteristics;
the target task model module is used for constructing a target task model based on the similar characterization model according to the user target characteristics under the target task and the user target labels acquired from the label provider, and is used for predicting the target task of the user; the model training party and the label provider carry out data interaction by adopting homomorphic encryption algorithm;
wherein the target task model module comprises:
the characteristic feature unit is used for taking the user target feature under the target task as the input of the similar characteristic model to obtain the similar characteristic feature of the user;
the combined feature unit is used for fusing the user target feature under the target task and the similar characterization feature to obtain a user combined feature;
And the target task model unit is used for constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider.
8. The apparatus of claim 7, wherein the characterization model module is specifically configured to:
training the deep learning model according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model.
9. The apparatus of claim 7, further comprising a similarity data module, the similarity data module comprising:
the sample extraction unit is used for sampling training samples of similar tasks according to the user tag value distribution of the similar tasks;
and the similar data unit is used for determining the user similar characteristics and the user similar labels according to the extracted training samples.
10. The apparatus of claim 7, further comprising a target feature module, the target feature module comprising:
the identification original text unit is used for determining the user identification original text according to the user identification ciphertext acquired from the tag provider;
and the target feature unit is used for determining the target feature of the user according to the user identification original text.
11. The apparatus of claim 7, wherein the target task model module further comprises:
The joint model unit is used for connecting the similar representation model with a prediction output layer of a target task to obtain a joint prediction model of joint modeling;
and the target task model unit is used for training the joint prediction model according to the user target characteristics under the target task and the user target labels acquired from the label provider to obtain the target task model.
12. A model joint training apparatus comprising:
the target label determining module is used for determining a user target label under a target task;
the training request sending module is used for sending a model training request carrying the user target label to a model training party and indicating the model training party to execute the following steps: performing knowledge distillation according to the user similar characteristics and the user similar labels under similar tasks to obtain a similar distillation model, and outputting a model in front of a middle layer of the similar distillation model as a similar characterization model by using the model in front of the middle layer of the similar distillation model as a similar characterization characteristic; constructing a target task model according to the user target characteristics under the target task and the user target labels based on the similar characterization model; based on the similar characterization model, constructing a target task model according to the user target characteristics and the user target labels under the target task comprises the following steps: taking the user target characteristics under the target task as the input of the similar characterization model to obtain the similar characterization characteristics of the user; fusing the user target characteristics under the target task and the similar characterization characteristics to obtain user joint characteristics; constructing a target task model according to the user joint characteristics and the user target labels acquired from the label provider;
And the model training party and the label provider adopt homomorphic encryption algorithm to conduct data interaction.
13. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.
14. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010866349.9A CN112070233B (en) | 2020-08-25 | 2020-08-25 | Model joint training method, device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010866349.9A CN112070233B (en) | 2020-08-25 | 2020-08-25 | Model joint training method, device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112070233A CN112070233A (en) | 2020-12-11 |
| CN112070233B true CN112070233B (en) | 2024-03-22 |
Family
ID=73659797
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010866349.9A Active CN112070233B (en) | 2020-08-25 | 2020-08-25 | Model joint training method, device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112070233B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334934A (en) * | 2017-06-07 | 2018-07-27 | 北京深鉴智能科技有限公司 | Convolutional neural networks compression method based on beta pruning and distillation |
| CN108898168A (en) * | 2018-06-19 | 2018-11-27 | 清华大学 | The compression method and system of convolutional neural networks model for target detection |
| CN109637546A (en) * | 2018-12-29 | 2019-04-16 | 苏州思必驰信息科技有限公司 | Knowledge distillating method and device |
| CN110223281A (en) * | 2019-06-06 | 2019-09-10 | 东北大学 | A kind of Lung neoplasm image classification method when in data set containing uncertain data |
| CN110246487A (en) * | 2019-06-13 | 2019-09-17 | 苏州思必驰信息科技有限公司 | Optimization method and system for single pass speech recognition modeling |
| CN110428052A (en) * | 2019-08-01 | 2019-11-08 | 江苏满运软件科技有限公司 | Construction method, device, medium and the electronic equipment of deep neural network model |
| CN111198940A (en) * | 2019-12-27 | 2020-05-26 | 北京百度网讯科技有限公司 | FAQ method, question-answer search system, electronic device, and storage medium |
| CN111460150A (en) * | 2020-03-27 | 2020-07-28 | 北京松果电子有限公司 | Training method, classification method and device of classification model and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11562287B2 (en) * | 2017-10-27 | 2023-01-24 | Salesforce.Com, Inc. | Hierarchical and interpretable skill acquisition in multi-task reinforcement learning |
-
2020
- 2020-08-25 CN CN202010866349.9A patent/CN112070233B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108334934A (en) * | 2017-06-07 | 2018-07-27 | 北京深鉴智能科技有限公司 | Convolutional neural networks compression method based on beta pruning and distillation |
| CN108898168A (en) * | 2018-06-19 | 2018-11-27 | 清华大学 | The compression method and system of convolutional neural networks model for target detection |
| CN109637546A (en) * | 2018-12-29 | 2019-04-16 | 苏州思必驰信息科技有限公司 | Knowledge distillating method and device |
| CN110223281A (en) * | 2019-06-06 | 2019-09-10 | 东北大学 | A kind of Lung neoplasm image classification method when in data set containing uncertain data |
| CN110246487A (en) * | 2019-06-13 | 2019-09-17 | 苏州思必驰信息科技有限公司 | Optimization method and system for single pass speech recognition modeling |
| CN110428052A (en) * | 2019-08-01 | 2019-11-08 | 江苏满运软件科技有限公司 | Construction method, device, medium and the electronic equipment of deep neural network model |
| CN111198940A (en) * | 2019-12-27 | 2020-05-26 | 北京百度网讯科技有限公司 | FAQ method, question-answer search system, electronic device, and storage medium |
| CN111460150A (en) * | 2020-03-27 | 2020-07-28 | 北京松果电子有限公司 | Training method, classification method and device of classification model and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| Cross-Modal Knowledge Distillation for Action Recognition;Fida Mohammad Thoker 等;2019 IEEE International Conference on Image Processing (ICIP);20190826;全文 * |
| 一种多标签统一域嵌入的推荐模型;张随雨;杨成;;哈尔滨工业大学学报;20200510(第05期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112070233A (en) | 2020-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102484617B1 (en) | Method and apparatus for generating model for representing heterogeneous graph node, electronic device, storage medium and program | |
| EP3882784A1 (en) | Event argument extraction method and apparatus and electronic device | |
| KR102528748B1 (en) | Method, apparatus, device and storage medium for constructing knowledge graph | |
| CN111259671B (en) | Semantic description processing method, device and equipment for text entity | |
| JP7076030B2 (en) | User interaction information processing model generation method and device, user interaction information processing method and device, equipment, medium, and program | |
| KR20210118360A (en) | Method, apparatus, electronic device, program and readable storage medium for creating a label marking model | |
| CN112102448B (en) | Virtual object image display method, device, electronic equipment and storage medium | |
| CN112650907A (en) | Search word recommendation method, target model training method, device and equipment | |
| CN111680517B (en) | Method, apparatus, device and storage medium for training model | |
| CN111078878B (en) | Text processing method, device, equipment and computer readable storage medium | |
| CN111241838B (en) | Semantic relation processing method, device and equipment for text entity | |
| EP3901905B1 (en) | Method and apparatus for processing image | |
| CN112529180B (en) | Method and apparatus for model distillation | |
| CN110795593A (en) | Voice packet recommendation method and device, electronic equipment and storage medium | |
| CN112380104A (en) | User attribute identification method and device, electronic equipment and storage medium | |
| CN111563198B (en) | Material recall method, device, equipment and storage medium | |
| CN111666771A (en) | Semantic label extraction device, electronic equipment and readable storage medium of document | |
| CN112699314A (en) | Hot event determination method and device, electronic equipment and storage medium | |
| CN112446728B (en) | Advertisement recall method, device, equipment and storage medium | |
| CN111125445B (en) | Community theme generation method and device, electronic equipment and storage medium | |
| CN111160552B (en) | News information recommendation processing method, device, equipment and computer storage medium | |
| CN112070233B (en) | Model joint training method, device, electronic equipment and storage medium | |
| CN112508830B (en) | Training method, device, equipment and storage medium of image processing model | |
| CN113591908B (en) | Event index matching method, model training method, device and electronic equipment | |
| CN111767988B (en) | Fusion method and device of neural network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |