CN110895508B - Method and device for generating traversal test path - Google Patents
Method and device for generating traversal test path Download PDFInfo
- Publication number
- CN110895508B CN110895508B CN201910984049.8A CN201910984049A CN110895508B CN 110895508 B CN110895508 B CN 110895508B CN 201910984049 A CN201910984049 A CN 201910984049A CN 110895508 B CN110895508 B CN 110895508B
- Authority
- CN
- China
- Prior art keywords
- test
- control
- traversal
- neural network
- network model
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
- Stored Programmes (AREA)
Abstract
The invention discloses a method and a device for generating a traversal test path. The method comprises the steps of collecting test sample data, training a neural network model according to the test sample data, transplanting the trained neural network model to test equipment, predicting and selecting a traversal path with the maximum number of trigger bugs according to the neural network model to serve as a traversal test path, and performing traversal test on a control set to be tested on the test equipment according to the traversal test path. The problem that in the prior art, the test efficiency is low and the test maintenance cost is high due to the fact that the test paths and parameters need to be adapted manually is solved, the traversal test path with the largest number of control Bug can be predicted intelligently by utilizing the neural network, and traversal test quality and test efficiency are improved. The method can be widely applied to the field of software traversal test.
Description
Technical Field
The invention relates to the field of software testing, in particular to a method and a device for generating a traversal testing path.
Background
In the field of automated testing, traversal testing is widely applied to various types of tests and plays an important role. The traversal test should discover the Bug of the application or control to be tested as much as possible, and exert the function to the maximum extent. In the traversal testing process, the traversal testing path has a large influence on the testing result, if the coverage of the testing path is high and the selection is reasonable, more bugs of the control to be tested can be triggered, and the bugs are directly used for optimizing the control, so that the selection of the reasonable traversal testing path is very important in the traversal testing process to trigger the bugs as much as possible. However, when many controls are tested, the testing paths and parameters need to be adapted manually, so that the testing efficiency is low and the testing and maintenance cost is very high. Therefore, it is necessary to generate a traversal test path capable of triggering the maximum number of bugs according to the attributes of the control to be tested.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art. Therefore, the invention aims to provide a method for generating a traversal test path capable of triggering the maximum number of bugs according to the neural network and the attributes of the control to be tested.
The technical scheme adopted by the embodiment of the invention is as follows:
in a first aspect, the present invention provides a method for generating a traversal test path, including:
collecting test sample data to form a test sample set and training a neural network model according to the test sample set;
transplanting the trained neural network model to test equipment;
and setting a control set to be tested, predicting and selecting a traversal path with the maximum number of trigger bugs as a traversal test path according to the neural network model, and performing traversal test on the control set to be tested on the test equipment according to the traversal test path.
Further, the test sample data comprises: the control attribute information comprises at least one of the following information: the control test package name, the entry function name, the control type, the control number and the control information.
Further, preprocessing the test sample data, wherein the preprocessing comprises feature extraction and data cleaning;
the feature extraction specifically comprises the following steps: clustering the non-digital control attribute information, and converting the clustered control attribute information into corresponding coding vectors according to a preset coding rule;
the data cleaning specifically comprises the following steps: and storing the coding vector and the control Bug number according to a preset format.
Further, the neural network model is trained by using a Tensorflow framework, and the training process specifically comprises the following steps:
setting parameters of the neural network model to build the neural network model;
inputting test sample data as a training sample set of the neural network model;
and training the neural network model through the training sample set and updating the network parameters. Further, the loss function of the neural network model is a cross entropy loss function.
Further, training the neural network model by adopting a gradient descent method.
Further, the traversing the test path refers to: and setting a click time sequence, predicting and selecting the control to be tested with the maximum number of bugs in the control set to be tested at each click time as a selection control at the current click time, wherein the traversal sequence of the selection controls forms the traversal test path.
In a second aspect, the present invention further provides an apparatus for generating a traversal test path, including:
a neural network model generation module: the neural network model is used for collecting test sample data and training the neural network model according to the test sample data;
a neural network model transplanting module: the neural network model is transplanted to a test device;
and a traversal path selecting module: the test device is used for setting a set of controls to be tested, performing traversal test on the controls to be tested on the test device, and predicting and selecting a traversal path with the maximum number of trigger bugs as a traversal test path according to the neural network model.
In a third aspect, the present invention further provides an apparatus for generating a traversal test path, including:
at least one processor; and a memory communicatively coupled to the at least one processor;
wherein the processor is operable to perform the method of any one of claims 1 to 7 by invoking a computer program stored in the memory.
In a fourth aspect, the present invention provides a computer-readable storage medium having stored thereon computer-executable instructions for causing a computer to perform the method of any of the first aspects.
The embodiment of the invention has the beneficial effects that:
according to the embodiment of the invention, test sample data is collected, the neural network model is trained according to the test sample data, the trained neural network model is transplanted to test equipment, then a control set to be tested is set, a traversal path with the maximum number of trigger bugs is predicted and selected according to the neural network model to serve as a traversal test path, and traversal test of the control set to be tested is carried out on the test equipment according to the traversal test path. The problem that in the prior art, the test efficiency is low and the test maintenance cost is high due to the fact that the test paths and parameters need to be adapted manually is solved, the traversal test path with the largest number of control Bug can be predicted intelligently by utilizing the neural network, and traversal test quality and test efficiency are improved. The method can be widely applied to the field of software traversal test.
Drawings
FIG. 1 is a flowchart illustrating an implementation of one embodiment of a method for generating a traversal test path in the present invention;
FIG. 2 is a block diagram of an embodiment of an apparatus for generating a traversal test path in the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following description will be made with reference to the accompanying drawings. It is obvious that the drawings in the following description are only some examples of the invention, and that for a person skilled in the art, other drawings and embodiments can be derived from them without inventive effort.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
The first embodiment is as follows:
the embodiment of the invention provides a method for generating a traversal test path, which applies a deep learning algorithm (such as a neural network model) to a traversal test, so that a control with the largest number of Bug can be selected for traversal, and the trained neural network model can intelligently predict and select the traversal test path with the highest Bug number.
Fig. 1 is a flowchart of an implementation of a method for generating a traversal test path according to this embodiment, as shown in fig. 1, the method includes the following steps:
s1: and generating a neural network model, specifically collecting test sample data to form a training sample set, and training the neural network model according to the training sample set.
S2: transplanting the neural network model to the test equipment means that the trained neural network model is transplanted to the test equipment to perform control traversal test, the test equipment can be selected to have android equipment, IOS equipment and the like, the trained neural network model is transplanted according to actual test requirements, and in the embodiment, an android terminal is selected as the test equipment to be explained.
S3: traversing the control set to be tested and selecting a traversal path, firstly setting the control set to be tested, carrying out traversal test on the control to be tested on test equipment, predicting the number of Bug of each control according to the transplanted neural network model, and selecting the traversal path with the maximum number of trigger Bug as the traversal test path.
Wherein the source of the test sample data in the S1 is optional, the control crawled from the Bug platform tests Bug data, and the test sample data comprises: the control attribute information comprises the following information: the control test package name, the name of the entry function (such as an Activity class name), the type of the control (such as a Button, a control type with different functions such as ImageView picture browsing and the like), the number of the control (used for distinguishing different controls), the information of the control (such as information of the boundary area of the control, whether the control is in the center of the interface, whether the control can slide and the like) and the like. The following description takes a set of specific examples as an example, where the test sample data includes control attribute information and a Bug number:
control test package name: com.example.test;
name of the entry function: MainActivity;
and (4) control type: button;
control information: 200 x 200 (space boundary area), slidable;
control numbering: x;
the number of Bug: 10, the number of the channels is 10;
in this embodiment, the attribute information of the controls is grouped into a set X, and the total number of bugs is found to be Y before clicking the controls, so that each control can obtain a group of data and record as (X, Y), which respectively corresponds to < control attribute information, Bug number >, and all test sample data are grouped together to obtain a vector (X, Y).
In one embodiment, assume that there are 1000 (x, y) pieces of data, as shown in table 1 below, for an example of test sample data:
| control information | Number of Bug |
| com.example.test,MainActivity,Button,200,200 | 6 |
| com.meetyou.app1,RecordActivity,ImageView,500,500 | 2 |
| com.example.test,WebViewActivity,WebView,800,1000 | 1 |
| ... | ... |
| com.meetyou.app2,WebViewActivity,Button,80,200 | 3 |
Table 1 test sample data example
The control attribute information in the table above is sequentially a control test package name, an entry function name, a control type, a control width in the control information, and a control height in the control information, and it can be understood that the control attribute information in this embodiment is not limited to the information provided in the table above, and the table above is only used as an example and is not limited thereto.
Then, using the collected test sample data as a training sample, wherein the training process specifically comprises the following steps:
setting parameters of a neural network model to build the neural network model;
inputting the preprocessed test sample data as a training sample set of the neural network model;
and training the neural network model through the training sample set and updating parameters of the neural network model to obtain the trained neural network model.
In this embodiment, the preprocessing includes feature extraction and data cleansing.
Wherein the feature extraction specifically comprises: clustering the non-digital control attribute information, and converting the clustered control attribute information into corresponding coding vectors according to a preset coding rule; the preset established coding rules can be defined according to actual control attribute information, the purpose is to convert the control attribute information into data, the subsequent processing is convenient, and different coding rules can be selected for different control attribute information.
For example, in table 1 above, the first two pieces of control attribute information are control test package names and entry function names, and belong to text information categories, so a clustering algorithm is selected to cluster the non-numeric control attribute information, and a K-Means clustering algorithm is optionally adopted, and the clustering idea of the K-Means clustering algorithm is as follows: for a given sample set, the sample set is divided into K clusters according to the distance between samples, and the control attribute information in the clusters is connected as close as possible, while the distance between the clusters is as large as possible.
The following description will take the clustering process of the entry function names as an example.
Assuming that 20 types of portal function names (Activity class names) exist, after K-Means clustering, 20 types of names can be clustered into A, B, C, D four categories, so that the Activity class names can be converted into corresponding coding vectors through datamation of preset coding rules, wherein the preset coding rules can be defined according to actual control attribute information and are used for datamation of the control attribute information so as to facilitate subsequent processing, and different coding rules can be selected for different control attribute information. For example, coded as (by way of example only, and not limitation): a is coded as (1,0,0,0), B is coded as (0,1,0,0), C is coded as (0,0,1,0), and D is coded as (0,0,0, 1). Assuming that the type to which the entry function name of a control belongs is a, the entry function name of the control can be converted into a code vector (1,0,0, 0).
Similarly, the names of the control test packages are clustered, and the names of the control test packages can be converted into a specific coding vector, for example, the names of the control test packages of com.
Preferably, as for the control types, since there are few types of commonly used controls, about 10, and there is not much correlation between the control types, encoding the control type of Button as (1,0,0,0,0, 0) directly without clustering, encoding the control type of ImageView as (0,1,0,0,0,0,0, 0) and so on.
In the embodiment, the control attribute information is clustered, and the clustered control attribute information is converted into the corresponding coding vector according to the preset coding rule, so that the subsequent analysis and processing are facilitated.
The data cleaning specifically comprises the following steps: and storing the coding vectors and the control Bug number according to a preset format, wherein the preset format refers to a format which is convenient to process by combining the coding vectors subjected to the digitization together, and the preset format is set according to actual requirements, for example, the data obtained after the characteristic extraction and the data cleaning in the table 1 is shown in the following table 2, and is an example of the data of the test sample after the preprocessing.
| Control information | Number of Bug |
| (1,0,0,0,0),(1,0,0,0),(1,0,0,0,0,0,0,0,0,0),200,200 | 6 |
| (0,0,1,0,0),(0,1,0,0),(0,1,0,0,0,0,0,0,0,0),500,500 | 2 |
| (1,0,0,0,0),(0,0,0,1),(0,0,1,0,0,0,0,0,0,0),800,1000 | 1 |
| ... | ... |
| (0,0,0,0,1),(0,0,0,1),(1,0,0,0,0,0,0,0,0,0),80,200 | 3 |
Table 2 sample data examples after pretreatment
The purpose of the test sample data preprocessing in the embodiment is to unify the data format, so that the data format can be used for training the neural network model.
In this embodiment, a tensoflow frame training neural network model may be selected, a gradient descent method may be further selected for model training, further, a loss function of the neural network model may be selected as a cross entropy loss function and other loss functions capable of replacing the effect of the cross entropy loss function, and further, a sigmoid activation function or a tanh activation function may be selected as the activation function.
Specifically, TensorFlow is a second-generation artificial intelligence learning system developed by Google based on DistBeief, the naming of the system is derived from the operation principle of the system, Tensor (Tensor) means an N-dimensional array, Flow (Flow) means calculation based on a data Flow graph, TensorFlow is a calculation process in which the Tensor flows from one end of the Flow graph to the other end, and TensorFlow is a system which transmits a complex data structure to an artificial intelligence neural network for analysis and processing.
The process of building the neural network includes conventional steps such as importing a tensrflow toolkit, generating a simulation data set, defining network parameters, setting input parameters and output tag values, defining a loss function and a back propagation algorithm, setting iteration times, and the like, which are not described herein in detail.
In this embodiment, the trained neural network model is saved as a PB format file, where the PB format file is a storage model format recommended by google, and the storage model format has language independence, can run independently, has a closed serialization format, can parse files using any programming language, and allows other languages and deep learning frameworks to read, continue training, and migrate. The method has the advantages that when the model is stored as a PB format file, variables of the model can be fixed, when the model is used in real offline test, data in the PB format cannot be updated and changed, the size of the model is reduced, and the method is more suitable for running on more test equipment at a mobile phone end.
Optionally, the transplantation process in the practical application of this embodiment is:
storing the PB file of the model under an asset directory folder of a source code project of a traversal test project;
putting the so file library and jar package of TensorFlow under the libs directory folder;
and loading a library file, operating the TensorFlow in a test item APP, and applying the neural network model to the Android end of the test equipment.
In this embodiment, the order of the click time is set, the control to be tested with the largest number of bugs in the control set to be tested at each click time is predicted and selected as the selection control at the current click time, and the traversal order of all the selection controls forms a traversal test path.
For example, assuming that t is the motor time of each control, when t is j-1, the click of the control K (j-1) is just completed, when t is j, the current set of controls to be tested is set as K, the number of bugs corresponding to each control in K is predicted through the neural network model and is marked as y, then the control with the largest y value is selected from the controls and is marked as Kymax as the selected control at the time j, and the traversal path of the selected control at each time is the traversal test path, which can be understood as being the optimal traversal path with the largest number of bugs.
According to the embodiment, test sample data is collected, a neural network model is trained according to the test sample data, the trained neural network model is transplanted to test equipment, then a traversal path with the maximum number of trigger bugs is predicted and selected according to the neural network model to serve as a traversal test path, and traversal test of a to-be-tested control set is performed on the test equipment according to the traversal test path. The traversing test quality and the test efficiency are improved.
Example two:
as shown in fig. 2, a block diagram of an apparatus for generating a traversal test path provided in this embodiment is used to execute the method according to the first embodiment, and includes the following modules:
the neural network model generation module 10: and the neural network model is used for collecting test sample data and training the neural network model according to the test sample data.
The neural network model transplanting module 20: the method is used for transplanting the neural network model to the test equipment.
Select traversal path module 30: and the control set to be tested is used for predicting and selecting the traversal path with the most number of triggered bugs as a traversal test path according to the neural network model, and performing traversal test on the control set to be tested on the test equipment according to the traversal test path.
Example three:
the invention also provides a device for generating traversal test paths, which comprises:
at least one processor, and a memory communicatively coupled to the at least one processor;
wherein the processor is configured to perform the method according to embodiment one by calling the computer program stored in the memory.
In addition, the present invention also provides a computer-readable storage medium, which stores computer-executable instructions for causing a computer to perform the method according to the first embodiment.
The embodiment of the invention overcomes the problems of low testing efficiency and high testing and maintaining cost caused by the fact that testing paths and parameters need to be adapted manually in the prior art, realizes the purpose of intelligently predicting the traversal testing path with the largest number of control bugs by utilizing a neural network, and improves the traversal testing quality and the testing efficiency. The method can be widely applied to the field of software traversal test.
The above embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same, although the present invention is described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention, and they should be construed as being included in the following claims and description.
Claims (7)
1. A method of generating a traversal test path, comprising:
collecting test sample data to form a test sample set and training a neural network model according to the test sample set;
transplanting the neural network model obtained by training to test equipment;
setting a to-be-tested control set, predicting and selecting a traversal path with the maximum number of trigger bugs as a traversal test path according to the neural network model, and performing traversal test on the to-be-tested control set on the test equipment according to the traversal test path;
the test sample data includes: the control attribute information comprises at least one of the following information: the control test package name, the entry function name, the control type, the control number and the control information;
the method also comprises the steps of preprocessing the test sample data, wherein the preprocessing comprises feature extraction and data cleaning;
the feature extraction specifically comprises the following steps: clustering the non-digital control attribute information, and converting the clustered control attribute information into corresponding coding vectors according to a preset coding rule;
the data cleaning specifically comprises the following steps: storing the coding vector and the control Bug number according to a preset format;
the traversal test path refers to: and setting a click time sequence, predicting and selecting the control to be tested with the maximum number of bugs in the control set to be tested at each click time as a selection control at the current click time, wherein the traversal sequence of the selection controls forms the traversal test path.
2. The method of claim 1, wherein the neural network model is trained using a Tensorflow framework, and the training process is specifically as follows:
setting parameters of the neural network model to build the neural network model;
inputting test sample data as a training sample set of the neural network model;
and training the neural network model through the training sample set and updating network parameters.
3. The method of claim 2, wherein the loss function of the neural network model is a cross-entropy loss function.
4. The method of claim 2, wherein the training of the neural network model is performed using a gradient descent method.
5. An apparatus for generating a traversal test path, comprising:
a neural network model generation module: the neural network model is used for collecting test sample data and training the neural network model according to the test sample data;
a neural network model transplanting module: for transplanting the neural network model onto a test device;
and a traversal path selecting module: the device is used for setting a control set to be tested, predicting and selecting a traversal path with the maximum number of trigger bugs as a traversal test path according to the neural network model, and performing traversal test on the control set to be tested on the test equipment according to the traversal test path;
the test sample data includes: the control attribute information comprises at least one of the following information: the control test package name, the entry function name, the control type, the control number and the control information;
the neural network model generation module is also used for preprocessing the test sample data, and the preprocessing comprises feature extraction and data cleaning;
the feature extraction specifically comprises the following steps: clustering the non-digital control attribute information, and converting the clustered control attribute information into corresponding coding vectors according to a preset coding rule;
the data cleaning specifically comprises the following steps: storing the coding vector and the control Bug number according to a preset format;
the traversal test path refers to: and setting a click time sequence, predicting and selecting the control to be tested with the maximum number of bugs in the control set to be tested at each click time as a selection control at the current click time, wherein the traversal sequence of the selection controls forms the traversal test path.
6. An apparatus for generating a traversal test path, comprising:
at least one processor; and a memory communicatively coupled to the at least one processor;
wherein the processor is adapted to perform the method of any one of claims 1 to 4 by invoking a computer program stored in the memory.
7. A computer-readable storage medium having stored thereon computer-executable instructions for causing a computer to perform the method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910984049.8A CN110895508B (en) | 2019-10-16 | 2019-10-16 | Method and device for generating traversal test path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910984049.8A CN110895508B (en) | 2019-10-16 | 2019-10-16 | Method and device for generating traversal test path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110895508A CN110895508A (en) | 2020-03-20 |
| CN110895508B true CN110895508B (en) | 2022-05-24 |
Family
ID=69785747
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910984049.8A Active CN110895508B (en) | 2019-10-16 | 2019-10-16 | Method and device for generating traversal test path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110895508B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6028994A (en) * | 1998-05-06 | 2000-02-22 | Advanced Micro Devices | Method for predicting performance of microelectronic device based on electrical parameter test data using computer model |
| CN108647520A (en) * | 2018-05-15 | 2018-10-12 | 浙江大学 | A kind of intelligent fuzzy test method and system based on fragile inquiry learning |
-
2019
- 2019-10-16 CN CN201910984049.8A patent/CN110895508B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6028994A (en) * | 1998-05-06 | 2000-02-22 | Advanced Micro Devices | Method for predicting performance of microelectronic device based on electrical parameter test data using computer model |
| CN108647520A (en) * | 2018-05-15 | 2018-10-12 | 浙江大学 | A kind of intelligent fuzzy test method and system based on fragile inquiry learning |
Non-Patent Citations (1)
| Title |
|---|
| "软件测试中选择有效测试技术的一种方法";李霁;《科技通报》;20150211(第1期);第217-220,224页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110895508A (en) | 2020-03-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110941467A (en) | Data processing method, device and system | |
| US11893364B2 (en) | Accelerating application modernization | |
| CN110991658A (en) | Model training method and device, electronic equipment and computer readable storage medium | |
| CN107463935A (en) | Application class methods and applications sorter | |
| CN111062520B (en) | Hostname feature prediction method based on random forest algorithm | |
| WO2020124240A1 (en) | Accurate and transparent path prediction using process mining | |
| CN111027629A (en) | Power distribution network fault outage rate prediction method and system based on improved random forest | |
| CN112527676A (en) | Model automation test method, device and storage medium | |
| CN108334895A (en) | Sorting technique, device, storage medium and the electronic device of target data | |
| CN109885481A (en) | A kind of the recording and playback test method and device of the game NGUI based on Unity technology | |
| CN112905451B (en) | Automatic testing method and device for application program | |
| CN110895508B (en) | Method and device for generating traversal test path | |
| CN114168446B (en) | Simulation evaluation method and device for mobile terminal operation algorithm model | |
| CN113312485B (en) | Log automatic classification method and device and computer readable storage medium | |
| US20220206773A1 (en) | Systems and methods for building and deploying machine learning applications | |
| CN114819106A (en) | Calculation graph optimization method and device, electronic equipment and computer readable medium | |
| CN115809688A (en) | Model debugging method and device, electronic equipment and storage medium | |
| CN104331507A (en) | Method and device for automatically finding and classifying machine data categories | |
| CN110796265B (en) | Interactive operation method, device, terminal equipment and medium of decision tree model | |
| WO2021262179A1 (en) | Automated machine learning: a unified, customizable, and extensible system | |
| CN114168216B (en) | Parameter tuning method, device and storage medium | |
| US20240013095A1 (en) | A system and method for the unification and optimization of machine learning inference pipelines | |
| US20230281482A1 (en) | Systems and methods for rule-based machine learning model promotion | |
| CN116475081B (en) | Industrial product sorting control method, device and system based on cloud edge cooperation | |
| CN111459917B (en) | Knowledge base management method, device and processing equipment |
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 |