Attention training processing method and apparatus

Abstract

The application discloses an attention training processing method and equipment, and belongs to the technical field of data processing. The method plans multiple training nodes for each training end, each training node is provided with an element training body which can be split into element blocks for training by the training end, the number of the element blocks is dynamically adjusted based on historical data of each training end, and real-time adjustment of training difficulty in the training process is realized. The application also supports collaborative training of each user in combination with an actual scene. When a collaborative request of a certain training end is received, the server calls training data of other training ends, selects one training end as a collaborative end, and the collaborative end collaboratively trains the training end which sends the collaborative request. The application takes the element training body as a training carrier, combines dynamic splitting and collaborative training, ensures that the training difficulty always matches the training end capability, strengthens the fusion of attention training and collaboration capability, and conforms to the actual scene.

Description

TECHNICAL FIELD

[0001] The present application relates to data processing technology, and in particular to an attention training processing method and device. BACKGROUND

[0002] Attention refers to the ability to purposefully concentrate mental activity on a certain thing for a long time, and attention training is a process of gradually strengthening the ability of continuous focus, precise orientation and anti-interference of the trainee by designing structured and goal-oriented tasks to guide the trainee to actively control attention resources.

[0003] With the development of technology, the form of attention training has evolved from traditional paper and pen static tasks to immersive training relying on VR and AR devices (such as target tracking in a virtual scene). For example, patent CN202010085920.3 discloses an attention training method based on VR and eye tracker, which realizes the collection and training of attention by immersing the trainee in a VR environment and combining eye movement data. This method only trains attention through eye movement, and only adjusts training in four dimensions: stability, breadth, transfer, and performance. The adjustment dimension is single and fixed, and cannot be dynamically adjusted according to the actual ability of the trainee during training.

[0004] Therefore, how to dynamically adjust the training difficulty according to the user's attention in attention training has become a problem to be solved. SUMMARY

[0005] The present application provides an attention training processing method and device, which can dynamically adjust the training difficulty according to the user's attention in attention training.

[0006] In a first aspect of the present application, an attention training processing method is provided, comprising:

[0007] determining a plurality of training nodes in a training link corresponding to each training end, the training nodes having corresponding element training bodies;

[0008] customizing and splitting the element training bodies based on the training data of the training end to obtain a plurality of element blocks of each training node;

[0009] responding to the cooperation request of the training end, selecting a cooperation end based on the training data of other training ends;

[0010] obtaining the training result of each training end according to the number of element blocks corresponding to each training end and the number of assembled blocks.

[0011] Optionally, in a possible implementation manner of the first aspect, the determination of a plurality of training nodes in a training link corresponding to each training end, the training nodes having corresponding element training bodies, comprises:

[0012] The training nodes corresponding to the training ends are constructed and trained, and the training nodes are arranged in sequence on the training link according to a preset order.

[0013] Optionally, in a possible implementation of the first aspect, the training data of the training end is used to customize the splitting of the element training body to obtain the element blocks of the training nodes, including:

[0014] The initial splitting number of the first element training body corresponding to each training end on the training link is called.

[0015] When the training end completes the assembly training of the corresponding element training body within a preset time according to the training data of the training end, the first splitting number is called from the preset splitting number table based on the completion time.

[0016] When the training end does not complete the assembly training of the element training body within a preset time, the second splitting number is called from the preset splitting number table based on the assembly number.

[0017] The initial splitting number, the first splitting number and the second splitting number are used to split the corresponding element training body to obtain the element blocks of the training nodes.

[0018] Optionally, in a possible implementation of the first aspect, the collaborative end is selected based on the training data of other training ends in response to the collaborative request of the training end, including:

[0019] The training end that does not issue a collaborative application is determined as an allocation end.

[0020] The historical training number of the element block in the element training body and the historical training times corresponding to the historical training number of each allocation end are called.

[0021] The classification set is constructed according to the historical training number and the historical training times, and the allocation end is selected as the collaborative end based on the classification set.

[0022] Optionally, in a possible implementation of the first aspect, the classification set is constructed according to the historical training number and the historical training times, and the allocation end is selected as the collaborative end based on the classification set, including:

[0023] The classification set corresponding to each historical training number is obtained by counting the historical training times of the allocation end.

[0024] The number of the element block in the element training body corresponding to the training end issuing the collaborative application is obtained as a reference number, and when there is a historical training number equal to the reference number, the corresponding classification set is determined as an equal set.

[0025] When it is determined that there is no classification set with the same number of historical training times, the classification sets are screened based on the number of historical training times in each classification set, and a collaborative set is obtained through the screening;

[0026] The distribution end is selected as the collaborative end according to the same set or the collaborative set.

[0027] Optionally, in a possible implementation manner of the first aspect, the screening of the classification sets based on the number of historical training times in each classification set comprises:

[0028] When it is determined that there is a classification set with the same number of historical training times, the corresponding classification set is obtained as the collaborative set;

[0029] When it is determined that there is no classification set with the same number of historical training times, the maximum number of historical training times in each classification set is extracted as a reference value, the difference between the reference value and other historical training times in each classification set is calculated, and the sum of all the differences is obtained as a selection parameter of each classification set;

[0030] The classification set corresponding to the smallest selection parameter is selected as the collaborative set.

[0031] Optionally, in a possible implementation manner of the first aspect, the method further comprises:

[0032] When it is determined that there are multiple classification sets with the same number of historical training times or multiple classification sets corresponding to the smallest selection parameter, the corresponding classification set is obtained as a first classification set;

[0033] The number of historical training times corresponding to the first classification set is obtained as a first number of historical training times, the absolute value of the difference between the first number of historical training times and the reference number is calculated, and the first classification set corresponding to the smallest absolute value is selected as the collaborative set.

[0034] Optionally, in a possible implementation manner of the first aspect, the selection of the distribution end as the collaborative end according to the same set or the collaborative set comprises:

[0035] The historical training time corresponding to the number of historical training times of each distribution end in the same set or the collaborative set is called, and the ratio of the historical training time to the number of historical training times is calculated to obtain a historical training efficiency;

[0036] The distribution end corresponding to the smallest historical training efficiency is selected as the collaborative end.

[0037] Optionally, in a possible implementation manner of the first aspect, the obtaining of the training result of each training end according to the number of element blocks corresponding to each training end and the number of assembled blocks comprises:

[0038] Obtain the element block assembled at the training node of each training end as an assembled block, count the number of all element blocks at the training node as the element number, and the assembly number corresponding to the assembled block;

[0039] The ratio of the assembly number to the element number is calculated to obtain the training result of each training end at the corresponding training node.

[0040] In a second aspect, the application provides an electronic device, comprising a memory, a processor and a computer program, the computer program being stored in the memory, and the processor executing the computer program to perform the method of the first aspect and various possible aspects related to the first aspect.

[0041] The beneficial effects of the application are as follows:

[0042] 1. The application takes the element training body as the core training carrier, and carries out attention training through the training form of element block assembly. The training end needs to identify the element block (the code of the element block), judge the position of the element block in the element training body, operate to complete the element block assembly, and finally combine into a complete element training body. This puzzle-like training mode focuses the user's attention on the whole process of element block assembly, and the user's participation is higher.

[0043] 2. The application customizes the element training body based on the training data of the training end, and matches the number of element blocks for each training node. If the training end completes the assembly within a preset time, the first split number is retrieved based on the completion time for the next element training body; if not, the second split number is retrieved based on the assembly number. This dynamic adjustment method makes the number of element blocks adapt to the current attention level of the training end, avoids the situation that the training end cannot continue training due to too many element blocks, and prevents the situation that the training difficulty is insufficient and the attention improvement effect is limited due to too few element blocks, so as to ensure that the training difficulty always matches the ability of the training end.

[0044] 3. The application can be trained cooperatively. When responding to the cooperative request of the training end, first, a classification set is constructed and a cooperative end is determined through screening. The efficiency evaluation of the distribution end is ensured to be in the same dimension during screening. First, the training times are counted according to the historical training number to form a classification set, and the same set with the same reference number is preferentially matched; if there is no same set, the classification set with the same historical training number or the smaller selection parameter (smaller difference in number within the set) is screened, if there are multiple candidate sets, the cooperative set is further determined through the absolute value of the difference between the historical training number and the reference number; finally, the historical training efficiency is calculated based on the ratio of the historical training time to the historical training number, and the distribution end with better efficiency is selected as the cooperative end. The whole screening process ensures the fairness and accuracy of the efficiency evaluation, so that the training end can obtain adaptive and efficient cooperative support. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 This is a schematic diagram illustrating the application scenario of the technical solution provided by the present invention;

[0046] Figure 2 A flowchart of an attention training processing method provided by the present invention;

[0047] Figure 3 A schematic diagram of a training node and element training body provided by the present invention;

[0048] Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided by the present invention. Detailed Implementation

[0049] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0050] like Figure 1 The diagram illustrates a scenario of the technical solution provided by this invention. This application scenario includes: a server, training terminals (in the example of three training terminals in the scenario diagram, one of which serves as a collaborative terminal), training nodes corresponding to the training terminals, and element training bodies corresponding to the training nodes. The server communicates with each training terminal, and each training terminal trains on the element training body at its corresponding training node. The server collects and saves the training data from each training terminal and performs customized splitting of the element training bodies at each training node based on the training data. When a collaborative request is received from a training terminal, the server retrieves the training data from other training terminals and selects one as the collaborative terminal. The collaborative terminal then performs collaborative training on the training terminal that issued the collaborative request. The training terminal can be a user using a VR device, and the training nodes and element training bodies can be nodes and training carriers set in a VR scene.

[0051] This invention provides an attention training processing method, such as... Figure 2 As shown, it includes:

[0052] S1, determine multiple training nodes in the training link corresponding to each training end, wherein each training node has a corresponding element training body.

[0053] It should be noted that most existing technologies train attention by having the eyes focus on a preset target using VR devices (such as adjusting the gaze direction by following the movement of the screen) or by combining it with computer-based human-computer interaction tests. These training methods lack a training platform that allows for multi-user collaborative interaction and dynamic adjustments. Simply recognizing the eyes or conducting human-computer tests cannot cover the multi-dimensional needs of attention in real-world scenarios, which require the integration of thought and action, and thus have low practicality.

[0054] Therefore, the application takes the element training body as the core training carrier, and realizes dynamic adjustment of training difficulty through split design of the element training body and the element block. The application plans multiple training nodes for each training end (user), each training node is provided with an element training body that can be split into element blocks for training of the training end, and the training end operates the element blocks to be assembled in the element training body. The application improves the initiative participation of the user, and is the fusion of thinking and action. The number of element blocks is dynamically adjusted based on historical data of each training end, so as to realize real-time adjustment of training difficulty in the training process. The application also supports collaborative training of each user in combination with an actual scene. The application takes the element training body as the training carrier, and improves the diversity and effectiveness of attention training in combination with dynamic split and collaborative training. Further, the application can realize training of the user through a VR device, so as to simulate the entire training scene by the VR device, and the user realizes assembly of the element blocks through corresponding actions.

[0055] It can be understood that first, training nodes are constructed for each training end. The training end can correspond to one training node, or can correspond to multiple training nodes. Each training node is provided with an element training body, and each element training body corresponds to the training node.

[0056] The training end represents a user participating in training, the training node is a basic node set up corresponding to the training end, the training end performs training at the corresponding training node, and the element training body is a training carrier for the training end to perform specific training work, which can be a complete jigsaw puzzle in a VR environment.

[0057] In some embodiments, (determining multiple training nodes in a training link corresponding to each training end, the training nodes having corresponding element training bodies) in step S1 includes S11:

[0058] S11, constructing training nodes corresponding to the training end, and arranging the training nodes on the training link in sequence according to a preset order.

[0059] It can be understood that, as shown in Figure 3 , the training nodes are arranged in sequence on the training link according to the preset training order, and the training end trains according to the arrangement order of the training nodes.

[0060] The training link is an ordered link composed of multiple training nodes.

[0061] S2, customizing and splitting the element training body based on training data of the training end, to obtain multiple element blocks of each training node.

[0062] It can be understood that, referring to Figure 3, the training difficulty and the user attention level are matched in real time, the challenge of the training is ensured, and the training interruption caused by the imbalance of the difficulty is avoided. For example, based on the training data of the first element training body corresponding to the A training end, the second element training body corresponding to the A training end is split, and then based on the training data of the second element training body, the third element training body corresponding to the A training end is split.

[0063] The training data refers to the data trained by the training end at the training node, including the completion state (all element blocks are completed / unfinished assembly) of the training node, the completion time (the total time length of the training end from starting assembly to completely matching the element training body), the assembly quantity (the number of element blocks assembled by the training end), etc. The element block refers to the independent unit after the element training body is split.

[0064] In some embodiments, the step S2 (customized splitting of the element training body based on the training data of the training end, to obtain a plurality of element blocks of each training node) includes S21-S24:

[0065] S21, the initial split quantity preset by each training end on the first element training body corresponding to the training link is called.

[0066] It can be understood that in the initial stage of attention training, each training end cannot split according to the training data due to the lack of training data of the training end, so the split quantity of the first element training body faced by each training end is preset, for example, the split quantity of the first element training body corresponding to the A training end in step S2 is preset.

[0067] The initial split quantity refers to the preset element block split quantity for the first element training body of each training end.

[0068] S22, according to the training data of the training end, when the training end completes the assembly training of the corresponding element training body within the preset time, the first split quantity is called from the preset split quantity table based on the completion time.

[0069] It can be understood that when determining the split quantity of each element training body, the training data of each training end in the last element training body is called, for example, when the split quantity of the second element training body corresponding to the A training end needs to be determined, the training data of the A training end in the first element training body is called. First, according to the training data, it is determined whether the training end in the last element training body is in a completed state, that is, it completes the assembly training of all element blocks within a preset time, for example, the A training end should complete the assembly of 15 element blocks within 20 seconds in the first element training body, and the A training end completes the assembly in 18 seconds, that is, the A training end completes the training within the preset time. 20 seconds represent the preset time, and 18 seconds represent the completion time.

[0070] It should be noted that in the attention training, the completed state of the training end in the training node can only indicate that the current difficulty is within the training end's ability range, but cannot distinguish the difference between "easy completion" and "hard completion". If the same number is used for all training ends, it will lead to insufficient difficulty increase for training ends with stronger ability, low attention improvement efficiency, and training ends with lower ability may not be able to smoothly train due to excessive quantity. Therefore, if the training end completes the assembly training of the element training body within the preset time, the first split quantity is called from the preset split quantity table based on the completion time.

[0071] Among them, the split quantity table is a preset record of the relationship between the time ratio (the ratio of the completion time to the preset time) and the first split quantity, for example, the split quantity table sets 3 time ratio intervals, the first split quantity is 20 blocks when the time ratio is less than or equal to 60%, the first split quantity is 18 blocks when the time ratio is 61%-80%, and the first split quantity is 15 blocks when the time ratio is 81%-100%, for example, the A training end completes the assembly within 18 seconds in the first element training body, and the preset time of the first element training body is 20 seconds, then the time ratio is 90%, and the first split quantity is 15 blocks, which means that the A training end is suitable for this intensity training and can continue to maintain.

[0072] Specifically, the preset time is a standard completion time length preset for the training node, the completion time refers to the total time length from the start of assembly to the complete assembly of the element training body, and the first split quantity refers to the split quantity set for the next element training body for the training end that completes the assembly training within the preset time.

[0073] Further, the assembling training refers to a training behavior of a training end (which can be through a VR device) for an element training body, specifically refers to a process that the training end visually recognizes, spatially judges (determines the position of the element block in the element training body), and finally combines into a complete element training body through operation interaction (VR handle grabbing and assembling element blocks). In the present application, the element block can be provided with a code, and the position of the element block in the element training body also has a code matching the element block, so that the element block and the position of the assembled element block correspond to each other.

[0074] S23, when it is determined that the training end does not complete the assembling training of the element training body within the preset time, a second split number is retrieved from the preset split number table based on the assembling number.

[0075] It can be understood that when the training end does not complete the assembling within the preset time, it means that the current difficulty may exceed its ability range, but the degree of not completing the assembling is different, some training ends may have a high degree of completion and assemble a large number of element blocks, and some training ends may have a low degree of completion and assemble a small number of element blocks, so targeted adjustment is needed, and therefore the second split number is retrieved from the preset split number table based on the assembling number.

[0076] The split number table further includes the relationship between the preset number ratio interval (the number ratio refers to the ratio of the assembling number to the number of element blocks in the element training body) and the second split number, for example, the split number table sets 3 number ratio intervals, the second split number is 10 blocks when the number ratio is less than or equal to 60%, the second split number is 12 blocks when the number ratio is 61%-80%, and the second split number is 15 blocks when the number ratio is 81%-100%, for example, the preset time of the A training end in the first element body is 20 seconds, the corresponding number of element blocks is 15 blocks, the A training end only completes 12 blocks in 20 seconds, the 12 blocks are the assembling number, the corresponding number ratio is 80%, and the second split number is 12 blocks, thereby reducing the training difficulty.

[0077] Specifically, the assembling number refers to the number of element blocks that the training end has assembled in the element training body, and the second split number refers to the split number set for the next element training body for the training end that does not complete the assembling training within the preset time.

[0078] S24, splitting the corresponding element training body based on the initial split number, the first split number and the second split number to obtain the element blocks of each training node.

[0079] It is not difficult to understand that after the number of element training bodies corresponding to each training end (the initial split number, the first split number and the second split number) is determined, the element training bodies are split according to the corresponding number to obtain the number of element blocks of the element training bodies on each training node, and the dynamic adjustment of the training difficulty in different training states is realized.

[0080] S3, in response to the cooperative request of the training end, the cooperative end is selected based on the training data of other training ends.

[0081] It should be noted that the present embodiment supports cooperative training of each user in combination with actual situations. In actual situations, the training end may not be able to carry out training, and other training ends need to be cooperated to ensure the continuation of the entire training. When the training end needs cooperation, a cooperative request is sent to the server, and the server retrieves the training data of other training ends according to the cooperative request to select other training ends as cooperative ends. After the training end is trained, there will be a number of element blocks assembled in each element training body. Some element training bodies may assemble 8 blocks, and some may assemble 10 blocks. The present embodiment will also count the number of corresponding assembly numbers, such as a training end assembling 8 blocks for 4 times and assembling 10 blocks for 2 times, which will also be reflected in the training data.

[0082] It can be understood that the selection of other training ends for cooperation is mainly based on the training efficiency of each cooperative end, so the training efficiency of each training end should be compared as much as possible based on the same dimension. The same assembly number and the same assembly number represent that the training difficulty faced by each training end in the initial stage is the same, and the familiarity of each training end based on the training difficulty is also the same, so analysis based on these two factors is more suitable. For example, A training end assembles 50 blocks for 10 times, and the total time is 300 seconds. Each time 50 blocks need 30 seconds on average. B training end assembles 3 blocks for 3 times, and the total time is 9 seconds. Each time 3 blocks need 3 seconds on average. The assembly number of A training end and B training end is different, indicating that they face different training difficulties, and they cannot be directly judged by these data.

[0083] Among them, the cooperative end refers to a training end selected from other training ends to assist in completing the assembly training.

[0084] In some embodiments, (in response to the cooperative request of the training end, the cooperative end is selected based on the training data of other training ends) in step S3 includes S31-S33:

[0085] S31, determining a training end that does not issue a cooperative application as a distribution end.

[0086] It is not difficult to understand that other training ends that do not issue a cooperative application are used as distribution ends.

[0087] S32, retrieve the historical training quantity of the element blocks in the element training body for each distribution end and the historical training times corresponding to the historical training quantity.

[0088] It can be understood that the historical training quantity of the element blocks in the element training body for each distribution end and the historical training times corresponding to the historical training quantity are retrieved, for example, the A training end historically assembles 10 element blocks 2 times, 20 element blocks 4 times, and 30 element blocks 1 time.

[0089] The historical training quantity refers to the number of element blocks assembled by the distribution end in each element training body in the past training, such as 10 blocks, 20 blocks, 30 blocks, etc. The historical training times refer to the total number of training completed by the distribution end for a certain historical training quantity, such as historically assembling 10 element blocks 2 times, 20 element blocks 4 times, and 30 element blocks 1 time, etc.

[0090] S33, constructing a classification set according to the historical training quantity and the historical training times, and selecting the distribution end as a collaborative end based on the classification set.

[0091] It can be understood that after retrieving the historical training quantity and the corresponding times of the distribution end, the distribution end is grouped according to the historical training quantity to form a plurality of classification sets, and the collaborative end is selected based on the classification set.

[0092] In some embodiments, step S33 (constructing a classification set according to the historical training quantity and the historical training times, and selecting the distribution end as a collaborative end based on the classification set) includes S331-S334:

[0093] S331, counting the historical training times of the distribution end to obtain a classification set corresponding to each historical training quantity.

[0094] It can be understood that the historical training times of each distribution end are counted corresponding to each historical training quantity, and a classification set is formed. Each classification set contains all distribution ends that have processed the corresponding historical training quantity and the corresponding training times, such as the a classification set corresponding to the historical training quantity of 10 blocks, the a classification set containing the A training end, the B training end, and the C training end, and the historical training times of the A training end corresponding to 10 blocks is 2 times, the historical training times of the B training end corresponding to 10 blocks is 2 times, and the historical training times of the C training end corresponding to 10 blocks is 3 times.

[0095] The classification set refers to a set formed by classifying the distribution end according to the historical training quantity.

[0096] S332, obtaining the number of element blocks in the element training body corresponding to the training end issuing the collaborative application as a reference quantity, and determining that there is a historical training quantity equal to the reference quantity to take the corresponding classification set as an equivalent set.

[0097] It can be understood that the number of element blocks in the element training body that the training end issuing a collaborative application currently needs to collaborate is obtained as a reference number, and the fine granularity of the element training body is determined. The classification set is analyzed again, and when there is a historical training number equal to the reference number, the corresponding classification set is taken as an equal set. For example, the number of element blocks corresponding to the training end issuing a collaborative application is 10 blocks (reference number), so first go to find if there is a 10-block number of the allocation end training in all classification sets. If there is, there must be a 10-block classification set, such as the a classification set in the S331 step, which contains A training end, B training end, and C training, corresponding to the number of times of 2, 2, and 3 respectively. Then the a classification set is selected as the equal set.

[0098] S333, when it is judged that there is no historical training number equal to the reference number, the classification sets are screened based on the historical training numbers in the classification sets, and a collaborative set is obtained by screening.

[0099] It should be noted that in the collaborative end screening process, it is not always possible to find a historical training number that completely matches the reference number. Therefore, the historical training numbers in each classification set are analyzed, and in the case of the same historical training number (the same fine granularity), the closer the historical training numbers corresponding to each allocation end, the more fair the evaluation of the efficiency of each allocation end. Therefore, the classification sets are screened based on the historical training numbers in the classification sets, and a collaborative set is obtained by screening.

[0100] In some embodiments, the step S333 (screening the classification sets based on the historical training numbers in the classification sets to obtain a collaborative set) includes S3331-S3333:

[0101] S3331, when it is determined that there is a classification set with the same historical training number, the corresponding classification set is obtained as a collaborative set.

[0102] It can be understood that when evaluating the training efficiency of each allocation end, the same dimension is as far as possible to be judged. Therefore, when there is no equal set, the historical training times in all classification sets are analyzed to determine the classification set with the same historical training times, and the corresponding classification set is obtained as the cooperative set. For example, the reference quantity is 5 pieces, the a classification set is a set corresponding to the historical training quantity of 10 pieces, including A, B and C three allocation ends, the number of times corresponding to 10 pieces of A allocation end is 2 times, the number of times corresponding to 10 pieces of B allocation end is 2 times, and the number of times corresponding to 10 pieces of C allocation end is 3 times, and there is a b classification set, the b classification set is a set corresponding to the historical training quantity of 20 pieces, including D, E and F three allocation ends, and the historical training times of the three allocation ends corresponding to 20 pieces are all 4 times, then the b classification set can be used as the cooperative set, because the historical training times of the allocation ends in the b classification set for training 20 pieces of element blocks are all 4 times, their efficiency can be accurately judged under the same standard, which can be understood as their familiarity with assembling 20 pieces of element blocks is the same.

[0103] S3332, when it is judged that there is no classification set with the same historical times, the maximum historical training times in each classification set is extracted as the reference value, the difference between the reference value and other historical training times in each classification set is calculated, and the sum of all the differences is obtained to get the selection parameter of each classification set.

[0104] It can be understood that if there is no classification set with the same historical times, the classification set with the historical training times close to each other in the internal set needs to be selected. For example, the a classification set is a set corresponding to the historical training quantity of 10 pieces, including A, B and C three allocation ends, the number of times corresponding to 10 pieces of A allocation end is 2 times, the number of times corresponding to 10 pieces of B allocation end is 2 times, and the number of times corresponding to 10 pieces of C allocation end is 3 times, and there is a c classification set, the c classification set is a set corresponding to the historical training quantity of 30 pieces, including G, H and I three allocation ends, and the historical training times of the three allocation ends corresponding to 20 pieces are 3 times, 5 times and 8 times respectively, the reference value of the a classification set is 3, the difference is 1 (3-2=1) and 1 (3-2=1) respectively, and the selection parameter is 2 (1+1=2), the reference value of the c classification set is 8, the difference is 5 (8-3=5) and 3 (8-5=3) respectively, and the selection parameter is 8 (5+3=8).

[0105] Among them, the reference value refers to the maximum value of the historical training times of the allocation ends in each classification set.

[0106] S3333, the classification set corresponding to the smallest selection parameter is selected as the cooperative set.

[0107] It is not difficult to understand that the smaller the selection parameter means that the overall proficiency of each allocation end in the corresponding classification set for the same historical training quantity is also closer, so the classification set corresponding to the smallest selection parameter is selected as the collaborative set.

[0108] Among them, the collaborative set refers to the absence of the same set, the selected classification set with the same or closest historical training times.

[0109] On the basis of the above embodiment, A1-A2 is further included:

[0110] A1, when there are multiple classification sets with the same historical training times or multiple classification sets corresponding to the smallest selection parameter, the corresponding classification set is taken as the first classification set.

[0111] It can be understood that there may be more than one classification set with the same historical times, and there may be more than one classification set corresponding to the smallest selection parameter, and when this situation occurs, the corresponding classification set is taken as the first classification set.

[0112] For example, the b classification set is a set corresponding to a historical training quantity of 20 blocks, including D, E, and F three allocation ends, and the historical training times of these three allocation ends corresponding to 20 blocks are all 4 times, and there is a d classification set, the d classification set is a set corresponding to a historical training quantity of 40 blocks, including J, K, and L three allocation ends, and the historical training times of these three allocation ends corresponding to 40 blocks are all 5 times, then the b classification set and the d classification set are both taken as the first classification set.

[0113] For another example, based on step S3332, the selection parameter of the c classification set is 2, and the selection parameter of the e classification set is also 2, then the c classification set and the e classification set are both taken as the first classification set.

[0114] A2, take the historical training quantity corresponding to the first classification set as the first historical training quantity, calculate the absolute value of the difference between the first historical training quantity and the reference quantity, and select the first classification set corresponding to the smallest absolute value as the collaborative set.

[0115] It can be understood that the historical training quantity corresponding to each set is extracted from all first classification sets as the first historical training quantity, and the absolute value of the difference between each first historical training quantity and the reference quantity is calculated, the smaller the absolute value, the closer the historical training quantity and the reference quantity (fine granularity is close), so the first classification set corresponding to the smallest absolute value is taken as the collaborative set.

[0116] S334, select the allocation end according to the same set or the collaborative set as the collaborative end.

[0117] In some embodiments, the step S334 (selecting the distribution end as the collaborative end according to the same set or the cooperative set) comprises S3341-S3342:

[0118] S3341, calling the historical training time of the historical training number corresponding to each distribution end in the same set or the cooperative set, calculating the ratio of the historical training time and the historical training number to obtain the historical training efficiency.

[0119] It can be understood that the other incompatible classification sets have been excluded by the same set or the cooperative set, and then the historical training efficiency of each distribution end in the same set or the cooperative set needs to be calculated. The historical training time of the historical training number corresponding to each distribution end in the same set or the cooperative set is called, and the ratio of the historical training time and the historical training number is calculated to obtain the historical training efficiency. For example, in step S3331, the b classification set is a set corresponding to a historical training number of 20 blocks, including D, E, and F three distribution ends. The historical training number corresponding to 20 blocks of the three distribution ends is 4 times, and the historical training time corresponding to the three distribution ends is 40 seconds, 80 seconds, and 120 seconds, respectively. Therefore, the historical training efficiency of the D, E, and F three distribution ends is 10 seconds / time, 20 seconds / time, and 30 seconds / time, respectively.

[0120] Among them, the historical training time refers to the total time of all historical training numbers corresponding to the historical training number of the distribution end.

[0121] S3342, selecting the distribution end corresponding to the smallest historical training efficiency as the collaborative end.

[0122] It can be easily understood that the smaller the value corresponding to the historical training efficiency, the higher the training efficiency of the distribution end, and the distribution end corresponding to the smallest historical training efficiency is selected as the collaborative end.

[0123] It should be noted that when the training number of the historical training number of each distribution end is too small, there may be only one time and no classification set can be formed. The following scheme can be used on the basis of the above embodiments, which comprises:

[0124] B1, determining the training end sending the collaborative application as the request end, and the remaining training ends as the distribution ends.

[0125] It can be easily understood that the training end sending the collaborative application is determined as the request end, and the remaining training ends are determined as the distribution ends.

[0126] B2, calling the reference number of element blocks in the element training body corresponding to the request end, the historical training number of element blocks in the element training body corresponding to each distribution end, and the corresponding historical operation time, calculating the ratio of the historical operation time and the historical training number to obtain the historical training speed of each distribution end.

[0127] It can be understood that the historical operation time refers to the time for the distribution end to complete the corresponding historical training quantity once, for example, the historical training quantity of the A distribution end is 10 blocks, and the historical operation time of the A distribution end is 10 seconds, so the historical training speed of the A distribution end is 1 second / block.

[0128] B3, calculate the ratio of the reference quantity to the historical training quantity of each distribution end to obtain an adjustment coefficient.

[0129] It should be noted that the training difficulty of each training end is different for different element block quantities, and the more the element block quantity, the higher the difficulty, so when measuring the efficiency of each distribution end, the difficulty of each distribution end should be balanced in the same dimension for analysis.

[0130] It can be understood that the adjustment coefficient is obtained by calculating the ratio of the reference quantity to the historical training quantity of each distribution end, for example, the reference quantity is 20, and the historical training quantity of the A distribution end is 10 blocks, so the adjustment coefficient of the A distribution end is 2 (20 / 10=2).

[0131] B4, based on the product of the historical training speed corresponding to each distribution end and the adjustment coefficient, the relative training speed is obtained, and the relative training speeds of each distribution end are compared, and the distribution end corresponding to the smallest speed value is selected as the collaborative end.

[0132] It can be understood that the relative training speed of each distribution end is obtained based on the product of the historical training speed corresponding to each distribution end and the adjustment coefficient, for example, the relative training speed of the A distribution end is 2 seconds / block, because the historical quantity of the A distribution end is 10 blocks, and the reference quantity is 20 blocks, if the A distribution end is based on the difficulty reference of 20 blocks, the efficiency of A should be slower than 1 second / block before, so the relative training speed is 2 seconds / block, and the time required for assembling one block is longer. Based on the reference quantity dimension, the relative training speeds of each distribution end are compared, and the distribution end corresponding to the smallest speed value is selected as the collaborative end, and the smallest speed value means that the time required for assembling an element block is the least, and then the efficiency is the highest.

[0133] S4, according to the number of element blocks corresponding to each training end and the number of assembled blocks, the training result of each training end is obtained.

[0134] In some embodiments, step S4 (obtaining the training result of each training end according to the number of element blocks corresponding to each training end and the number of assembled blocks) includes S41-S42:

[0135] S41, obtaining the element blocks assembled at the training node of each training end as the assembled blocks, counting the number of all element blocks at the training node as the element quantity, and the assembly quantity corresponding to the assembled blocks.

[0136] It can be understood that the assembled block refers to the element block assembled by each training end at the training node within a preset time, and the assembly number refers to the number of element blocks assembled by the training end in the element training body, i.e., the number of assembled blocks. The element number refers to the number of element blocks at each training node.

[0137] S42, calculating the ratio of the assembly number and the element number to obtain the training result of each training end at the corresponding training node.

[0138] It can be understood that the training result reflects the completion percentage of each training end at the training node within a preset time.

[0139] Referring to Figure 4 is a hardware structure schematic diagram of an electronic device provided by an embodiment of the present application. The electronic device 40 comprises a processor 41, a memory 42 and a computer program; wherein

[0140] The memory 42 is used for storing the computer program, and the memory can also be a flash memory. The computer program is, for example, an application program, a functional module and the like for implementing the above method.

[0141] The processor 41 is used for executing the computer program stored in the memory to implement each step of the device in the above method. For details, refer to the related description in the method embodiment.

[0142] Optionally, the memory 42 can be independent or integrated with the processor 41.

[0143] When the memory 42 is a device independent of the processor 41, the device can further comprise:

[0144] The bus 43 is used for connecting the memory 42 and the processor 41.

[0145] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the above embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An attention training processing method, characterized in that, include: Multiple training nodes are identified in the training link corresponding to each training end, and each training node has a corresponding element training body; Based on the training data from the training endpoint, the element training body is customized and split into multiple element blocks for each training node, including: Retrieve the preset initial split number of the first element training body corresponding to each training end in the training link; Based on the training data from the training end, when it is determined that the training end has completed the assembly training of the corresponding element training body within a preset time, the first split quantity is retrieved from the preset split quantity table based on the completion time. If the training end fails to complete the assembly training of the element training body within the preset time, the second split quantity is retrieved from the preset split quantity table based on the assembly quantity. The corresponding element training body is split based on the initial split number, the first split number, and the second split number to obtain the element blocks of each training node; In response to the collaboration request from the training end, collaboration ends are selected based on the training data from other training ends, including: The training endpoints that have not issued a collaboration request are designated as the allocation endpoints; Retrieve the historical training count of element blocks in the element training body from each distribution end and the historical training count of the corresponding historical training count; A classification set is constructed based on the historical training quantity and the historical training count, and an allocation end is selected as the collaboration end based on the classification set. The training results for each training end are obtained based on the number of element blocks and the number of assembled blocks corresponding to each training end.

2. The method according to claim 1, characterized in that, The determination of multiple training nodes in the training link corresponding to each training end, wherein each training node has a corresponding element training body, includes: Construct training nodes that correspond to the training end, and arrange the training nodes sequentially on the training link according to a pre-set order.

3. The method according to claim 1, characterized in that, The step of constructing a classification set based on the historical training quantity and the historical training count, and selecting an allocation end as a coordinating end based on the classification set, includes: The historical training counts at the statistical allocation end yield the classification set corresponding to each historical training count. The number of element blocks in the element training body corresponding to the training end that issued the collaboration request is used as a reference number. When it is determined that there is a historical training number that is equal to the reference number, the corresponding classification set is used as the equivalent set. When it is determined that there is no historical training quantity equal to the reference quantity, the classification set is filtered based on the historical training times in each classification set to obtain the collaborative set; The allocation end is selected as the collaboration end based on the equal set or collaboration set.

4. The method according to claim 3, characterized in that, The classification sets are filtered based on the historical training counts in each classification set, resulting in a collaborative set, which includes: When it is determined that there exists a set of classifications with the same number of historical training iterations, the corresponding set of classifications is obtained as the collaborative set; When it is determined that there is no category set with the same number of historical training times, the largest historical training time in each category set is extracted as a reference value. The difference between the reference value and other historical training times in each category set is calculated. The sum of all differences is calculated to obtain the selection parameters for each category set. The category set corresponding to the smallest selection parameter is selected as the collaborative set.

5. The method according to claim 4, characterized in that, Also includes: When it is determined that there are multiple classification sets with the same number of historical training times or multiple classification sets corresponding to the smallest selection parameter, the corresponding classification set is taken as the first classification set; Obtain the historical training quantity corresponding to the first classification set as the first historical training quantity, calculate the absolute value of the difference between the first historical training quantity and the reference quantity, and select the first classification set corresponding to the smallest absolute value as the collaborative set.

6. The method according to claim 5, characterized in that, The step of selecting the allocation endpoint as the cooperating endpoint based on the equal set or cooperative set includes: Retrieve the historical training time corresponding to the historical training counts for each allocation end in the same set or cooperative set, and calculate the ratio of historical training time to historical training counts to obtain the historical training efficiency. The allocation end corresponding to the lowest historical training efficiency is selected as the collaboration end.

7. The method according to claim 1, characterized in that, The training results for each training terminal are obtained based on the number of element blocks and the number of assembled blocks corresponding to each training terminal, including: Obtain the element blocks that have been assembled at the training node of each training end as the assembly blocks, count the total number of element blocks at the training node as the element quantity, and the assembly quantity corresponding to the assembly blocks. Calculate the ratio of the number of components to the number of elements to obtain the training results of each training end at the corresponding training node.

8. An electronic device, characterized in that, include: The method comprises a memory, a processor, and a computer program, wherein the computer program is stored in the memory and the processor executes the computer program to perform the method according to any one of claims 1 to 7.

Images