CN110895718A - Method and system for training machine learning model - Google Patents
Method and system for training machine learning model Download PDFInfo
- Publication number
- CN110895718A CN110895718A CN201811041753.1A CN201811041753A CN110895718A CN 110895718 A CN110895718 A CN 110895718A CN 201811041753 A CN201811041753 A CN 201811041753A CN 110895718 A CN110895718 A CN 110895718A
- Authority
- CN
- China
- Prior art keywords
- machine learning
- learning model
- configuration
- training
- defining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
A method and system for training a machine learning model is provided. The method comprises the following steps: obtaining a configuration of a training process for defining a machine learning model; and analyzing the acquired configuration by using a model training framework aiming at the machine learning model, and executing the analyzed processing logic for training the machine learning model to train the machine learning model, wherein the configuration for limiting the training process of the machine learning model comprises at least one of the following configurations: an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model; an input arrangement for defining input data for the machine learning algorithm; a parameter configuration for defining parameters of a machine learning model; and an environment configuration for defining an environment when training the machine learning model. According to the method and system, the machine learning model can be trained based on the configuration of the training process used for defining the machine learning model.
Description
Technical Field
The present invention relates generally to the field of machine learning, and more particularly to a method and system for training a machine learning model.
Background
With the advent of massive amounts of data, people tend to use machine learning techniques to mine value from the data. Machine learning is a necessary product of the development of artificial intelligence research to a certain stage, and aims to improve the performance of the system by means of calculation and by using experience. In a computer system, "experience" is usually in the form of "data" from which a "model" can be generated by a machine learning algorithm, i.e., the empirical data is provided to the machine learning algorithm, and a model can be generated based on the empirical data; when a new situation is faced, the trained model is used for obtaining a corresponding prediction result.
At present, the basic process of generating a model based on empirical data mainly comprises:
1. importing a data set (e.g., a data table) containing historical data records;
2. completing feature engineering, wherein various processing is carried out on the attribute information of the data records in the data set to obtain various features, and a feature vector formed by the features can be used as a machine learning sample;
3. and training a model, wherein the model is learned based on the machine learning samples obtained through the feature engineering according to a set machine learning algorithm (such as a logistic regression algorithm, a decision tree algorithm, a neural network algorithm and the like).
The bottom layer of the model training framework of the current machine learning platform or system is mostly realized by using C language, and the model training framework has the characteristics of high operation efficiency, but has higher use threshold and is not suitable for being directly used by users. Thus, the underlying code is typically one-layer encapsulated by a developer using a scripting language (e.g., python), such that a user may implement the processing logic provided by the calling model training framework by writing executable code (i.e., scripts) for calling the underlying interface of the model training framework using the scripting language. However, considering the complexity of the algorithm itself and the continuous expansion of the algorithm supported by the framework, an effective user interaction mode is lacking at present, which makes it difficult for the user to conveniently develop various machine learning models.
Disclosure of Invention
Exemplary embodiments of the present invention are directed to providing a method and system for training a machine learning model, which can train out the machine learning model based on a configuration of a training process for defining the machine learning model.
In accordance with an exemplary embodiment of the present invention, there is provided a method for training a machine learning model, comprising: obtaining a configuration of a training process for defining a machine learning model; and analyzing the acquired configuration by using a model training framework aiming at the machine learning model, and executing the analyzed processing logic for training the machine learning model to train the machine learning model, wherein the configuration for limiting the training process of the machine learning model comprises at least one of the following configurations: an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model; an input arrangement for defining input data for the machine learning algorithm; a parameter configuration for defining parameters of a machine learning model; and an environment configuration for defining an environment when training the machine learning model.
Optionally, the algorithm is configured to define a configuration of a computational graph representing computational logic of the machine learning algorithm; and/or, inputting input data configured to define a manner in which input data of the machine learning algorithm is generated from raw input data; and/or, the parameter configuration is used to define at least one of: initial values of parameters of the machine learning model or a generation mode of the initial values, an updating mode of the parameters of the machine learning model and a data type of the parameters of the machine learning model; and/or the environment configuration is for defining at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a graphic processor is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
Optionally, the machine learning algorithm used to train the machine learning model comprises a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
Optionally, the step of obtaining a configuration of a training process for defining the machine learning model comprises: obtaining a script for defining a training logic of a machine learning model; a configuration for defining a training process for the machine learning model is generated based on the obtained script.
Optionally, the step of obtaining a configuration of a training process for defining the machine learning model comprises: providing a graphical interface to a user for setting a configuration for defining a training process of a machine learning model; and receiving an input operation performed on the graphical interface by a user for setting the configuration, and acquiring the configuration set by the user according to the input operation.
According to another exemplary embodiment of the invention, a system for training a machine learning model is provided, comprising: configuration acquisition means that acquires a configuration for defining a training process of a machine learning model; and the training device analyzes the acquired configuration by using a model training framework aiming at the machine learning model and executes the analyzed processing logic for training the machine learning model so as to train the machine learning model, wherein the configuration for limiting the training process of the machine learning model comprises at least one of the following configurations: an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model; an input arrangement for defining input data for the machine learning algorithm; a parameter configuration for defining parameters of a machine learning model; and an environment configuration for defining an environment when training the machine learning model.
Optionally, the algorithm is configured to define a configuration of a computational graph representing computational logic of the machine learning algorithm; and/or, inputting input data configured to define a manner in which input data of the machine learning algorithm is generated from raw input data; and/or, the parameter configuration is used to define at least one of: initial values of parameters of the machine learning model or a generation mode of the initial values, an updating mode of the parameters of the machine learning model and a data type of the parameters of the machine learning model; and/or the environment configuration is for defining at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a graphic processor is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
Optionally, the machine learning algorithm used to train the machine learning model comprises a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
Optionally, the configuration acquiring means acquires a script for defining a training logic of the machine learning model; and generating a configuration for a training process defining the machine learning model based on the retrieved script.
Optionally, the configuration acquisition means provides a user with a graphical interface for setting a configuration for defining a training process of the machine learning model; and receiving an input operation performed on the graphical interface by a user for setting the configuration, and acquiring the configuration set by the user according to the input operation.
According to another exemplary embodiment of the invention, a system is provided comprising at least one computing device and at least one storage device storing instructions, wherein the instructions, when executed by the at least one computing device, cause the at least one computing device to perform the method for training a machine learning model as described above.
According to another exemplary embodiment of the invention, a computer-readable storage medium storing instructions is provided, which when executed by at least one computing device, cause the at least one computing device to perform the method for training a machine learning model as described above.
According to the method and the system for training the machine learning model, the machine learning model can be trained based on the configuration of the training process for limiting the machine learning model, on one hand, developers do not need to further package bottom layer codes of a model training framework, and the workload of the developers is reduced; on the other hand, even if the user does not understand the programming language, the machine learning model meeting the requirements can be trained by the method, so that the use threshold of model training is reduced, and the usability of the model training is improved. In addition, the method for training a machine learning model according to an exemplary embodiment of the present invention is applicable to various machine learning algorithms, and a machine learning model can be uniformly trained by the method regardless of a conventional machine learning algorithm with a small number of parameters and a simple hierarchy or an artificial neural network algorithm with a more complex arithmetic logic.
Additional aspects and/or advantages of the present general inventive concept will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the general inventive concept.
Drawings
The above and other objects and features of exemplary embodiments of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings which illustrate exemplary embodiments, wherein:
FIG. 1 shows a flow diagram of a method for training a machine learning model according to an exemplary embodiment of the invention;
FIG. 2 illustrates a block diagram of a system for training a machine learning model according to an exemplary embodiment of the present invention.
Detailed Description
Reference will now be made in detail to the embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the like elements throughout. The embodiments are described below in order to explain the present invention by referring to the figures.
FIG. 1 shows a flowchart of a method for training a machine learning model according to an exemplary embodiment of the invention. Here, the method may be executed by a computer program, or may be executed by a hardware device or an aggregation of hardware and software resources dedicated to performing machine learning, big data computing, data analysis, or the like, for example, by a machine learning platform or framework for implementing a machine learning related business.
Referring to fig. 1, in step S10, a configuration of a training process for defining a machine learning model is acquired. Here, the configuration of the training process for defining the machine learning model includes at least one of the following configurations: algorithm configuration, input configuration, parameter configuration and environment configuration.
In step S20, the acquired configuration is analyzed using a model training framework for the machine learning model, and the analyzed processing logic for training the machine learning model is executed to train the machine learning model. In particular, a model training framework is used to parse a configuration of a training process for defining a machine learning model to derive processing logic for training the machine learning model, and then the parsed processing logic is executed to implement the process of training the machine learning model.
As an example, when the method is performed by a machine learning platform for performing a machine learning process, the model training framework for the machine learning model may be a model training framework of the machine learning platform, where the model training framework of the machine learning platform may indicate a subject logical architecture for the machine learning platform to implement model training.
Hereinafter, the algorithm configuration, the input configuration, the parameter configuration, and the environment configuration will be described in detail.
In particular, the algorithm is configured with operational logic for defining a machine learning algorithm for training a machine learning model.
As an example, an algorithm configuration may be used to define a configuration of a computational graph representing the operational logic of the machine learning algorithm. Here, the computation graph may be a directed acyclic graph. For example, an algorithmic configuration may be used to define the nodes and connection relationships between the nodes comprised by the computational graph.
It should be understood that the machine learning algorithms used to train the machine learning model may be various types and combinations of machine learning algorithms. As an example, the machine learning algorithm used to train the machine learning model may include a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
Input data configured to define the machine learning algorithm. For example, the input of a node in a computational graph representing the operational logic of the machine learning algorithm may be the input data.
As an example, the input configuration may be used to define the manner in which the input data for the machine learning algorithm is generated from the raw input data. Here, the raw input data may be machine learning samples obtained by feature engineering.
In essence, the input arrangement declares one or more input interfaces, each representing a conversion process from raw input data to a desired data format for that interface, the converted data being used as input data for the machine learning algorithm (e.g., as computational input for a computational graph). According to the exemplary embodiment of the invention, the model training framework can learn all required input interfaces and corresponding data formats thereof before executing the processing logic corresponding to the machine learning algorithm by analyzing the input configuration, and as an example, the processing logic corresponding to the input configuration obtained by analyzing can be executed first, and then the processing logic corresponding to the algorithm configuration obtained by analyzing can be executed, so that the converted input data can be stored locally in advance, rather than being converted every time when required, and the execution efficiency can be improved by optimizing the execution process. Moreover, for a user, only input interfaces required by the machine learning algorithm need to be considered during configuration, and how to optimize the execution process does not need to be considered.
The parameters are configured to define parameters of a machine learning model. Regarding parameters of the machine learning model, on the one hand, the parameters are the output of the entire machine learning model training process; on the other hand, for some machine learning algorithms (e.g., artificial neural network algorithms) used for training a machine learning model, parameters are also input to the machine learning algorithm (e.g., the input to a node in a computational graph representing the operational logic of the machine learning algorithm may be input data and parameters), but unlike the input data of the machine learning algorithm, the parameters themselves are not fixed values, but are updated iteratively as the algorithm is operated, and the updating method of the parameters also needs to be specified.
As an example, the parameter configuration may be used to define at least one of: initial values of parameters of the machine learning model or a generation method of the initial values, an update method of the parameters of the machine learning model, and a data type of the parameters of the machine learning model. For example, the data type of the parameter may include sparse and/or continuous, and if the data type of the parameter is continuous, the value of the parameter may be a value with certain continuity; if the data type of the parameter is sparse, the value of the parameter has no continuity.
The environment configuration is used to define an environment when training the machine learning model. As an example, the environment configuration may be used to define at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a Graphic Processing Unit (GPU) is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
As an example, step S10 may include: the configuration of the training process for defining the machine learning model is obtained directly from the outside. As another example, step S10 may include: providing a graphical interface to a user for setting a configuration for defining a training process of a machine learning model; and receiving an input operation performed on the graphical interface by a user for setting the configuration, and acquiring the configuration set by the user according to the input operation.
By way of example, the graphical interface may include: the system comprises a graphical interface part for setting algorithm configuration, a graphical interface part for setting input configuration, a graphical interface part for setting parameter configuration and a graphical interface part for setting environment configuration. According to the mode of the invention, the user can conveniently configure the process of training the machine learning model from four aspects of algorithm, input, parameters and environment, and even if the configuration of the training process for limiting the machine learning model which needs to be set is more, the user can systematically and logically configure through the graphical interface, thereby improving the configuration efficiency.
As an example, the graphical interface may comprise at least one of: a selection input type interface for setting a configuration, an editing type interface for editing a configuration, and a canvas type interface for setting a directed acyclic graph.
As an example, the graphical interface for setting the algorithm configuration may be a canvas-type interface for setting a computational graph. For example, the algorithm configuration set by the user may be obtained according to the operation of the user setting and configuring the nodes of the computational graph through the canvas-type interface. The graphical interface for setting input configurations, the graphical interface for setting parameter configurations, and the graphical interface for setting environment configurations may be selection input type interfaces for setting configurations and/or editing type interfaces for editing configurations.
According to the exemplary embodiment of the invention, even if a user is not familiar with the underlying programming language, the configuration of the training process for limiting the machine learning model can be well set, so that the machine learning model meeting the requirement can be trained, and the use threshold for training the machine learning model is reduced.
As an example, step S10 may include: obtaining a script for defining a training logic of a machine learning model; then, a configuration for defining a training process of the machine learning model is generated based on the acquired script. In this way, the method for training a machine learning model according to the exemplary embodiment of the present invention can support both the configuration of the training process for defining the machine learning model and the script for defining the training logic of the machine learning model.
FIG. 2 illustrates a block diagram of a system for training a machine learning model according to an exemplary embodiment of the present invention. As shown in fig. 2, a system for training a machine learning model according to an exemplary embodiment of the present invention includes: the acquisition means 10 and the training means 20 are configured.
In particular, the configuration acquisition means 10 are used to acquire a configuration of a training process for defining a machine learning model. Here, the configuration of the training process for defining the machine learning model includes at least one of the following configurations: algorithm configuration, input configuration, parameter configuration and environment configuration.
The training apparatus 20 is configured to analyze the obtained configuration using a model training framework for the machine learning model, and execute the analyzed processing logic for training the machine learning model to train the machine learning model. Specifically, the training apparatus 20 uses the model training framework to parse the configuration of the training process for defining the machine learning model to obtain the processing logic for training the machine learning model, and then executes the parsed processing logic to implement the process of training the machine learning model.
Hereinafter, the algorithm configuration, the input configuration, the parameter configuration, and the environment configuration will be described in detail.
In particular, the algorithm is configured with operational logic for defining a machine learning algorithm for training a machine learning model.
As an example, an algorithm configuration may be used to define a configuration of a computational graph representing the operational logic of the machine learning algorithm. Here, the computation graph may be a directed acyclic graph. For example, an algorithmic configuration may be used to define the nodes and connection relationships between the nodes comprised by the computational graph.
It should be appreciated that the machine learning algorithm used to train the machine learning model may be various types of machine learning algorithms. As an example, the machine learning algorithm used to train the machine learning model may include a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
Input data configured to define the machine learning algorithm. For example, the input of a node in a computational graph representing the operational logic of the machine learning algorithm may be the input data.
As an example, the input configuration may be used to define the manner in which the input data for the machine learning algorithm is generated from the raw input data. Here, the raw input data may be machine learning samples obtained by feature engineering.
In essence, the input arrangement declares one or more input interfaces, each representing a conversion process from raw input data to a desired data format for that interface, the converted data being used as input data for the machine learning algorithm (e.g., as computational input for a computational graph). According to an exemplary embodiment of the present invention, the model training framework may obtain all required input interfaces and data formats thereof before executing the processing logic corresponding to the machine learning algorithm by analyzing the input configuration, and as an example, the training apparatus 20 may execute the processing logic corresponding to the input configuration obtained by analyzing first and then execute the processing logic corresponding to the algorithm configuration obtained by analyzing second, so that the converted input data may be stored locally in advance, instead of being converted every time when required, and by optimizing the execution process, the execution efficiency may be improved. Moreover, for a user, only input interfaces required by the machine learning algorithm need to be considered during configuration, and how to optimize the execution process does not need to be considered.
The parameters are configured to define parameters of a machine learning model. Regarding parameters of the machine learning model, on the one hand, the parameters are the output of the entire machine learning model training process; on the other hand, for some machine learning algorithms (e.g., artificial neural network algorithms) used for training a machine learning model, parameters are also input to the machine learning algorithm (e.g., the input to a node in a computational graph representing the operational logic of the machine learning algorithm may be input data and parameters), but unlike the input data of the machine learning algorithm, the parameters themselves are not fixed values, but are updated iteratively as the algorithm is operated, and the updating method of the parameters also needs to be specified.
As an example, the parameter configuration may be used to define at least one of: initial values of parameters of the machine learning model or a generation method of the initial values, an update method of the parameters of the machine learning model, and a data type of the parameters of the machine learning model. For example, the data type of the parameter may include sparse and/or continuous, and if the data type of the parameter is continuous, the value of the parameter may be a value with certain continuity; if the data type of the parameter is sparse, the value of the parameter has no continuity.
The environment configuration is used to define an environment when training the machine learning model. As an example, the environment configuration may be used to define at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a Graphic Processing Unit (GPU) is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
As an example, the configuration acquisition means 10 may directly acquire the configuration of the training process for defining the machine learning model from the outside. As another example, the configuration acquisition apparatus 10 may provide a graphical interface for setting a configuration for defining a training process of the machine learning model to a user; and receiving an input operation performed on the graphical interface by a user for setting the configuration, and acquiring the configuration set by the user according to the input operation.
By way of example, the graphical interface may include: the system comprises a graphical interface part for setting algorithm configuration, a graphical interface part for setting input configuration, a graphical interface part for setting parameter configuration and a graphical interface part for setting environment configuration. According to the mode of the invention, the user can conveniently configure the process of training the machine learning model from four aspects of algorithm, input, parameters and environment, and even if the configuration of the training process for limiting the machine learning model which needs to be set is more, the user can systematically and logically configure through the graphical interface, thereby improving the configuration efficiency.
As an example, the graphical interface may comprise at least one of: a selection input type interface for setting a configuration, an editing type interface for editing a configuration, and a canvas type interface for setting a directed acyclic graph.
As an example, the graphical interface for setting the algorithm configuration may be a canvas-type interface for setting a computational graph. For example, the algorithm configuration set by the user may be obtained according to the operation of the user setting and configuring the nodes of the computational graph through the canvas-type interface. The graphical interface for setting input configurations, the graphical interface for setting parameter configurations, and the graphical interface for setting environment configurations may be selection input type interfaces for setting configurations and/or editing type interfaces for editing configurations.
According to the exemplary embodiment of the present invention, even if the user does not understand the programming language, the configuration of the training process for defining the machine learning model can be well set, so that the machine learning model satisfying the requirement can be trained, and the use threshold for training the machine learning model is reduced.
As an example, the configuration acquisition means 10 may acquire a script for defining training logic of the machine learning model; and generating a configuration for a training process defining the machine learning model based on the retrieved script. In this way, the system for training a machine learning model according to an exemplary embodiment of the present invention can support both the configuration of the training process for defining the machine learning model and the script for defining the training logic of the machine learning model.
It should be appreciated that a specific implementation of the system for training a machine learning model according to an exemplary embodiment of the present invention may be implemented with reference to the related specific implementation described in conjunction with fig. 1, and will not be described herein again.
The apparatus included in a system for training a machine learning model according to an exemplary embodiment of the present invention may be software, hardware, firmware, or any combination thereof, each configured to perform a specific function. These means may correspond, for example, to a dedicated integrated circuit, to pure software code, or to a module combining software and hardware. Further, one or more functions implemented by these apparatuses may also be collectively performed by components in a physical entity device (e.g., a processor, a client, a server, or the like).
It is to be understood that the method for training a machine learning model according to an exemplary embodiment of the present invention may be implemented by a program recorded on a computer-readable medium, for example, according to an exemplary embodiment of the present invention, there may be provided a computer-readable medium for uniformly performing feature extraction, wherein a computer program for performing the following method steps is recorded on the computer-readable medium: obtaining a configuration of a training process for defining a machine learning model; and analyzing the acquired configuration by using a model training framework aiming at the machine learning model, and executing the analyzed processing logic for training the machine learning model to train the machine learning model, wherein the configuration for limiting the training process of the machine learning model comprises at least one of the following configurations: an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model; an input arrangement for defining input data for the machine learning algorithm; a parameter configuration for defining parameters of a machine learning model; and an environment configuration for defining an environment when training the machine learning model.
The computer program in the computer-readable medium may be executed in an environment deployed in a computer device such as a client, a host, a proxy device, a server, etc., and it should be noted that the computer program may also be used to perform additional steps other than the above steps or perform more specific processing when the above steps are performed, and the content of the additional steps and the further processing are described with reference to fig. 1, and will not be described again to avoid repetition.
It should be noted that the system for training a machine learning model according to an exemplary embodiment of the present invention may completely rely on the execution of a computer program to implement the corresponding functions, i.e., each device corresponds to each step in the functional architecture of the computer program, so that the entire system is called by a special software package (e.g., lib library) to implement the corresponding functions.
On the other hand, the respective means included in the system for training a machine learning model according to an exemplary embodiment of the present invention may also be implemented by hardware, software, firmware, middleware, microcode, or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segments to perform the corresponding operations may be stored in a computer-readable medium such as a storage medium, so that a processor may perform the corresponding operations by reading and executing the corresponding program code or code segments.
For example, exemplary embodiments of the present invention may also be implemented as a computing device comprising a storage component having stored therein a set of computer-executable instructions that, when executed by the processor, perform a method for training a machine learning model.
In particular, the computing devices may be deployed in servers or clients, as well as on node devices in a distributed network environment. Further, the computing device may be a PC computer, tablet device, personal digital assistant, smart phone, web application, or other device capable of executing the set of instructions described above.
The computing device need not be a single computing device, but can be any device or collection of circuits capable of executing the instructions (or sets of instructions) described above, individually or in combination. The computing device may also be part of an integrated control system or system manager, or may be configured as a portable electronic device that interfaces with local or remote (e.g., via wireless transmission).
In the computing device, the processor may include a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a programmable logic device, a special purpose processor system, a microcontroller, or a microprocessor. By way of example, and not limitation, processors may also include analog processors, digital processors, microprocessors, multi-core processors, processor arrays, network processors, and the like.
Certain operations described in the method for training a machine learning model according to the exemplary embodiments of the present invention may be implemented by software, certain operations may be implemented by hardware, and further, the operations may be implemented by a combination of hardware and software.
The processor may execute instructions or code stored in one of the memory components, which may also store data. Instructions and data may also be transmitted and received over a network via a network interface device, which may employ any known transmission protocol.
The memory component may be integral to the processor, e.g., having RAM or flash memory disposed within an integrated circuit microprocessor or the like. Further, the storage component may comprise a stand-alone device, such as an external disk drive, storage array, or any other storage device usable by a database system. The storage component and the processor may be operatively coupled or may communicate with each other, such as through an I/O port, a network connection, etc., so that the processor can read files stored in the storage component.
Further, the computing device may also include a video display (such as a liquid crystal display) and a user interaction interface (such as a keyboard, mouse, touch input device, etc.). All components of the computing device may be connected to each other via a bus and/or a network.
The operations involved in a method for training a machine learning model according to an exemplary embodiment of the present invention may be described as various interconnected or coupled functional blocks or functional diagrams. However, these functional blocks or functional diagrams may be equally integrated into a single logic device or operated on by non-exact boundaries.
For example, as described above, a computing device for training a machine learning model according to an exemplary embodiment of the present invention may include a storage component and a processor, wherein the storage component has stored therein a set of computer-executable instructions that, when executed by the processor, perform the steps of: obtaining a configuration of a training process for defining a machine learning model; and analyzing the acquired configuration by using a model training framework aiming at the machine learning model, and executing the analyzed processing logic for training the machine learning model to train the machine learning model, wherein the configuration for limiting the training process of the machine learning model comprises at least one of the following configurations: an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model; an input arrangement for defining input data for the machine learning algorithm; a parameter configuration for defining parameters of a machine learning model; and an environment configuration for defining an environment when training the machine learning model.
While exemplary embodiments of the invention have been described above, it should be understood that the above description is illustrative only and not exhaustive, and that the invention is not limited to the exemplary embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the invention. Therefore, the protection scope of the present invention should be subject to the scope of the claims.
Claims (10)
1. A method for training a machine learning model, comprising:
obtaining a configuration of a training process for defining a machine learning model; and
using a model training framework for the machine learning model to analyze the obtained configuration, and executing the analyzed processing logic for training the machine learning model to train the machine learning model,
wherein the configuration of the training process for defining the machine learning model comprises at least one of:
an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model;
an input arrangement for defining input data for the machine learning algorithm;
a parameter configuration for defining parameters of a machine learning model; and
an environment configuration for defining an environment when training the machine learning model.
2. The method of claim 1, wherein,
an algorithm configuration for defining a configuration of a computational graph representing operational logic of the machine learning algorithm;
and/or, inputting input data configured to define a manner in which input data of the machine learning algorithm is generated from raw input data;
and/or, the parameter configuration is used to define at least one of: initial values of parameters of the machine learning model or a generation mode of the initial values, an updating mode of the parameters of the machine learning model and a data type of the parameters of the machine learning model;
and/or the environment configuration is for defining at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a graphic processor is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
3. The method of claim 1, wherein the machine learning algorithm used to train the machine learning model comprises a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
4. The method of claim 1, wherein obtaining a configuration of a training process for defining a machine learning model comprises:
obtaining a script for defining a training logic of a machine learning model;
a configuration for defining a training process for the machine learning model is generated based on the obtained script.
5. The method of claim 1, wherein obtaining a configuration of a training process for defining a machine learning model comprises:
providing a graphical interface to a user for setting a configuration for defining a training process of a machine learning model;
and receiving an input operation performed on the graphical interface by a user for setting the configuration, and acquiring the configuration set by the user according to the input operation.
6. A system for training a machine learning model, comprising:
configuration acquisition means that acquires a configuration for defining a training process of a machine learning model; and
a training device for analyzing the obtained configuration by using a model training framework for the machine learning model and executing the analyzed processing logic for training the machine learning model to train the machine learning model,
wherein the configuration of the training process for defining the machine learning model comprises at least one of:
an algorithm configuration for defining an arithmetic logic of a machine learning algorithm for training a machine learning model;
an input arrangement for defining input data for the machine learning algorithm;
a parameter configuration for defining parameters of a machine learning model; and
an environment configuration for defining an environment when training the machine learning model.
7. The system of claim 6, wherein,
an algorithm configuration for defining a configuration of a computational graph representing operational logic of the machine learning algorithm;
and/or, inputting input data configured to define a manner in which input data of the machine learning algorithm is generated from raw input data;
and/or, the parameter configuration is used to define at least one of: initial values of parameters of the machine learning model or a generation mode of the initial values, an updating mode of the parameters of the machine learning model and a data type of the parameters of the machine learning model;
and/or the environment configuration is for defining at least one of: the method comprises the steps of storing paths of original input data, storing paths of parameters of a trained machine learning model, the number of operation threads used by processing logic obtained by analysis, whether a graphic processor is used for executing the processing logic obtained by analysis, whether a cluster is used for executing the processing logic obtained by analysis and the cluster used by the processing logic obtained by analysis.
8. The system of claim 6, wherein the machine learning algorithm used to train the machine learning model comprises a neural network algorithm, a logistic regression algorithm, or a decision tree algorithm.
9. A system comprising at least one computing device and at least one storage device storing instructions that, when executed by the at least one computing device, cause the at least one computing device to perform the method for training a machine learning model of any of claims 1 to 5.
10. A computer-readable storage medium storing instructions that, when executed by at least one computing device, cause the at least one computing device to perform the method for training a machine learning model of any of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811041753.1A CN110895718A (en) | 2018-09-07 | 2018-09-07 | Method and system for training machine learning model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811041753.1A CN110895718A (en) | 2018-09-07 | 2018-09-07 | Method and system for training machine learning model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110895718A true CN110895718A (en) | 2020-03-20 |
Family
ID=69785056
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811041753.1A Pending CN110895718A (en) | 2018-09-07 | 2018-09-07 | Method and system for training machine learning model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110895718A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340240A (en) * | 2020-03-25 | 2020-06-26 | 第四范式（北京）技术有限公司 | Method and device for realizing automatic machine learning |
| CN111523676A (en) * | 2020-04-17 | 2020-08-11 | 第四范式（北京）技术有限公司 | Method and device for assisting machine learning model to be online |
| CN111831322A (en) * | 2020-04-15 | 2020-10-27 | 中国人民解放军军事科学院战争研究院 | Machine learning parameter configuration method for multi-level user |
| WO2023044631A1 (en) * | 2021-09-22 | 2023-03-30 | Siemens Aktiengesellschaft | A device, system, method and storage medium for ai application deployment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550746A (en) * | 2015-12-08 | 2016-05-04 | 北京旷视科技有限公司 | Training method and training device of machine learning model |
| CN106250986A (en) * | 2015-06-04 | 2016-12-21 | 波音公司 | Advanced analysis base frame for machine learning |
| CN106250987A (en) * | 2016-07-22 | 2016-12-21 | 无锡华云数据技术服务有限公司 | A kind of machine learning method, device and big data platform |
| CN106779088A (en) * | 2016-12-06 | 2017-05-31 | 北京物思创想科技有限公司 | Perform the method and system of machine learning flow |
| CN107766946A (en) * | 2017-09-28 | 2018-03-06 | 第四范式（北京）技术有限公司 | Generate the method and system of the assemblage characteristic of machine learning sample |
| CN108021986A (en) * | 2017-10-27 | 2018-05-11 | 平安科技（深圳）有限公司 | Electronic device, multi-model sample training method and computer-readable recording medium |
-
2018
- 2018-09-07 CN CN201811041753.1A patent/CN110895718A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106250986A (en) * | 2015-06-04 | 2016-12-21 | 波音公司 | Advanced analysis base frame for machine learning |
| CN105550746A (en) * | 2015-12-08 | 2016-05-04 | 北京旷视科技有限公司 | Training method and training device of machine learning model |
| CN106250987A (en) * | 2016-07-22 | 2016-12-21 | 无锡华云数据技术服务有限公司 | A kind of machine learning method, device and big data platform |
| CN106779088A (en) * | 2016-12-06 | 2017-05-31 | 北京物思创想科技有限公司 | Perform the method and system of machine learning flow |
| CN107766946A (en) * | 2017-09-28 | 2018-03-06 | 第四范式（北京）技术有限公司 | Generate the method and system of the assemblage characteristic of machine learning sample |
| CN108021986A (en) * | 2017-10-27 | 2018-05-11 | 平安科技（深圳）有限公司 | Electronic device, multi-model sample training method and computer-readable recording medium |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111340240A (en) * | 2020-03-25 | 2020-06-26 | 第四范式（北京）技术有限公司 | Method and device for realizing automatic machine learning |
| CN111831322A (en) * | 2020-04-15 | 2020-10-27 | 中国人民解放军军事科学院战争研究院 | Machine learning parameter configuration method for multi-level user |
| CN111831322B (en) * | 2020-04-15 | 2023-08-01 | 中国人民解放军军事科学院战争研究院 | Multi-level user-oriented machine learning parameter configuration method |
| CN111523676A (en) * | 2020-04-17 | 2020-08-11 | 第四范式（北京）技术有限公司 | Method and device for assisting machine learning model to be online |
| CN111523676B (en) * | 2020-04-17 | 2024-04-12 | 第四范式（北京）技术有限公司 | Method and device for assisting machine learning model to be online |
| WO2023044631A1 (en) * | 2021-09-22 | 2023-03-30 | Siemens Aktiengesellschaft | A device, system, method and storage medium for ai application deployment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11948003B2 (en) | System and method for automated production and deployment of packaged AI solutions | |
| KR102493449B1 (en) | Edge computing test methods, devices, electronic devices and computer-readable media | |
| WO2019129060A1 (en) | Method and system for automatically generating machine learning sample | |
| CN108304201B (en) | Object updating method, device and equipment | |
| CN110895718A (en) | Method and system for training machine learning model | |
| US20170090875A1 (en) | Declarative design-time experience platform for code generation | |
| CN108008942B (en) | Method and system for processing data records | |
| JP7324831B2 (en) | DEPLOYMENT METHOD, DEPLOYMENT DEVICE AND ELECTRONIC DEVICE OF OPERATORS IN DEEP LEARNING FRAMEWORK | |
| JP7059508B2 (en) | Video time series motion detection methods, devices, electronic devices, programs and storage media | |
| WO2020248511A1 (en) | Method for realizing control over terminal utilization on the basis of run-time model | |
| US11861469B2 (en) | Code generation for Auto-AI | |
| EP4024228A1 (en) | System and method for batch and real-time feature calculation | |
| US10409567B2 (en) | Trimming unused dependencies using package graph and module graph | |
| CN111158651A (en) | Method, device and system for constructing artificial intelligence application | |
| WO2020038376A1 (en) | Method and system for uniformly performing feature extraction | |
| CN115237436A (en) | Application deployment method and device, electronic equipment and readable storage medium | |
| CN108898229B (en) | Method and system for constructing machine learning modeling process | |
| CN110868324A (en) | Service configuration method, device, equipment and storage medium | |
| WO2021129812A1 (en) | Method and system for running artificial intelligence application, and device | |
| CN108681426B (en) | Method and system for performing feature processing on data | |
| CN114764296A (en) | Machine learning model training method and device, electronic equipment and storage medium | |
| CN111125451B (en) | Data production processing method and device, electronic equipment and storage medium | |
| US11281477B2 (en) | Assigning computing resources to execution of evaluators for image animation | |
| CN114968225A (en) | Micro-service unified construction method, environment generation method and related equipment | |
| CN108960433B (en) | Method and system for running machine learning modeling process |
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 |