WO2021213234A1 - 一种机器学习应用的提供方法、装置、电子设备及存储介质 - Google Patents

一种机器学习应用的提供方法、装置、电子设备及存储介质 Download PDF

Info

Publication number
WO2021213234A1
WO2021213234A1 PCT/CN2021/087310 CN2021087310W WO2021213234A1 WO 2021213234 A1 WO2021213234 A1 WO 2021213234A1 CN 2021087310 W CN2021087310 W CN 2021087310W WO 2021213234 A1 WO2021213234 A1 WO 2021213234A1
Authority
WO
WIPO (PCT)
Prior art keywords
machine learning
learning
circle
display position
image
Prior art date
Application number
PCT/CN2021/087310
Other languages
English (en)
French (fr)
Inventor
徐昀
唐继正
李琦
张世健
Original Assignee
第四范式(北京)技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 第四范式(北京)技术有限公司 filed Critical 第四范式(北京)技术有限公司
Publication of WO2021213234A1 publication Critical patent/WO2021213234A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning

Definitions

  • the present disclosure relates to the field of artificial intelligence technology, and in particular to a method, device, electronic device, and storage medium for providing machine learning applications.
  • the embodiments of the present disclosure provide a method, device, electronic device, and storage medium for providing a machine learning application, so that users can quickly grasp the operation of the machine learning product.
  • an embodiment of the present disclosure provides a method for providing a machine learning application, including: displaying a learning circle image at a first display position on a display screen, wherein the learning circle image is provided with an identification of a machine learning step Information, and the side part of the learning circle image displays the configuration operation entry corresponding to the step; receiving the configuration information input by the user through the configuration operation entry and corresponding to the step; Under the condition of the learning circle start instruction, move the learning circle image from the first display position to the second display position on the display screen.
  • At the second display position at least part of the learning circle image is displayed in In the display screen, and at least part of the displayed learning circle image is provided with identification information of the machine learning steps; based on the configuration information corresponding to the steps, each of the machine learning steps is executed in sequence to Obtain and provide the machine learning application.
  • an embodiment of the present disclosure further provides an apparatus for providing a machine learning application, including: a first display module configured to display a learning circle image at a first display position on the display screen, wherein the learning circle image The identification information of the steps of machine learning is provided in the machine learning step, and the side part of the learning circle image displays the configuration operation entry corresponding to the step; the configuration module is configured to receive the user input through the configuration operation entry, and The configuration information corresponding to the steps; the second display module is configured to move the learning circle image from the first display position on the display screen to A second display position, in which at least part of the learning circle image is displayed on the display screen, and at least part of the displayed learning circle image is provided with an identification of the step of machine learning Information; an execution module configured to sequentially execute each of the machine learning steps based on the configuration information corresponding to the steps to obtain and provide the machine learning application.
  • the embodiments of the present disclosure also provide an electronic device, the electronic device includes: a housing, a processor, a memory, a circuit board, and a power circuit, wherein the circuit board is arranged inside a space enclosed by the housing, and the processing
  • the processor and the memory are configured on the circuit board;
  • the power supply circuit is configured to supply power to the various circuits or devices of the above electronic equipment;
  • the memory is configured to store executable program codes;
  • the processor reads the executable program codes stored in the memory, It is configured to execute the method for providing a machine learning application described in any of the foregoing embodiments.
  • the embodiments of the present disclosure also provide a computer-readable storage medium, the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors, In order to realize the method for providing a machine learning application described in any of the foregoing embodiments.
  • an embodiment of the present disclosure also provides an application program that is executed to implement the method provided in any embodiment of the present disclosure.
  • the method, device, electronic device, and storage medium for providing machine learning applications provided by the embodiments of the present disclosure are based on the Cooper learning circle theory in pedagogy.
  • the machine learning process is divided into different steps, and the different step identifiers are displayed in Learning circle image.
  • the learning circle image with the identification information of the machine learning step is first displayed in the first display position on the display screen. After corresponding configuration is performed through the configuration operation portal, the learning circle image is automatically configured after receiving the learning circle start instruction.
  • the first display position on the display screen moves to the second display position.
  • the learning circle image still provides the identification information of the machine learning steps; based on the configuration information corresponding to the steps, the steps can be sequentially Perform each of the machine learning steps, obtain a machine learning model, and apply the obtained machine learning model online.
  • the learning circle image is always displayed (at least partially) in the interface displayed on the display screen, and the learning circle image is always provided with the identification of the machine learning step Information, so that users can understand the learning circle, use the learning circle, and through the change process of the learning circle image, more intuitively understand the process of machine learning, so as to facilitate faster grasp of the operation mode of machine learning products.
  • FIG. 1 is a schematic flowchart of a method for providing a machine learning application according to an embodiment of the present disclosure
  • Fig. 2 shows an interface for displaying a learning circle image on the display screen for the first time in an embodiment
  • Figure 3 shows the process of the image of the learning circle moving from the center of the interface to the left after the user has read the introduction of the guide information
  • FIG. 4 shows an image of the learning circle displayed on the first display position of the display screen in an embodiment
  • Figure 5 shows the process of moving the image of the learning circle from the first display position to the second display position on the display screen after the learning circle is started
  • FIG. 6 shows a schematic diagram after the image of the learning circle is moved from the first display position to the second display position
  • Figure 7 shows the status of the learning circle after it is online, and the status of the learning circle is updated to "in service"
  • FIG. 8 is a schematic flowchart of a method for providing a machine learning application according to another embodiment of the present disclosure.
  • FIG. 9 is a block diagram of an apparatus for providing a machine learning application according to an embodiment of the present disclosure.
  • FIG. 10 is a block diagram of an apparatus for providing a machine learning application according to another embodiment of the present disclosure.
  • FIG. 11 is a schematic structural diagram of an embodiment of the electronic device of the present disclosure.
  • the embodiments of the present disclosure provide a method and device for providing machine learning applications, and present the steps of machine learning to users in the form of learning circles, so that users can quickly grasp the operation of machine learning products.
  • the machine learning product mentioned in the embodiments of the present disclosure can be an artificial intelligence AI application development tool.
  • the user can use the product to build a machine learning model, and can directly provide the built machine learning model to the user, or follow The user needs to deploy the model directly to the online application.
  • Fig. 1 is a schematic flowchart of a method for providing a machine learning application according to an embodiment of the present disclosure; referring to Fig. 1, the method for providing a machine learning application in this embodiment includes steps S100 to S106:
  • the learning circle in this disclosure is based on the Cooper learning circle theory in pedagogy.
  • the process of machine learning refers to the process of human learning and is defined as a learning circle.
  • machine learning that is, a machine learning model (also called a machine learning model) is defined as a learning circle.
  • the training process of artificial intelligence model is divided into different steps, and these steps are in the form or pattern of machine learning in a continuous loop.
  • the learning circle image is an image of the learning circle presented based on the above-mentioned learning circle principle, and may also be called a learning circle graphic or a learning circle pattern.
  • the learning circle image may include identification information of each step of machine learning, and may also reflect execution relationships such as sequence and cycle between steps.
  • the shape of the learning circle image can be specifically presented as a circle shape.
  • the identification information of each step of machine learning can be shown in accordance with the machine learning
  • the sequence of execution of each step is arranged end to end to form a circle image (or the form of circle).
  • the step identification information of each step of machine learning can be arranged in a fan shape on the circle image.
  • the shape of the learning circle image is not limited to the shape of the circle.
  • the identification information of each step of machine learning can be arranged in the first quadrant to the first quadrant to In the fourth quadrant, a four-quadrant image is formed.
  • the identification information of each step of the machine learning can also be arranged in a curve or arc according to the sequence of execution of the steps of the machine learning, and so on.
  • the steps of machine learning provided in the learning circle image may include the steps of collecting behavior data, collecting feedback data, training the machine learning model, and applying the machine learning model.
  • the behavior data refers to historical behavior data, and each piece of behavior data corresponds to a historical “behavior”. For example, in a basketball game, shooting is this "behavior", and a piece of behavioral data represents a historical shot in a basketball game in the past.
  • the data that can be used to describe the behavior of shooting such as "shooting player, posture, position, longitude, latitude, opponent, assist, distance to the basket, remaining time, opponent team" are all included in the behavior data. information. "Kobe, jumper, three-pointer position, 33.7693, -118.1798, Patterson (opponent), Payton (assist), 28 feet, 1 second, Blazers (against the team)" This is a piece of behavioral data. Corresponding to Bryant's final 1-second shot in the 2004 regular season finale.
  • Feedback data is the "feedback" of historical behavior and represents the results produced by behavioral data.
  • the result of pitching is a hit or a miss. If the pitch is scored, the feedback data is "hit".
  • a large amount of behavioral data and the number of feedbacks collected are used as training data, and supervised machine learning is performed to obtain a machine learning model.
  • Applying a machine learning model is to apply a trained machine learning model to a business system to make predictions to obtain prediction results.
  • the identification information can be presented in a variety of ways, such as graphics and text.
  • the identification information of the step of the machine learning may specifically include at least one of the step name and the step icon.
  • the identification information is a machine learning step name, such as a Chinese name or an English name, etc.; in another example, the identification information is a machine learning step icon, such as arrows, symbols, etc., different steps Different shapes or different colors of icons can be used to distinguish and represent; in another example, the identification information is a combination of machine learning step names and step icons.
  • the name of the step for collecting behavior data is "behavior”
  • the name of the step for collecting feedback data is "feedback”
  • the name of the step for training the machine learning model is "learning”
  • the application machine The step name of the learning model is "Apply”.
  • the display screen that displays the image of the learning circle can be the display screen of a desktop computer (commonly known as a desktop computer), a notebook computer, or a tablet computer, the display screen of a server, or the display screen of an industry-specific computing device.
  • the display screen can be a touch screen or a non-touch screen.
  • the first display position on the display screen may be the central display area of the display screen, or other display areas deviated from the central display area by a certain distance.
  • the display of the first display position of the learning circle image on the display screen may be triggered by the user.
  • the user can create a learning circle image by clicking the "create learning circle” button in the operation interface displayed on the display screen, that is, start the display of the learning circle image at the first display position on the display screen.
  • the displaying the learning circle image at the first display position on the display screen may include: under the condition that the user issues (for example, by clicking the "create learning circle” button) a learning circle creation instruction, the electronic device After receiving the learning circle creation instruction issued by the user, the learning circle image is displayed on the first display position of the display screen of the electronic device.
  • the guidance information about the learning circle can be automatically played in the interface displayed on the display screen, and the guidance information can include the flow of the learning circle and Information such as basic concepts is convenient for users to quickly understand and master the operation methods of the learning circle.
  • Fig. 2 shows an interface where a learning circle image is displayed on the display screen for the first time in an embodiment.
  • the learning circle image can be moved from the first display position to a predetermined direction (such as to the left) until the learning circle image is moved to the first display position of the display screen.
  • Figure 3 shows how the image of the learning circle moves from the center of the interface to the left after the user has read the introduction of the guide information.
  • Figure 4 shows an image of the learning circle displayed at the first display position of the display screen in an embodiment.
  • the learning circle image at the first display position in addition to providing identification information of the machine learning step, it can also Provides the execution relationship between the steps of machine learning, so that users can clearly understand the sequence and cycle of the steps.
  • the sequential execution relationship between adjacent steps can be indicated by the arrow pointing, and the step at the starting point of the arrow is executed before the step at the end of the arrow.
  • the size of the number can be used to indicate the execution relationship between the steps.
  • the steps corresponding to the small numbers are executed before the steps corresponding to the large numbers.
  • the step of collecting behavior data corresponding to the number "1" is before the number
  • the step of collecting feedback data corresponding to "2" is executed.
  • S102 Receive configuration information corresponding to the step that is input by the user through the configuration operation portal.
  • the side of the learning circle image displays configuration operation portals corresponding to the machine learning steps.
  • the configuration operation portal Through the configuration operation portal, the data or parameters required to perform each step can be configured.
  • the collection of the result (that is, feedback data) generated by the behavior data can be configured. Since the feedback data corresponds to the behavior data, it is possible to configure the behavior data through the behavior data configuration operation portal while automatically configuring the feedback data, that is, there is no need to separately set the feedback data configuration operation portal.
  • the frequency of machine learning Through the learning configuration operation entry corresponding to the steps of training the machine learning model, the frequency of machine learning, the time range of training data used for machine learning, the split ratio of training data and test data, the accuracy of machine learning, and machine learning can be adjusted. Time, etc. to be configured.
  • the identification information of the first step when the user inputs configuration information through the configuration operation portal of the first step of the steps, at least one of the following may be highlighted:
  • the way of highlighting can be highlight display, underline display, increase font size display, bold font display, etc.
  • the highlighting manner of the identification information in the first step may be the same as or different from the highlighting manner of the configuration operation portal in the first step.
  • the first step here generally refers to any of the multiple steps of the machine learning model.
  • the identification information of the step of collecting behavior data in Fig. 4 is highlighted in a highlighted manner.
  • the learning circle image at the first display position also provides machine learning status prompt information.
  • the user inputs configuration information through the configuration operation portal.
  • the state prompt information of the machine learning is displayed in the first preset manner in the learning circle image.
  • the first preset mode may be to display the words "in configuration" in the center area of the learning circle. In one embodiment, it may also be displayed on the side or around the words "in configuration” with a scrolling effect. Bar or circular pattern to increase the dynamic effect of the status prompt information and enhance the user experience.
  • Figure 4 shows the configuration page of the learning circle.
  • the current state of the learning circle-in configuration is displayed.
  • the upper navigation bar prompts the user the name of the current learning circle and the buttons for global operations.
  • the user can start or delete the current learning circle; showing the four step cards on the right allows the user to adjust the specific parameters of the four steps of the learning circle.
  • the start learning circle button in the upper right corner of the navigation to start the machine learning process in the learning circle.
  • S104 Move the learning circle image from the first display position on the display screen to the second display position.
  • the learning circle can be started to perform the steps of machine learning.
  • the learning circle image can be moved from the first display position on the display screen to the second display position, and the learning circle is displayed in the second display position.
  • the state after startup Specifically, under the condition of receiving the learning circle start instruction issued by the user, initialize the machine learning resources, and after the initialization is completed, move the learning circle image from the first display position on the display screen to the second display position , To reflect the current state of machine learning through the change of the position of the learning circle.
  • the learning circle image moves from the first display position to the second display position before the learning circle image
  • the status prompt information of the machine learning is displayed.
  • the first preset mode may be to display the words "Starting" in the central area of the learning circle. In one embodiment, it may also display a scrolling effect on the side or around the words "Starting". Bar or circular pattern to increase the dynamic effect of the status prompt information and enhance the user experience.
  • Figure 5 shows the process of moving the learning circle image from the first display position to the second display position on the display screen after the learning circle is started.
  • FIG. 6 shows a schematic diagram of the learning circle image after moving from the first display position to the second display position.
  • the learning circle image After the learning circle image is moved from the first display position to the second display position on the display screen, in the second display position, the learning circle image can all be displayed on the display screen, and the machine is provided in the displayed learning circle image Identification information of the learning step.
  • the second display position only part of the learning circle image (for example, 1/3, 1/2, or 2/3, etc.) is displayed on the display screen.
  • the learning circle Approximately 2/3 of the image is displayed on the display.
  • the identification information of the machine learning step is displayed in the partial image of the learning circle image displayed on the display screen. Specifically, the identification information of each step of the machine learning is arranged end to end according to the sequence of execution of the steps of the machine learning, and is displayed in the partial image of the learning circle image displayed on the display screen.
  • the learning circle image at the second display position at least part of the execution relationship between the steps may also be provided.
  • the display mode of the execution relationship between the steps in the learning circle image in the second display position may be consistent with the display mode in the learning circle image in the first display position.
  • S106 Based on the configuration information corresponding to the steps, sequentially execute each of the machine learning steps to obtain and provide the machine learning application.
  • each of the machine learning steps can be sequentially executed to obtain and provide the Machine learning applications.
  • sequentially executing each of the machine learning steps to obtain and provide the machine learning application may include: collecting behavior based on the configuration information corresponding to the step of collecting behavior data Data; collect feedback data based on the configuration information corresponding to the step of collecting feedback data; based on the configuration information corresponding to the step of training the machine learning model, use the collected behavior data and the feedback data as training data to perform machine learning Train to obtain a machine learning model; apply the obtained machine learning model online.
  • Figure 7 shows the status of the learning circle after it is online, and the status of the learning circle is updated to "in service”. At the same time, the online performance of the machine learning model is shown. At the top of the display interface, the global operation buttons of the learning circle have changed; at this time, the user can choose to apply to take the learned machine learning model offline or manually launch a new machine learning model.
  • the process of machine learning model training is divided into different steps, and the different steps are displayed in the learning circle image.
  • the learning circle image with the identification information of the machine learning step is first displayed on the first display position on the display screen; after corresponding configuration is performed through the configuration operation portal, the learning circle image is automatically
  • the first display position on the display screen moves to the second display position.
  • the learning circle image still provides the identification information of the machine learning steps; based on the configuration information corresponding to the steps, the steps can be sequentially Perform each of the machine learning steps, obtain a machine learning model, and apply the obtained machine learning model online.
  • the learning circle image is always displayed (at least partially) in the interface displayed on the display screen, and the learning circle image is always provided with the identification of the machine learning step Information, so that users can understand the learning circle, use the learning circle, and through the change process of the learning circle image, understand the process of machine learning more intuitively, so as to facilitate the quicker grasp of the operation mode of machine learning products.
  • the identification information of the step corresponding to the currently executed machine learning step is highlighted.
  • the way of highlighting can be highlight display, underline display, increase font size display, bold font display, etc.
  • the identification information of the step of training the machine learning model in FIG. 6 and the identification information of the step of online application of the machine learning model in FIG. 7 are respectively highlighted in a highlighted manner.
  • the identification information of the steps corresponding to the steps of machine learning is also highlighted in sequence accordingly, making it easier for users to understand and master the learning circle.
  • the learning circle image at the second display position also provides machine learning status prompt information. Specifically, when each of the machine learning steps is executed in sequence At the same time, the state prompt information of the machine learning corresponding to the step is displayed in the second preset manner in the learning circle image.
  • the second preset method may be to display "learning in progress" in the center area of the learning circle. "", refer to Figure 6; when the machine learning model is applied online, the second preset mode may be to display the words "in application” in the central area of the learning circle, refer to Figure 7.
  • a bar or circular pattern with a scrolling effect can also be displayed on the side or around the words "learning” and “applying” to increase the dynamic effect of the status prompt information and enhance the user experience.
  • the second step of the machine learning step when executed, display on the side of the partial image of the learning circle image
  • the current execution status information of the second step here generally refers to any of the multiple steps of the machine learning model.
  • the learning state information can be displayed on the side of the partial image of the learning circle image.
  • the learning status information may include information such as the status prompt of the current learning circle in learning, the optimal effect of the current learning, the time that has been running, the estimated remaining time, the number of running rounds, the trend relationship between the running effect and the behavior data, and so on.
  • Application status information may include information such as the cumulative forecast accuracy rate, the forecast failure rate, the 7-day forecast number, the total forecast number, the number of forecast failures, the application launch time, the forecast accuracy rate and the trend relationship of the behavior data.
  • FIG. 8 is a schematic flowchart of a method for providing a machine learning application according to another embodiment of the present disclosure.
  • the self-learning of the machine learning model can also be started. Through the self-learning, the currently applied machine learning model can be optimized or replaced. Specifically, referring to FIG. 8, based on the embodiment shown in FIG. 1, in an embodiment, after the obtained machine learning model is applied online, the method may further include the steps:
  • S108 Collect new behavior data based on the configuration information corresponding to the step of collecting behavior data.
  • S110 Collect new feedback data based on the configuration information corresponding to the step of collecting feedback data.
  • S112 Perform self-learning based on the new behavior data and the new feedback data, and update or replace the obtained machine learning model.
  • the process of machine learning is divided into four steps: behavior, feedback, learning, and application.
  • the four steps continue to loop, and the machine learning model will get better and better as it continues to repeat the four steps in sequence.
  • FIG. 9 is a block diagram of an apparatus for providing a machine learning application in an embodiment of the present disclosure.
  • the apparatus for providing a machine learning application in this embodiment includes: a first display module 10, a configuration module 20, a second display module 30, and an execution Module 40; wherein the first display module 10 is configured to display a learning circle image at a first display position on the display screen, wherein the learning circle image is provided with identification information of the step of machine learning, and the learning The side part of the circle image displays the configuration operation entry corresponding to the step.
  • the configuration module 20 is configured to receive configuration information corresponding to the steps input by the user through the configuration operation portal.
  • the side of the learning circle image displays configuration operation portals corresponding to the machine learning steps.
  • the configuration operation portal Through the configuration operation portal, the data or parameters required to perform each step can be configured.
  • the collection of the result generated by the behavior data (that is, the feedback data) can be configured. Since the feedback data corresponds to the behavior data, it is possible to configure the behavior data through the behavior data configuration operation portal while automatically configuring the feedback data, that is, there is no need to separately set the feedback data configuration operation portal.
  • the frequency of machine learning Through the learning configuration operation entry corresponding to the steps of training the machine learning model, the frequency of machine learning, the time range of training data used for machine learning, the split ratio of training data and test data, the accuracy of machine learning, and machine learning can be adjusted. Time, etc. to be configured.
  • the second display module 30 is configured to move the learning circle image from the first display position on the display screen to the second display position under the condition of receiving the learning circle start instruction issued by the user, In the second display position, at least part of the learning circle image is displayed on the display screen, and identification information of the machine learning step is provided in at least part of the displayed learning circle image.
  • the execution module 40 is configured to sequentially execute each of the machine learning steps based on the configuration information corresponding to the steps to obtain and provide the machine learning application.
  • each of the machine learning steps can be sequentially executed to obtain and provide the Machine learning applications.
  • the device of this embodiment can be used to implement the technical solution of the method embodiment shown in FIG.
  • the learning circle image at the first display position also provides the execution relationship between the steps, so that the user can clearly understand the sequence of execution of the steps.
  • the sequential execution relationship between adjacent steps can be indicated by the arrow pointing, and the step at the starting point of the arrow is executed before the step at the end of the arrow.
  • the size of the number can be used to indicate the execution relationship between the steps. The steps corresponding to the small numbers are executed before the steps corresponding to the large numbers. For example, the step of collecting behavior data corresponding to the number "1" is before the number The step of collecting feedback data corresponding to "2" is executed.
  • the first display module 10 is configured to: display identification information of each step of machine learning in a first display position on the display screen, and combine the identification information of each step of machine learning, According to the sequence of execution between the steps of machine learning, they are arranged end to end to form a circle image; or, the identification information of each step of machine learning is displayed in the first display position on the display screen, and each step of machine learning The identification information of the steps is correspondingly arranged in the first quadrant to the fourth quadrant according to the sequence of execution of the steps of the machine learning to form a four-quadrant image.
  • the first display module 10 may include: a first instruction receiving sub-module 100 configured to receive a learning circle creation instruction issued by a user; and a display sub-module 102 configured to Under the condition that the instruction receiving submodule receives the learning circle creation instruction issued by the user, the learning circle image is displayed on the first display position of the display screen.
  • the learning circle image at the first display position also provides machine learning status prompt information.
  • the machine learning application providing device also It may include a status information prompt module 50 configured to display the status prompt information of machine learning in a first preset manner in the learning circle image when the user inputs configuration information through the configuration operation portal.
  • the first preset mode may be to display the words "in configuration" in the central area of the learning circle. In one embodiment, it may also be displayed on the side or around the words "in configuration" with a scrolling effect.
  • the bar-shaped or circular pattern in order to increase the dynamic effect of the status prompt information and enhance the user experience.
  • the first display module 10 may further include: a first highlighting sub-module 104 configured to pass the first step of the step by the user When inputting configuration information, highlight at least one of the following: the identification information of the first step in the learning circle image, and the configuration operation entry of the first step.
  • a first highlighting sub-module 104 configured to pass the first step of the step by the user When inputting configuration information, highlight at least one of the following: the identification information of the first step in the learning circle image, and the configuration operation entry of the first step.
  • the second display module 30 may include: a second instruction receiving sub-module 300 configured to receive a learning circle start instruction issued by a user; and an initialization sub-module 302 configured to The second instruction receiving sub-module 300 initializes the machine learning resources under the condition that the learning circle start instruction issued by the user is received, and after the initialization is completed, the learning circle image is displayed on the display screen. The first display position is moved to the second display position to reflect the current state of the machine learning through the change of the position of the learning circle.
  • the learning circle image is displayed in the learning circle image before moving from the first display position to the second display position.
  • the first preset mode may be to display the words "Starting" in the central area of the learning circle. In one embodiment, it may also display a scrolling effect on the side or around the words "Starting". The bar-shaped or circular pattern in order to increase the dynamic effect of the status prompt information and enhance the user experience.
  • the learning circle image After the learning circle image is moved from the first display position to the second display position on the display screen, in the second display position, the learning circle image can all be displayed on the display screen, and the machine is provided in the displayed learning circle image Identification information of the learning step.
  • the learning circle image In some examples, at the second display position, only part of the learning circle image (for example, 1/3, 1/2, or 2/3, etc.) is displayed on the display screen.
  • the learning circle image is about 2/ Part 3 is displayed on the display.
  • the identification information of the machine learning step is displayed in the partial image of the learning circle image displayed on the display screen. Specifically, the identification information of each step of the machine learning is arranged end to end according to the sequence of execution of the steps of the machine learning, and is displayed in the partial image of the learning circle image displayed on the display screen.
  • the learning circle image at the second display position at least part of the execution relationship between the steps may also be provided.
  • the display mode of the execution relationship between the steps in the learning circle image in the second display position may be consistent with the display mode in the learning circle image in the first display position.
  • the second display module 30 may also be configured as: the execution module 40 executes the second step in the machine learning step.
  • the current execution status information of the second step is displayed on the side of the partial image of the learning circle image.
  • the second step here generally refers to any of the multiple steps of the machine learning model.
  • the learning state information can be displayed on the side of the partial image of the learning circle image.
  • the learning status information may include information such as the status prompt of the current learning circle in learning, the optimal effect of the current learning, the time that has been running, the estimated remaining time, the number of running rounds, the trend relationship between the running effect and the behavior data, and so on.
  • Application status information may include information such as the cumulative forecast accuracy rate, the forecast failure rate, the 7-day forecast number, the total forecast number, the number of forecast failures, the application launch time, the forecast accuracy rate and the trend relationship of the behavior data.
  • the execution module 40 may include: a first execution sub-module 400 configured to collect behavior data based on configuration information corresponding to the step of collecting behavior data; and a second execution sub-module 402 configured to Collect feedback data based on the configuration information corresponding to the step of collecting feedback data; the third execution sub-module 404 is configured to collect the behavior data and the feedback based on the configuration information corresponding to the step of training the machine learning model The data is training data, and machine learning training is performed to obtain a machine learning model; the fourth execution sub-module 406 is configured to apply the obtained machine learning model online.
  • the status information prompt module 50 may also be configured to display in the learning circle image in a second preset manner while the execution module 40 sequentially executes each of the machine learning steps.
  • the second preset method may be to display "learning in progress" in the center area of the learning circle. "", refer to Figure 6; when the machine learning model is applied online, the second preset mode may be to display the words "in application” in the central area of the learning circle, refer to Figure 7.
  • a bar or circular pattern with scrolling effect can also be displayed on or around the side of the words "learning” and “applying", so as to increase the dynamic effect of the status prompt information and enhance the user experience.
  • the execution module 40 may further include: a second highlighting submodule 406 , Is configured to highlight the identification information of the step corresponding to the currently executed machine learning step in the learning circle image while the execution module 40 sequentially executes each of the machine learning steps.
  • the way of highlighting can be highlight display, underline display, increase font size display, bold font display, etc.
  • the identification information of the step of training the machine learning model in FIG. 6 and the identification information of the step of online application of the machine learning model in FIG. 7 are respectively highlighted in a highlighted manner.
  • the apparatus for providing machine learning applications may further include: a self-learning module 60 configured to obtain After the machine learning model is applied online, new behavior data is collected based on the configuration information corresponding to the step of collecting behavior data; new feedback data is collected based on the configuration information corresponding to the step of collecting feedback data; Self-learning is performed on the behavior data of and the new feedback data, and the obtained machine learning model is updated or replaced.
  • a self-learning module 60 configured to obtain After the machine learning model is applied online, new behavior data is collected based on the configuration information corresponding to the step of collecting behavior data; new feedback data is collected based on the configuration information corresponding to the step of collecting feedback data; Self-learning is performed on the behavior data of and the new feedback data, and the obtained machine learning model is updated or replaced.
  • the process of machine learning is divided into four steps: behavior, feedback, learning, and application.
  • the four steps continue to loop, and the machine learning model will get better and better as it continues to repeat the four steps in sequence.
  • FIG. 11 is a schematic structural diagram of an embodiment of the electronic device of the present disclosure, which can implement the process of the embodiment shown in FIG. 1 and FIG. It may include: a housing 41, a processor 42, a memory 43, a circuit board 44, and a power supply circuit 45, wherein the circuit board 44 is arranged inside the space enclosed by the housing 41, and the processor 42 and the memory 43 are arranged on the circuit board 44
  • the power supply circuit 45 is configured to supply power to the various circuits or devices of the above electronic equipment;
  • the memory 43 is configured to store executable program code;
  • the processor 42 is configured to execute by reading the executable program code stored in the memory 43
  • the electronic device exists in many forms, including but not limited to: a desktop computer (commonly known as a desktop computer), a notebook computer, or a tablet computer, a server, or an industry-specific computing device.
  • a desktop computer commonly known as a desktop computer
  • notebook computer or a tablet computer
  • server or an industry-specific computing device.
  • the embodiments of the present disclosure also provide a computer-readable storage medium, the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to realize any of the foregoing.
  • the embodiments of the present disclosure also provide an application program that is executed to implement the method for providing a machine learning application described in any of the foregoing embodiments.
  • the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment.
  • each unit/module can be implemented in the same one or more software, the same one or more hardware, or the same one or more software and hardware combinations.
  • the program can be stored in a computer readable storage medium. During execution, it may include the procedures of the above-mentioned method embodiments.
  • the storage medium may be a magnetic disk, an optical disc, a read-only memory (Read-Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种机器学习应用的提供方法、装置、电子设备及存储介质,涉及人工智能技术领域,适用于机器学习产品,便于用户较快掌握机器学习产品的操作。方法包括:在显示屏上的第一显示位置显示学习圈图像(100);接收用户通过配置操作入口输入的、与所述步骤对应的配置信息(102);将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显示的至少部分的学习圈图像中提供有所述机器学习的步骤的标识信息(104);基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用(106)。

Description

一种机器学习应用的提供方法、装置、电子设备及存储介质
本公开要求于2020年4月20日提交中国专利局、申请号为202010315660.4,发明名称为“一种机器学习应用的提供方法、装置、电子设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本公开中。
技术领域
本公开涉及人工智能技术领域,尤其涉及一种机器学习应用的提供方法、装置、电子设备及存储介质。
背景技术
机器学习在各行各业的应用越来越多,但是机器学习还是一个需要很强专业技能的领域。一个机器学习应用从问题定义,到建模,再到模型上线服务,整个流程非常的复杂。
就机器学习产品的用户来讲,特别是一些对机器学习了解有限的新手用户,在使用具体的机器学习产品时,因不够了解机器学习流程,增加了他们的认知和操作门槛,使得对机器学习产品的操作过程难以较快掌握。
发明内容
有鉴于此,本公开实施例提供一种机器学习应用的提供方法、装置、电子设备及存储介质,便于用户较快掌握机器学习产品的操作。
第一方面,本公开实施例提供一种机器学习应用的提供方法,包括:在显示屏上的第一显示位置显示学习圈图像,其中,所述学习圈图像中提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口;接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息;在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显示的至少部分的学习圈图像中提供有所述机器学习的步骤的标识信息;基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
第二方面,本公开实施例还提供一种机器学习应用的提供装置,包括:第一显示模块,被配置为在显示屏上的第一显示位置显示学习圈图像,其中,所述学习圈图像中提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口;配置模块,被配置为接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息;第二显示模块,被配置为在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显 示的至少部分的学习圈图像中提供有所述机器学习的步骤的标识信息;执行模块,被配置为基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
第三方面,本公开实施例还提供一种电子设备,所述电子设备包括:壳体、处理器、存储器、电路板和电源电路,其中,电路板安置在壳体围成的空间内部,处理器和存储器配置在电路板上;电源电路,被配置为为上述电子设备的各个电路或器件供电;存储器被配置为存储可执行程序代码;处理器通过读取存储器中存储的可执行程序代码,被配置为执行前述任一实施例所述的机器学习应用的提供方法。
第四方面,本公开实施例还提供一种计算机可读存储介质,所述计算机可读存储介质存储有一个或者多个程序,所述一个或者多个程序可被一个或者多个处理器执行,以实现前述任一实施例所述的机器学习应用的提供方法。
第五方面,本公开实施例还提供一种应用程序,所述应用程序被执行以实现本公开任一实施例提供的方法。
本公开实施例提供的机器学习应用的提供方法、装置、电子设备及存储介质,基于教育学中的库伯学习圈理论,将机器学习的过程划分为不同的步骤,将不同的步骤标识展现在学习圈图像中。带有机器学习的步骤的标识信息的学习圈图像,先展示在显示屏上的第一显示位置,通过配置操作入口进行相应配置后,在接收到学习圈启动指令的条件下,学习圈图像自显示屏上的第一显示位置移动至第二显示位置,在第二显示位置,学习圈图像中仍然提供有所述机器学习的步骤的标识信息;基于与所述步骤对应的配置信息,可依次执行各所述机器学习的步骤,获取机器学习模型并将获得的机器学习模型上线应用。学习圈图像在从第一显示位置到第二显示位置的变化过程中,始终显示(至少有部分显示)在显示屏显示的界面中,并且在学习圈图像中始终提供有机器学习的步骤的标识信息,这样可以使用户了解学习圈、使用学习圈,并通过学习圈图像的变化过程,更加直观地理解到机器学习的过程,从而便于较快掌握机器学习产品的操作方式。
附图说明
为了更清楚地说明本公开实施例或相关技术中的技术方案,下面将对实施例或相关技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本公开的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其它的附图。
图1为本公开一实施例机器学习应用的提供方法流程示意图;
图2展示了一实施例中显示屏上首次显示学习圈图像的界面;
图3展示当用户看完引导信息的介绍后,学习圈图像从界面中央向左移动的过程;
图4展示了一实施例中显示屏的第一显示位置显示的学习圈图像;
图5展示了当学习圈启动后,学习圈图像自显示屏上的第一显示位置移动至第二显示位置的过程;
图6展示学习圈图像自第一显示位置移动至第二显示位置后的示意图;
图7展示了学习圈上线后的状态,学习圈的状态更新为“服务中”;
图8为本公开另一实施例机器学习应用的提供方法流程示意图;。
图9为本公开一实施例机器学习应用的提供装置的方框图;
图10为本公开另一实施例机器学习应用的提供装置的方框图;
图11为本公开电子设备一个实施例的结构示意图。
具体实施方式
下面结合附图对本公开实施例进行详细描述。应当明确,所描述的实施例仅仅是本公开一部分实施例,而不是全部的实施例。基于本公开中的实施例,本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其它实施例,都属于本公开保护的范围。
本公开实施例提供一种机器学习应用的提供方法及装置,将机器学习的步骤以学习圈的形式展现给用户,便于用户较快掌握机器学习产品的操作。
本公开实施例中所提及的机器学习产品,可以是一种人工智能AI应用开发工具,用户可以利用该产品构建机器学习模型,可以直接将所构建的机器学习模型提供给用户,亦可按照用户需求将模型直接部署上线应用。
图1为本公开一实施例机器学习应用的提供方法流程示意图;参看图1,本实施例机器学习应用的提供方法,包括步骤S100至步骤S106:
S100、在显示屏上的第一显示位置显示学习圈图像。
在显示屏上的第一显示位置显示的学习圈图像中,提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口。
本公开中的学习圈,是基于教育学中的库伯学习圈理论,将机器学习的过程参考人的学习的过程,定义为学习圈,具体是将机器学习,即机器学习模型(也可称为人工智能模型)训练的过程划分为不同的步骤,这些步骤在不停的循环的机器学习形态或样式。
学习圈图像,是基于上述学习圈的原理呈现的学习圈的图像,也可称为学习圈图形、或学习圈样式等。具体的,学习圈图像中可以包括机器学习的各个步骤的标识信息,还可以体现步骤之间的顺序和循环等执行关系。
学习圈图像的形态可以有多种。为了能让用户更直观地理解、更方便地操作,在一个例子中,学习圈图像的形态可具体呈现为圈的形态,具体地,可将机器学习的各 步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,形成圈的图像(或称为圈的形态)。机器学习的每个步骤的步骤标识信息在圈的图像上可呈扇形布置。
学习圈图像的形态,不局限于圈的形态,在另一个例子中,可将机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,对应布置在第一象限至第四象限中,形成四象限的图像。在又一个例子中,还可将机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,布置成曲线形或弧形,等等。
学习圈图像中提供的机器学习的步骤,可包括:收集行为数据的步骤、收集反馈数据的步骤、训练机器学习模型的步骤、应用机器学习模型的步骤。
其中的行为数据,是指历史行为数据,每一条行为数据对应一个历史“行为”。比如在篮球比赛中,投篮就是这个“行为”,一条行为数据就代表了过去某场篮球比赛中的一次历史投篮。那些可以用来描述投篮这个行为的数据,比如“投篮队员、姿势、位置、经度、维度、对手、助攻、与篮筐距离、剩余时间、对战球队……”都属于行为数据里所包含的信息。“科比、跳投、三分球位置、33.7693、-118.1798、帕特森(对手)、佩顿(助攻)、28英尺、1秒、开拓者(对战球队)……”这就是一条行为数据,对应科比在04年常规赛收官战上,最后1秒的投篮。
反馈数据,是对历史行为的“反馈”,代表着行为数据所产生的结果。在篮球比赛中,投球所产生的结果,就是命中或不命中。投球进了,反馈数据就是“命中”。
训练机器学习模型,是以收集的大量的行为数据和反馈数量作为训练数据,进行有监督机器学习,以得到机器学习模型。
应用机器学习模型,是将经过训练得到的机器学习模型应用于业务系统中,进行预测,以获得预测结果。
在学习圈图像中,标识信息可以通过图形加文字等多种方式呈现。机器学习的步骤的标识信息,具体可包括步骤名称、步骤图标中的至少一项。在一个例子中,所述标识信息是机器学习的步骤名称,如中文名称或英文名称等;在另一个例子中,所述标识信息是机器学习的步骤图标,如箭头、符号等,不同的步骤可以用不同形状或不同颜色的图标区分表示;在又一个例子中,所述标识信息是机器学习的步骤名称和步骤图标的组合。
就机器学习的步骤名称而言,在一个例子中,收集行为数据的步骤名称为“行为”,收集反馈数据的步骤名称为“反馈”、训练机器学习模型的步骤名称为“学习”、应用机器学习模型的步骤名称为“应用”。
显示学习圈图像的显示屏,可以是桌面计算机(俗称台式电脑)、笔记本电脑、或平板电脑的显示屏,也可以是服务器的显示屏,还可以是行业专用计算设备的显示屏。 显示屏可以是触摸屏或非触摸屏。
显示屏上的第一显示位置,可以是显示屏的中央显示区域,也可以是偏离中央显示区域一定距离的其它显示区域。
学习圈图像在显示屏上的第一显示位置的显示,可以是由用户触发的。在一个例子中,用户可通过点击显示屏上所显示的操作界面中的“创建学习圈”按钮,来创建学习圈图像,即启动学习圈图像在显示屏上的第一显示位置的显示。具体地,所述在显示屏上的第一显示位置显示学习圈图像可包括:在用户下发(比如通过点击“创建学习圈”按钮下发)了学习圈创建指令的条件下,电子设备在接收到用户下发的学习圈创建指令后,在电子设备的显示屏的第一显示位置显示学习圈图像。
在电子设备的显示屏上首次显示学习圈图像(用户在首次进入显示界面)时,在显示屏所显示的界面中可自动播放有关于学习圈的引导信息,引导信息可包括学习圈的流程和基本理念等信息,便于用户较快理解和掌握学习圈的操作方法。
在电子设备的显示屏上首次显示学习圈图像时,优先选择将学习圈图像在显示屏上的中央显示区域显示,这样可将学习圈图像作为界面中的核心焦点,便于用户更好地理解和记住学习圈图像。图2展示了一实施例中显示屏上首次显示学习圈图像的界面。
在引导信息自动播放完毕,可将学习圈图像自首次显示位置向预定方向(如向左)移动,直至将学习圈图像移动至显示屏的第一显示位置。图3展示当用户看完引导信息的介绍后,学习圈图像从界面中央向左移动的过程。
图4展示了一实施例中显示屏的第一显示位置显示的学习圈图像,参看图4,第一显示位置的学习圈图像中,除了提供有机器学习的步骤的标识信息之外,还可提供有机器学习的各步骤之间的执行关系,以便用户能够清楚了解各步骤之间的先后执行顺序和循环方式。在一例子中,可以用箭头指向的方式表示相邻步骤之间的先后执行关系,处于箭头起始点的步骤,先于处于箭头终点的步骤执行。
在另一个例子中,可以用数字大小表示各步骤之间的执行关系,小数字对应的步骤,先于大数字对应的步骤执行,比如数字“1”对应的收集行为数据的步骤,先于数字“2”对应的收集反馈数据的步骤执行。
S102、接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息。
参看图4,学习圈图像的侧部显示有与机器学习步骤对应的配置操作入口。通过配置操作入口,可对执行各步骤所需的数据或参数进行配置。
通过与收集行为数据的步骤相对应的行为数据配置操作入口,可选择收集哪些行为数据。
通过与收集反馈数据的步骤相对应的反馈数据配置操作入口,可对行为数据所产 生的结果(即反馈数据)的收集进行配置。由于反馈数据与行为数据相对应,可以在通过所述行为数据配置操作入口,对行为数据进行配置的同时,自动对反馈数据进行配置,亦即可以不用单独设置反馈数据配置操作入口。
通过与训练机器学习模型的步骤相对应的学习配置操作入口,可对机器学习的频率、机器学习所用训练数据的时间范围、训练数据与测试数据的拆分比例,机器学习的准确率、机器学习的时间等进行配置。
通过与应用机器学习模型的步骤相对应的应用配置操作入口,可对是否自动上线机器学习模型,或是否自动启动机器学习模型的自学习等进行配置。
为了便于向用户提供关于配置操作的引导,在一个例子中,在用户通过所述步骤中的第一步骤的配置操作入口输入配置信息时,可突出显示以下至少一项:学习圈图像中所述第一步骤的标识信息、第一步骤的配置操作入口。突出显示的方式可以是高亮显示、加下划线显示、加大字号显示、加粗字体显示等。第一步骤的标识信息的突出显示方式,与第一步骤的配置操作入口的突出显示方式,可以相同,也可以不同。这里的第一步骤是泛指机器学习模型的多个步骤中的任一步骤。图4中收集行为数据的步骤的标识信息,以高亮显示的方式突出显示。
为使用户了解学习圈当前所处的状态,在一个例子中,第一显示位置的学习圈图像中还提供有机器学习的状态提示信息,具体地,在用户通过所述配置操作入口输入配置信息时,在学习圈图像中以第一预设方式显示机器学习的状态提示信息。在一具体例子中,第一预设方式可以是在学习圈的中心区域显示“配置中”的字样,在一实施方式中,还可在“配置中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态效果,增强用户体验。
图4展示了学习圈的配置页面,在圈的中央显示学习圈当前的状态----配置中。同时,在上方的导航栏提示用户当前学习圈的名称和全局操作的按钮。用户可以启动或删除当前的学习圈;在右侧展示四个步骤卡片可以让用户调整学习圈四个步骤的具体参数。当用户确定好配置后,可以点击右上角导航的启动学习圈按钮,让学习圈开始机器学习的过程。
S104、将学习圈图像自显示屏上的第一显示位置移动至第二显示位置。
在通过所述配置操作入口进行配置之后,就可以启动学习圈,以执行机器学习的步骤。
在一个例子中,在接收到用户下发的学习圈启动指令的条件下,则可将学习圈图像自显示屏上的第一显示位置移动至第二显示位置,在第二显示位置展示学习圈启动后的状态。具体地,在接收到用户下发的学习圈启动指令的条件下,初始化机器学习资源,并在所述初始化完成后,将学习圈图像自显示屏上的第一显示位置移动至第二 显示位置,以通过学习圈的位置的变化体现出机器学习当前所处的状态。
为使用户了解学习圈当前所处的状态,在一个例子中,在接收到用户下发的学习圈启动指令后,学习圈图像从第一显示位置移动到第二显示位置之前,在学习圈图像中以第一预设方式显示机器学习的状态提示信息。在一具体例子中,第一预设方式可以是在学习圈的中心区域显示“启动中”的字样,在一实施方式中,还可在“启动中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态效果,增强用户体验。
图5展示了当学习圈启动后,学习圈图像自显示屏上的第一显示位置移动至第二显示位置的过程。图6展示学习圈图像自第一显示位置移动至第二显示位置后的示意图。
当学习圈图像自显示屏上的第一显示位置移动至第二显示位置后,在第二显示位置,学习圈图像可以全部显示在显示屏中,在显示的学习圈图像中提供有所述机器学习的步骤的标识信息。在一些例子中,在第二显示位置,学习圈图像仅有部分(比如有1/3、1/2或2/3等)显示在显示屏中,图6所示的实施例中,学习圈图像大约有2/3的部分显示在显示屏中。机器学习的步骤的标识信息,显示在显示屏中所显示的所述学习圈图像的部分图像中。具体地,机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,显示在显示屏中所显示的所述学习圈图像的部分图像中。
在第二显示位置的学习圈图像中,还可提供有至少部分步骤之间的执行关系。步骤之间的执行关系在第二显示位置的学习圈图像中的显示方式,可与在第一显示位置的学习圈图像中的显示方式一致。
S106、基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
在将学习圈图像自显示屏上的第一显示位置移动至第二显示位置后,即可基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。具体地,基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用,可包括:基于与收集行为数据的步骤对应的配置信息,收集行为数据;基于与收集反馈数据的步骤对应的配置信息,收集反馈数据;基于与训练机器学习模型的步骤对应的配置信息,以收集的所述行为数据和所述反馈数据为训练数据,进行机器学习训练,获得机器学习模型;将获得的所述机器学习模型上线应用。
图6中,学习圈图像自第一显示位置移动至第二显示位置后,学习圈的状态会更新为“学习中”,同时,在顶部,学习圈的全局操作按钮有所变化;此时用户可以选择 申请将学习的机器学习模型上线或停止学习圈的学习;在右侧,展示了此时学习圈学习步骤的状态。用户可以点击学习圈外围的四个步骤来切换查看四个步骤的状态。当学习圈学习有了结果,产生了机器学习模型后,用户可以点击右上角的申请上线按钮,将学习圈的机器学习模型上线服务以应用。此时用户依然可以通过底部的按钮切换回配置页面查看学习圈的配置。
图7展示了学习圈上线后的状态,学习圈的状态更新为“服务中”。同时展示了机器学习模型在线上的表现。在显示界面的顶部,学习圈的全局操作按钮有所变化;此时用户可以选择申请将学习的机器学习模型下线或手动上线一个新的机器学习模型。
本公开实施例中,基于教育学中的库伯学习圈理论,将机器学习模型训练的过程划分为不同的步骤,将不同的步骤展现在学习圈图像中。带有机器学习的步骤的标识信息的学习圈图像,先展示在显示屏上的第一显示位置;通过配置操作入口进行相应配置后,在接收到学习圈启动指令的条件下,学习圈图像自显示屏上的第一显示位置移动至第二显示位置,在第二显示位置,学习圈图像中仍然提供有所述机器学习的步骤的标识信息;基于与所述步骤对应的配置信息,可依次执行各所述机器学习的步骤,获取机器学习模型并将获得的机器学习模型上线应用。学习圈图像在从第一显示位置到第二显示位置的变化过程中,始终显示(至少有部分显示)在显示屏显示的界面中,并且在学习圈图像中始终提供有机器学习的步骤的标识信息,这样可以使用户了解学习圈、使用学习圈,并通过学习圈图像的变化过程,更加直观地理解到机器学习的过程,从而便于较快掌握机器学习产品的操作方式。
为增强用户对机器学习的步骤执行顺序的认识,便于用于了解机器学习当前处于哪个步骤,在一个例子中,在依次执行各所述机器学习的步骤的同时,还可将学习圈图像中与当前所执行的机器学习的步骤相对应的步骤的标识信息,突出显示。突出显示的方式可以是高亮显示、加下划线显示、加大字号显示、加粗字体显示等。图6中训练机器学习模型的步骤的标识信息,以及图7中机器学习模型上线应用的步骤的标识信息,分别以高亮显示的方式突出显示。
机器学习的各步骤顺序执行的过程中,机器学习的步骤相对应的步骤的标识信息,也相应地顺序突出显示,使得用户更容易了解和掌握学习圈。
为使用户了解学习圈当前所处的状态,在一个例子中,第二显示位置的学习圈图像中还提供有机器学习的状态提示信息,具体地,在依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息。在一具体例子中,当执行收集行为数据的步骤时、执行收集反馈数据的步骤时、以及执行训练机器学习模型的步骤时,第二预设方式可以是在学习圈的中心区域显示“学习中”的字样,参看图6所示;在机器学习模型上线应用时,第二预设方 式可以是在学习圈的中心区域显示“应用中”的字样,参看图7所示。在一实施方式中,还可在“学习中”和“应用中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态效果,增强用户体验。
为了使用户直观地了解当前执行步骤的一些执行状态信息,在一个例子中,可在执行所述机器学习的步骤中的第二步骤时,在所述学习圈图像的部分图像的侧部,显示所述第二步骤当前的执行状态信息。这里的第二步骤是泛指机器学习模型的多个步骤中的任一步骤。
比如,参看图6所示,在执行训练机器学习模型的步骤时,即学习圈处于“学习状态”时,可在学习圈图像的部分图像的侧部,显示学习状态信息。学习状态信息可包括当前学习圈处于学习中的状态提示、当前学习的最优效果、已经运行的时间、预计剩余时间、已经运行轮数、运行效果与行为数据的走势关系等信息。
又比如,参看图7所示,在执行应用机器学习模型的步骤时,即学习圈处于“应用状态”时,可在学习圈图像的部分图像的侧部,显示应用状态信息。应用状态信息可包括累计预测准确率、预测失效率、7日预测数量、预测总数量、预测失败数、应用上线时间、预测正确率与行为数据的走势关系等信息。
图8为本公开另一实施例机器学习应用的提供方法流程示意图。
在将获得的所述机器学习模型上线应用之后,还可开启机器学习模型的自学习,通过自学习,可对当前应用的机器学习模型进行优化或替换。具体地,参看图8所示,在图1所示实施例的基础上,在一实施方式中,在将获得的所述机器学习模型上线应用之后,所述方法还可包括步骤:
S108、基于与收集行为数据的步骤对应的配置信息,收集新的行为数据。
S110、基于与收集反馈数据的步骤对应的配置信息,收集新的反馈数据。
S112、基于所述新的行为数据和所述新的反馈数据,进行自学习,对已获得的所述机器学习模型进行更新或替换。
本实施例中,基于教育学中的库伯学习圈理论,将机器学习的过程划分为行为、反馈、学习、应用四个步骤。四个步骤不断循环,机器学习模型在不断依次重复四个步骤的过程中会变得越来越好。
图9为本公开一实施例机器学习应用的提供装置的方框图,参看图9,本实施例机器学习应用的提供装置,包括:第一显示模块10、配置模块20、第二显示模块30和执行模块40;其中,第一显示模块10,被配置为在显示屏上的第一显示位置显示学习圈图像,其中,所述学习圈图像中提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口。
本实施例中,关于学习圈、学习圈图像、机器学习的步骤、及机器学习的步骤的 标识信息,可参看图1所示方法实施例的相关解释,在此不再赘述。
配置模块20,被配置为接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息。
参看图4,学习圈图像的侧部显示有与机器学习步骤对应的配置操作入口。通过配置操作入口,可对执行各步骤所需的数据或参数进行配置。
通过与收集行为数据的步骤相对应的行为数据配置操作入口,可选择收集哪些行为数据。
通过与收集反馈数据的步骤相对应的反馈数据配置操作入口,可对行为数据所产生的结果(即反馈数据)的收集进行配置。由于反馈数据与行为数据相对应,可以在通过所述行为数据配置操作入口,对行为数据进行配置的同时,自动对反馈数据进行配置,亦即可以不用单独设置反馈数据配置操作入口。
通过与训练机器学习模型的步骤相对应的学习配置操作入口,可对机器学习的频率、机器学习所用训练数据的时间范围、训练数据与测试数据的拆分比例,机器学习的准确率、机器学习的时间等进行配置。
通过与应用机器学习模型的步骤相对应的应用配置操作入口,可对是否自动上线机器学习模型,或是否自动启动机器学习模型的自学习等进行配置。
第二显示模块30,被配置为在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显示的至少部分的学习圈图像中提供有所述机器学习的步骤的标识信息。
执行模块40,被配置为基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
在将学习圈图像自显示屏上的第一显示位置移动至第二显示位置后,即可基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
本实施例的装置,可以用于执行图1所示方法实施例的技术方案,其实现原理和技术效果类似,此处不再赘述。
在一实施例中,第一显示位置的学习圈图像中还提供有所述步骤之间的执行关系,以便用户能够清楚了解各步骤之间的先后执行顺序。在一例子中,可以用箭头指向的方式表示相邻步骤之间的先后执行关系,处于箭头起始点的步骤,先于处于箭头终点的步骤执行。在另一个例子中,可以用数字大小表示各步骤之间的执行关系,小数字对应的步骤,先于大数字对应的步骤执行,比如数字“1”对应的收集行为数据的步骤,先于数字“2”对应的收集反馈数据的步骤执行。
在一实施例中,所述第一显示模块10,被配置为:在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,形成圈的图像;或者,在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,对应布置在第一象限至第四象限中,形成四象限的图像。
参看图10,在一实施例中,所述第一显示模块10可包括:第一指令接收子模块100,被配置为接收到用户下发的学习圈创建指令;显示子模块102,被配置为在指令接收子模块接收到用户下发的学习圈创建指令的条件下,则在所述显示屏的第一显示位置显示学习圈图像。
为使用户了解学习圈当前所处的状态,在一个例子中,第一显示位置的学习圈图像中还提供有机器学习的状态提示信息,具体地,所述的机器学习应用的提供装置,还可包括状态信息提示模块50,被配置为:在用户通过所述配置操作入口输入配置信息时,在所述学习圈图像中以第一预设方式显示机器学习的状态提示信息。在一具体例子中,第一预设方式可以是在学习圈的中心区域显示“配置中”的字样,在一实施方式中,还可在“配置中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态效果,增强用户体验。
为了便于向用户提供关于配置操作的引导,在一个例子中,所述第一显示模块10,还可包括:第一突出显示子模块104,被配置为在用户通过所述步骤中的第一步骤的配置操作入口输入配置信息时,突出显示以下至少一项:所述学习圈图像中所述第一步骤的标识信息、第一步骤的配置操作入口。第一步骤的标识信息、第一步骤的配置操作入口中至少一项的突出显示方式,可参看图1所示方法实施例中的相关描述,在此不再赘述。
参看图10,在一实施例中,第二显示模块30,可包括:第二指令接收子模块300,被配置为接收到用户下发的学习圈启动指令;初始化子模块302,被配置为在所述第二指令接收子模块300接收到用户下发的学习圈启动指令的条件下,初始化机器学习资源,并在所述初始化完成后,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,以通过学习圈的位置的变化体现出机器学习当前所处的状态。
使用户了解学习圈当前所处的状态,在一个例子中,在接收到用户下发的学习圈启动指令后,学习圈图像从第一显示位置移动到第二显示位置之前,在学习圈图像中以第一预设方式显示机器学习的状态提示信息。在一具体例子中,第一预设方式可以是在学习圈的中心区域显示“启动中”的字样,在一实施方式中,还可在“启动中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态 效果,增强用户体验。
当学习圈图像自显示屏上的第一显示位置移动至第二显示位置后,在第二显示位置,学习圈图像可以全部显示在显示屏中,在显示的学习圈图像中提供有所述机器学习的步骤的标识信息。在一些例子中,在第二显示位置,学习圈图像仅有部分(比如有1/3、1/2或2/3等)显示在显示屏中,图6中,学习圈图像大约有2/3的部分显示在显示屏中。机器学习的步骤的标识信息,显示在显示屏中所显示的所述学习圈图像的部分图像中。具体地,机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,显示在显示屏中所显示的所述学习圈图像的部分图像中。
在第二显示位置的学习圈图像中,还可提供有至少部分步骤之间的执行关系。步骤之间的执行关系在第二显示位置的学习圈图像中的显示方式,可与在第一显示位置的学习圈图像中的显示方式一致。
为了使用户直观地了解当前执行步骤的一些执行状态信息,在一个例子中,所述第二显示模块30还可被配置为:在所述执行模块40执行所述机器学习的步骤中的第二步骤时,在所述学习圈图像的部分图像的侧部,显示所述第二步骤当前的执行状态信息。这里的第二步骤是泛指机器学习模型的多个步骤中的任一步骤。
比如,参看图6所示,在执行训练机器学习模型的步骤时,即学习圈处于“学习状态”时,可在学习圈图像的部分图像的侧部,显示学习状态信息。学习状态信息可包括当前学习圈处于学习中的状态提示、当前学习的最优效果、已经运行的时间、预计剩余时间、已经运行轮数、运行效果与行为数据的走势关系等信息。
又比如,参看图7所示,在执行应用机器学习模型的步骤时,即学习圈处于“应用状态”时,可在学习圈图像的部分图像的侧部,显示应用状态信息。应用状态信息可包括累计预测准确率、预测失效率、7日预测数量、预测总数量、预测失败数、应用上线时间、预测正确率与行为数据的走势关系等信息。
在一个例子中,所述执行模块40,可包括:第一执行子模块400,被配置为基于与收集行为数据的步骤对应的配置信息,收集行为数据;第二执行子模块402,被配置为基于与收集反馈数据的步骤对应的配置信息,收集反馈数据;第三执行子模块404,被配置为基于与训练机器学习模型的步骤对应的配置信息,以收集的所述行为数据和所述反馈数据为训练数据,进行机器学习训练,获得机器学习模型;第四执行子模块406,被配置为将获得的所述机器学习模型上线应用。
在一个例子中,所示状态信息提示模块50,还可被配置为在所述执行模块40依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息。在一具体例子中,当执行收集行为数据的步骤时、执行收集反馈数据的步骤时、以及执行训练机器学习模型的步骤时,第二预设方 式可以是在学习圈的中心区域显示“学习中”的字样,参看图6所示;在机器学习模型上线应用时,第二预设方式可以是在学习圈的中心区域显示“应用中”的字样,参看图7所示。在一实施方式中,还可在“学习中”和“应用中”字样的侧部或周围显示具有滚动效果的条形或环形图案,以增加状态提示信息的动态效果,增强用户体验。
参看图10,为增强用户对机器学习的步骤执行顺序的认识,便于用于了解机器学习当前处于哪个步骤,在一个例子中,所述执行模块40,还可包括:第二突出显示子模块406,被配置为在所述执行模块40依次执行各所述机器学习的步骤的同时,将所述学习圈图像中与当前所执行的机器学习的步骤相对应的步骤的标识信息,突出显示。突出显示的方式可以是高亮显示、加下划线显示、加大字号显示、加粗字体显示等。图6中训练机器学习模型的步骤的标识信息,以及图7中机器学习模型上线应用的步骤的标识信息,分别以高亮显示的方式突出显示。
在将获得的所述机器学习模型上线应用之后,还可开启机器学习模型的自学习,通过自学习,可对当前应用的机器学习模型进行优化或替换。具体地,参看图10所示,在图9所示实施例的基础上,所述的机器学习应用的提供装置,还可包括:自学习模块60,被配置为在所述执行模块40将获得的所述机器学习模型上线应用之后,基于与收集行为数据的步骤对应的配置信息,收集新的行为数据;基于与收集反馈数据的步骤对应的配置信息,收集新的反馈数据;基于所述新的行为数据和所述新的反馈数据,进行自学习,对已获得的所述机器学习模型进行更新或替换。
本实施例中,基于教育学中的库伯学习圈理论,将机器学习的过程划分为行为、反馈、学习、应用四个步骤。四个步骤不断循环,机器学习模型在不断依次重复四个步骤的过程中会变得越来越好。
本公开实施例还提供一种电子设备,图11为本公开电子设备一个实施例的结构示意图,可以实现本公开图1及图8所示实施例的流程,如图11所示,上述电子设备可以包括:壳体41、处理器42、存储器43、电路板44和电源电路45,其中,电路板44安置在壳体41围成的空间内部,处理器42和存储器43配置在电路板44上;电源电路45,被配置为为上述电子设备的各个电路或器件供电;存储器43被配置为存储可执行程序代码;处理器42通过读取存储器43中存储的可执行程序代码,被配置为执行前述任一实施例所述的机器学习应用的提供方法。
处理器42对上述步骤的具体执行过程以及处理器42通过运行可执行程序代码来进一步执行的步骤,可以参见本公开图1及图8所示实施例的描述,在此不再赘述。
该电子设备以多种形式存在,包括但不限于:桌面计算机(俗称台式电脑)、笔记本电脑、或平板电脑,也可以是服务器,还可以是行业专用计算设备等。
本公开实施例还提供一种计算机可读存储介质,所述计算机可读存储介质存储有 一个或者多个程序,所述一个或者多个程序可被一个或者多个处理器执行,以实现前述任一实施例所述的机器学习应用的提供方法。
本公开的实施例还提供一种应用程序,所述应用程序被执行以实现前述任一实施例所述的机器学习应用的提供方法。
需要说明的是,在本文中,诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来,而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
本说明书中的各个实施例均采用相关的方式描述,各个实施例之间相同相似的部分互相参见即可,每个实施例重点说明的都是与其他实施例的不同之处。
尤其,对于装置实施例而言,由于其基本相似于方法实施例,所以描述的比较简单,相关之处参见方法实施例的部分说明即可。
为了描述的方便,描述以上装置是以功能分为各种单元/模块分别描述。当然,在实施本公开时可以把各单元/模块的功能在同一个或多个软件、同一个或多个硬件、同一个或多个软件和硬件的结合中实现。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程,是可以通过计算机程序来指令相关的硬件来完成,所述的程序可存储于一计算机可读取存储介质中,该程序在执行时,可包括如上述各方法的实施例的流程。其中,所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory,ROM)或随机存储记忆体(Random Access Memory,RAM)等。
以上所述,仅为本公开的具体实施方式,但本公开的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本公开揭露的技术范围内,可轻易想到的变化或替换,都应涵盖在本公开的保护范围之内。因此,本公开的保护范围应以权利要求的保护范围为准。

Claims (32)

  1. 一种机器学习应用的提供方法,包括:
    在显示屏上的第一显示位置显示学习圈图像,其中,所述学习圈图像中提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口;
    接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息;
    在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显示的至少部分的学习圈图像中提供有所述机器学习的步骤的标识信息;
    基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
  2. 根据权利要求1所述的机器学习应用的提供方法,其中,所述方法还包括以下至少一项:
    所述第一显示位置的学习圈图像中还提供有所述步骤之间的执行关系,
    所述第二显示位置的学习圈图像中还提供有至少部分步骤之间的执行关系。
  3. 根据权利要求1所述的机器学习应用的提供方法,其中,
    所述机器学习的步骤包括:收集行为数据的步骤、收集反馈数据的步骤、训练机器学习模型的步骤、应用机器学习模型的步骤;
    所述机器学习的步骤的标识信息包括步骤名称、步骤图标中的至少一项。
  4. 根据权利要求3所述的机器学习应用的提供方法,其中,所述收集行为数据的步骤名称、收集反馈数据的步骤名称、训练机器学习模型的步骤名称、应用机器学习模型的步骤名称,依次为行为、反馈、学习、应用。
  5. 根据权利要求1所述的机器学习应用的提供方法,其中,所述在显示屏上的第一显示位置显示学习圈图像,包括:
    在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,形成圈的图像;或者,
    在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,对应布置在第一象限至第四象限中,形成四象限的图像。
  6. 根据权利要求1所述的机器学习应用的提供方法,其中,所述方法还包括:
    在用户通过所述步骤中的第一步骤的配置操作入口输入配置信息时,突出显示以下至少一项:所述学习圈图像中所述第一步骤的标识信息,第一步骤的配置操作入口。
  7. 根据权利要求1所述的机器学习应用的提供方法,其中,所述在显示屏上的第一显示位置显示学习圈图像包括:
    在接收到用户下发的学习圈创建指令的条件下,在所述显示屏的第一显示位置显示学习圈图像。
  8. 根据权利要求1所述的机器学习应用的提供方法,其中,所述在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,包括:
    在接收到用户下发的学习圈启动指令的条件下,初始化机器学习资源,并在所述初始化完成后,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置。
  9. 根据权利要求1所述的机器学习应用的提供方法,其中,
    在所述第二显示位置,所述学习圈图像的部分图像显示在所述显示屏中;其中,所述机器学习的步骤的标识信息,显示在显示屏中所显示的所述学习圈图像的部分图像中。
  10. 根据权利要求9所述的机器学习应用的提供方法,其中,所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,显示在显示屏中所显示的所述学习圈图像的部分图像中。
  11. 根据权利要求3所述的机器学习应用的提供方法,其中,所述基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用,包括:
    基于与收集行为数据的步骤对应的配置信息,收集行为数据;
    基于与收集反馈数据的步骤对应的配置信息,收集反馈数据;
    基于与训练机器学习模型的步骤对应的配置信息,以收集的所述行为数据和所述反馈数据为训练数据,进行机器学习训练,获得机器学习模型;
    将获得的所述机器学习模型上线应用。
  12. 根据权利要求1或11所述的机器学习应用的提供方法,其中,在依次执行各所述机器学习的步骤的同时,所述方法还包括:
    将所述学习圈图像中与当前所执行的机器学习的步骤相对应的步骤的标识信息,突出显示。
  13. 根据权利要求11所述的机器学习应用的提供方法,其中,在将获得的所述机器学习模型上线应用之后,所述方法还包括:
    基于与收集行为数据的步骤对应的配置信息,收集新的行为数据;
    基于与收集反馈数据的步骤对应的配置信息,收集新的反馈数据;
    基于所述新的行为数据和所述新的反馈数据,进行自学习,对已获得的所述机器学习模型进行更新或替换。
  14. 根据权利要求1所述的机器学习应用的提供方法,其中,所述方法还包括:
    在执行所述机器学习的步骤中的第二步骤时,在所述学习圈图像的部分图像的侧部,显示所述第二步骤当前的执行状态信息。
  15. 根据权利要求1所述的机器学习应用的提供方法,其中,
    所述第一显示位置的学习圈图像中还提供有机器学习的状态提示信息,所述方法还包括:在以下至少一种情况下,在所述学习圈图像中以第一预设方式显示机器学习的状态提示信息:在用户通过所述配置操作入口输入配置信息时;在接收到用户下发的学习圈启动指令后,所述学习圈图像从第一显示位置移动到第二显示位置之前;
    所述第二显示位置的学习圈图像中还提供有机器学习的状态提示信息,所述方法还包括:在依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息;
    所述第一显示位置的学习圈图像中还提供有机器学习的状态提示信息,所述方法还包括:在以下至少一种情况下,在所述学习圈图像中以第一预设方式显示机器学习的状态提示信息:在用户通过所述配置操作入口输入配置信息时;在接收到用户下发的学习圈启动指令后,所述学习圈图像从第一显示位置移动到第二显示位置之前;
    所述第二显示位置的学习圈图像中还提供有机器学习的状态提示信息,所述方法还包括:在依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息。
  16. 一种机器学习应用的提供装置,包括:
    第一显示模块,被配置为在显示屏上的第一显示位置显示学习圈图像,其中,所述学习圈图像中提供有机器学习的步骤的标识信息,且所述学习圈图像的侧部显示有与所述步骤对应的配置操作入口;
    配置模块,被配置为接收用户通过所述配置操作入口输入的、与所述步骤对应的配置信息;
    第二显示模块,被配置为在接收到用户下发的学习圈启动指令的条件下,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置,在所述第二显示位置,所述学习圈图像至少有部分显示在所述显示屏中,且在显示的至少部分的学习圈 图像中提供有所述机器学习的步骤的标识信息;
    执行模块,被配置为基于与所述步骤对应的配置信息,依次执行各所述机器学习的步骤,以获取并提供所述机器学习应用。
  17. 根据权利要求16所述的机器学习应用的提供装置,其中,所述装置包括以下至少一项:
    所述第一显示位置的学习圈图像中还提供有所述步骤之间的执行关系,所述第二显示位置的学习圈图像中还提供有至少部分步骤之间的执行关系。
  18. 根据权利要求16所述的机器学习应用的提供装置,其中,
    所述机器学习的步骤包括:收集行为数据的步骤、收集反馈数据的步骤、训练机器学习模型的步骤、应用机器学习模型的步骤;
    所述机器学习的步骤的标识信息包括步骤名称、步骤图标中的至少一项。
  19. 根据权利要求18所述的机器学习应用的提供装置,其中,所述收集行为数据的步骤名称、收集反馈数据的步骤名称、训练机器学习模型的步骤名称、应用机器学习模型的步骤名称,依次为行为、反馈、学习、应用。
  20. 根据权利要求16所述的机器学习应用的提供装置,其中,所述第一显示模块,被配置为:
    在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,形成圈的图像;或者,
    在显示屏上的第一显示位置显示机器学习的各步骤的标识信息,并将所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,对应布置在第一象限至第四象限中,形成四象限的图像。
  21. 根据权利要求16所述的机器学习应用的提供装置,其中,所述第一显示模块包括:
    第一突出显示子模块,被配置为在用户通过所述步骤中的第一步骤的配置操作入口输入配置信息时,突出显示以下至少一项:所述学习圈图像中所述第一步骤的标识信息,第一步骤的配置操作入口。
  22. 根据权利要求16所述的机器学习应用的提供装置,其中,所述第一显示模块包括:
    第一指令接收子模块,被配置为接收到用户下发的学习圈创建指令;
    显示子模块,被配置为在指令接收子模块接收到用户下发的学习圈创建指令的条件下,则在所述显示屏的第一显示位置显示学习圈图像。
  23. 根据权利要求16所述的机器学习应用的提供装置,其中,所述第二显示模块, 包括:
    第二指令接收子模块,被配置为接收到用户下发的学习圈启动指令;
    初始化子模块,被配置为在所述第二指令接收子模块接收到用户下发的学习圈启动指令的条件下,初始化机器学习资源,并在所述初始化完成后,将所述学习圈图像自所述显示屏上的第一显示位置移动至第二显示位置。
  24. 根据权利要求16所述的机器学习应用的提供装置,其中,
    在所述第二显示位置,所述学习圈图像的部分图像显示在所述显示屏中;其中,所述机器学习的步骤的标识信息,显示在显示屏中所显示的所述学习圈图像的部分图像中。
  25. 根据权利要求24所述的机器学习应用的提供装置,其中,所述机器学习的各步骤的标识信息,按照机器学习的各步骤之间的先后执行顺序,首尾排列,显示在显示屏中所显示的所述学习圈图像的部分图像中。
  26. 根据权利要求18所述的机器学习应用的提供装置,其中,所述执行模块,包括:
    第一执行子模块,被配置为基于与收集行为数据的步骤对应的配置信息,收集行为数据;
    第二执行子模块,被配置为基于与收集反馈数据的步骤对应的配置信息,收集反馈数据;
    第三执行子模块,被配置为基于与训练机器学习模型的步骤对应的配置信息,以收集的所述行为数据和所述反馈数据为训练数据,进行机器学习训练,获得机器学习模型;
    第四执行子模块,被配置为将获得的所述机器学习模型上线应用。
  27. 根据权利要求16或26所述的机器学习应用的提供装置,其中,所述执行模块包括:
    第二突出显示子模块,被配置为在所述执行模块依次执行各所述机器学习的步骤的同时,将所述学习圈图像中与当前所执行的机器学习的步骤相对应的步骤的标识信息,突出显示。
  28. 根据权利要求26所述的机器学习应用的提供装置,其中,还包括:
    自学习模块,被配置为在所述执行模块将获得的所述机器学习模型上线应用之后,基于与收集行为数据的步骤对应的配置信息,收集新的行为数据;基于与收集反馈数据的步骤对应的配置信息,收集新的反馈数据;基于所述新的行为数据和所述新的反馈数据,进行自学习,对已获得的所述机器学习模型进行更新或替换。
  29. 根据权利要求16所述的机器学习应用的提供装置,其中,所述第二显示模块 还被配置为:
    在所述执行模块执行所述机器学习的步骤中的第二步骤时,在所述学习圈图像的部分图像的侧部,显示所述第二步骤当前的执行状态信息。
  30. 根据权利要求16所述的机器学习应用的提供装置,其中,还包括状态信息提示模块,被配置为:
    在以下至少一种情况下,在所述学习圈图像中以第一预设方式显示机器学习的状态提示信息:在用户通过所述配置操作入口输入配置信息时;在接收到用户下发的学习圈启动指令后,所述学习圈图像从第一显示位置移动到第二显示位置之前;
    在所述执行模块依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息;
    在以下至少一种情况下,在所述学习圈图像中以第一预设方式显示机器学习的状态提示信息:在用户通过所述配置操作入口输入配置信息时;在接收到用户下发的学习圈启动指令后,所述学习圈图像从第一显示位置移动到第二显示位置之前;在所述执行模块依次执行各所述机器学习的步骤的同时,在所述学习圈图像中以第二预设方式显示与所述步骤对应的机器学习的状态提示信息。
  31. 一种电子设备,所述电子设备包括:壳体、处理器、存储器、电路板和电源电路,其中,电路板安置在壳体围成的空间内部,处理器和存储器配置在电路板上;电源电路,被配置为为上述电子设备的各个电路或器件供电;存储器被配置为存储可执行程序代码;处理器通过读取存储器中存储的可执行程序代码,被配置为执行前述权利要求1至15任一项所述的机器学习应用的提供方法。
  32. 一种计算机可读存储介质,所述计算机可读存储介质存储有一个或者多个程序,所述一个或者多个程序可被一个或者多个处理器执行,以实现前述权利要求1至15任一项所述的机器学习应用的提供方法。
PCT/CN2021/087310 2020-04-20 2021-04-14 一种机器学习应用的提供方法、装置、电子设备及存储介质 WO2021213234A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010315660.4A CN111931942A (zh) 2020-04-20 2020-04-20 一种机器学习应用的提供方法、装置、电子设备及存储介质
CN202010315660.4 2020-04-20

Publications (1)

Publication Number Publication Date
WO2021213234A1 true WO2021213234A1 (zh) 2021-10-28

Family

ID=73316330

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/087310 WO2021213234A1 (zh) 2020-04-20 2021-04-14 一种机器学习应用的提供方法、装置、电子设备及存储介质

Country Status (2)

Country Link
CN (1) CN111931942A (zh)
WO (1) WO2021213234A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111931942A (zh) * 2020-04-20 2020-11-13 第四范式(北京)技术有限公司 一种机器学习应用的提供方法、装置、电子设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103648045A (zh) * 2013-12-20 2014-03-19 乐视致新电子科技(天津)有限公司 通过侧边导航栏切换显示界面的方法及装置
CN108830383A (zh) * 2018-05-30 2018-11-16 第四范式(北京)技术有限公司 用于展示机器学习建模过程的方法及系统
US20190102098A1 (en) * 2017-09-29 2019-04-04 Coupa Software Incorporated Configurable machine learning systems through graphical user interfaces
CN110766164A (zh) * 2018-07-10 2020-02-07 第四范式(北京)技术有限公司 用于执行机器学习过程的方法和系统
CN110930256A (zh) * 2019-09-30 2020-03-27 北京九章云极科技有限公司 一种量化分析方法及量化分析系统
CN111931942A (zh) * 2020-04-20 2020-11-13 第四范式(北京)技术有限公司 一种机器学习应用的提供方法、装置、电子设备及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103648045A (zh) * 2013-12-20 2014-03-19 乐视致新电子科技(天津)有限公司 通过侧边导航栏切换显示界面的方法及装置
US20190102098A1 (en) * 2017-09-29 2019-04-04 Coupa Software Incorporated Configurable machine learning systems through graphical user interfaces
CN108830383A (zh) * 2018-05-30 2018-11-16 第四范式(北京)技术有限公司 用于展示机器学习建模过程的方法及系统
CN110766164A (zh) * 2018-07-10 2020-02-07 第四范式(北京)技术有限公司 用于执行机器学习过程的方法和系统
CN110930256A (zh) * 2019-09-30 2020-03-27 北京九章云极科技有限公司 一种量化分析方法及量化分析系统
CN111931942A (zh) * 2020-04-20 2020-11-13 第四范式(北京)技术有限公司 一种机器学习应用的提供方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN111931942A (zh) 2020-11-13

Similar Documents

Publication Publication Date Title
Wigdor et al. Brave NUI world: designing natural user interfaces for touch and gesture
US8909653B1 (en) Apparatus, systems and methods for interactive dissemination of knowledge
Suma et al. Adapting user interfaces for gestural interaction with the flexible action and articulated skeleton toolkit
CN110090444B (zh) 游戏中行为记录创建方法、装置、存储介质及电子设备
JP2016504638A (ja) タブレットコンピュータ
US11423795B2 (en) Cognitive training utilizing interaction simulations targeting stimulation of key cognitive functions
WO2021213234A1 (zh) 一种机器学习应用的提供方法、装置、电子设备及存储介质
CN114502971A (zh) 用于提供自动化产品测试的电子产品测试系统
MacNamara Evaluating the effectiveness of the gestalt principles of perceptual observation for virtual reality user interface design
US10198150B2 (en) Cross database data selection and correlation interface
Pihlajamäki From Desktop to Mobile: UI Patterns for User Interface Adaptation in Games
CN112755510A (zh) 一种移动端云游戏控制方法、系统和计算机可读存储介质
US20220366810A1 (en) Application onboarding tutorial system
Beer et al. Hello app inventor!: Android programming for kids and the rest of us
Paquette et al. Menu Controller: Making existing software more accessible for people with motor impairments
Cerqueira et al. Serious game interaction techniques applied to an operational satellite simulator
KR102606891B1 (ko) 게임 콘텐츠와 결합한 흐름 기반 교육용 프로그래밍 언어를 이용한 코딩 교육 시스템 및 방법
US20090276703A1 (en) Profile-based help for metaverse applications
Hung Usability and learnability improvements for the TaleBlazer game editor
US20230244506A1 (en) Conversational Assistant Control of a Graphical User Interface
KR102199247B1 (ko) 버그 제보를 위한 게임 서비스 제공 시스템
KR102027358B1 (ko) 아이템 생성 관련 게임 서비스 제공 방법 및 장치
Feng et al. MUD: Towards a Large-Scale and Noise-Filtered UI Dataset for Modern Style UI Modeling
Sansoe Testing Virtual Reality Interaction Techniques and Metaphors: The VR InteracTest Approach
Malankar Learning Android Game Development

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21792495

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21792495

Country of ref document: EP

Kind code of ref document: A1