Information generation method and apparatus

Abstract

The application discloses an information generation method and device, and relates to the technical field of information processing. The method comprises the following steps: in response to obtaining a user performance prediction request, extracting work data of a to-be-predicted user from different data sources; inputting the work data of the to-be-predicted user into a prediction model, wherein in the prediction model, a corresponding branch network is activated based on the post type corresponding to the to-be-predicted user; and a specified feature is determined by the branch network; the output specified feature is input into a shared coding network after being processed to extract a general feature; a prediction result under different performance indicators of each post is output based on the general feature through an output head; and based on the prediction result, a score under different performance indicators of users of different posts is determined, and a performance comprehensive value is determined based on the score. The intelligent prediction of performance is realized.

Description

Technical Field

[0001] This invention relates to the field of information processing technology, and specifically to an information generation method and apparatus. Background Technology

[0002] User performance is usually calculated periodically according to performance criteria. For example, different companies set performance appraisal rules for different users, and then the performance is calculated manually on the performance calculation day.

[0003] This method cannot predict user performance in advance, and therefore cannot intervene in users beforehand. Therefore, how to achieve intelligent prediction of user performance is a technical problem to be solved. Summary of the Invention

[0004] The main objective of this invention is to provide a data processing method and apparatus to address the shortcomings of related technologies.

[0005] To achieve the above objectives, according to a first aspect of the present invention, an information generation method is provided, comprising: extracting work data of users to be predicted from different data sources; inputting the work data of users to be predicted into a prediction model, wherein, in the prediction model, a corresponding branch network is activated based on the job type corresponding to the user to be predicted; and the branch network determines specified features; the output specified features are processed and input into a shared coding network to extract general features; outputting prediction results for different performance indicators of each job based on the general features through an output header; determining scores for different performance indicators of users in different jobs based on the prediction results, and determining a comprehensive performance value based on the scores.

[0006] Optionally, outputting prediction results for different performance indicators for each position based on the general features through the output head includes: processing the position type to obtain a position type embedding vector, where e j =E·OneHot(j), where E is the job embedding matrix, j∈{1,2,...,N} j} represents the job type label; the specified feature h branch The concatenated feature h is obtained by concatenating the embedding vector of the job type with the concatenated feature h. cond =[h branch ;e j ].

[0007] Optionally, outputting prediction results for different performance indicators for each position based on the general features via the output header includes: concatenating a specified feature with the general features to obtain a target concatenated feature; and outputting prediction results based on the target concatenated feature. Where h is the target feature, and σ = ReLU is the activation function. W is the bias term; W is the dynamic weight, W = W base +ΔW·e j Among them, W base The basic weights of the shared encoding network are represented by ΔW, which is a dynamically adjusted tensor.

[0008] Optionally, when training the prediction model, the method includes: collecting historical multi-source heterogeneous data of users under different job types, wherein different multi-source heterogeneous data are associated with job types, and different multi-source heterogeneous data are labeled with different performance indicator information as training samples; determining the branch network corresponding to the data based on the job type, and using the training samples associated with the job type as the input of the corresponding branch network, and using the corresponding performance indicator information as the output to train the prediction model; wherein, during training, the parameters of each branch network are updated based on the loss function, and the parameters of the shared coding network are updated based on the loss function.

[0009] Optionally, when determining the loss function, the task-specific loss corresponding to different job types j includes classification loss. Regression loss Based on the parameters of the shared coding network, determine the regularization loss; based on the total task loss and the regularization loss, determine the loss function: Where α and β are the weights of the regression task and the classification task, respectively.

[0010] Optionally, the prediction model activates a branch network based on the job type corresponding to the user to be predicted, and the branch network outputs specified features including: determining the data types contained in the training samples corresponding to different job types; if the training samples corresponding to the same job type contain both numerical and text types, then the numerical and text features extracted by each branch network are fused to obtain the specified features: h fused =w num ·h num +w text ·h text h num For numerical characteristics, h text For text features; w num w text The weights corresponding to different job types, w num w represents the attention weights corresponding to the numerical features. text For the attention weights corresponding to the text features; where, during training, the attention weights are based on the job type embedding vector e. j Adjust modal fusion weights: w num w text =MLP(e j Based on the modal fusion weights, the fusion features, i.e. the specified features, are determined.

[0011] Optionally, the prediction model activates a branch network based on the job type corresponding to the user to be predicted. The branch network outputs specified features, including: if the training samples corresponding to the same job type are numerical or textual, then for training samples of numerical type, the first branch network is activated, wherein the training samples are normalized by the first branch network and then input into a fully connected layer to extract numerical features; or, for training samples of text type, the second branch network is activated, wherein the CLS vector is extracted from the training samples by the second branch network and then input into a fully connected layer for dimensionality reduction to obtain textual features.

[0012] Optionally, the method further includes: randomly generating an initial development path based on the development actions of each performance indicator in the development action library, wherein the development action includes the name of the development action, the required time, the required cost, the improvement of any performance indicator after the completion of the development action, and the coverage of the development action to various needs of the enterprise; filtering the initial development path based on the objective function to obtain the development path corresponding to different positions and different users, wherein each development path includes three stages: short-term, medium-term and long-term, and each stage includes m development actions.

[0013] Optionally, when constructing the objective function, the method includes: determining the score of any user in any position under different performance indicators, and the difference between these scores and the target thresholds of each performance indicator corresponding to that position; and calculating the corresponding weakness weight based on the difference. in, Based on the weight of the weakest link And the increase Δs of any performance indicator after the completion of the development action. i Determine the first objective function Based on the set weights of each enterprise's needs j The degree to which development actions in the development path cover firm demand j (c) j Determine the second objective function The sum of the contributions of all development actions in the development path to firm demand j is the coverage c. j The third objective function is determined based on the total time cost T required for the user to complete all development actions in the development path and the direct capital cost K paid by the company for the user's development path, f3=-(α·T+β·k), where α and β are cost coefficients; the objective function composed of the first objective function, the second objective function and the third objective function is optimized based on the preset total time and total cost as constraints.

[0014] According to a second aspect of the present invention, a user performance intelligent prediction device is provided, comprising: a data acquisition unit for extracting work data of users to be predicted from different data sources; a prediction unit for inputting the work data of users to be predicted into a prediction model, wherein, in the prediction model, a corresponding branch network is activated based on the job type corresponding to the user to be predicted; and a specified feature is determined by the branch network; the output specified feature is processed and input into a shared coding network to extract a general feature; and a prediction result under different performance indicators for each job is output through an output head based on the general feature; and an output unit for determining the score of users in different job positions under different performance indicators based on the prediction result, and determining a comprehensive performance value based on the score.

[0015] This embodiment presents an information generation method and apparatus. The method includes, in response to receiving a user performance prediction request, extracting work data of the user to be predicted from different data sources; inputting the work data of the user to be predicted into a prediction model, wherein, in the prediction model, a corresponding branch network is activated based on the job type corresponding to the user to be predicted; and the branch network determines specified features; the output specified features are processed and input into a shared coding network to extract general features; the prediction results for different performance indicators of each job are output through an output header based on the general features; based on the prediction results, scores for different performance indicators of users in different jobs are determined, and a comprehensive performance value is determined based on the scores. This achieves intelligent performance prediction. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the information generation method according to an embodiment of the present invention;

[0018] Figure 2 This is an application diagram of the information generation method according to an embodiment of the invention;

[0019] Figure 3 This is a schematic diagram of the electronic device structure according to an embodiment of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] According to embodiments of the present invention, an information generation method is provided, such as... Figure 1 As shown, steps 101 to 103 are included below:

[0024] Step 101: Extract the work data of the users to be predicted from different data sources.

[0025] In this step, users can trigger performance prediction operations for different users on the terminal interface. Once triggered, a performance prediction request is generated and sent to the server. The server retrieves data from various data sources, including but not limited to: work data from the user's business systems (e.g., patent management systems, which are merely examples); communication data from the user's social media software; telephone record data from the telephone system; and data from different development environments (e.g., code data from different program development environments). These data sources are illustrative, and data can be obtained from different sources as needed. Different user roles require different data; therefore, work data is collected from different data sources for different roles. For example, sales users can extract sales-related data such as workload from the business system and telephone record data from the telephone system; development users can collect code-related data from different development environments, such as code volume, code comment rate, and code error rate.

[0026] Step 102: Input the work data of the user to be predicted into the prediction model, wherein, in the prediction model, the corresponding branch network is activated based on the job type of the user to be predicted; and the branch network determines the specified features; the output specified features are input into the shared coding network to extract the useful features; and the prediction results under different performance indicators for each job are output through the output head based on the general features.

[0027] Step 103: Based on the prediction results, determine the scores for different performance indicators for users in different positions, and determine the comprehensive performance value based on the scores.

[0028] In step 102, after the work data of the user to be predicted is input into the prediction model, different branch networks of the prediction model are activated based on the job type label. The data types of work data differ for different job types, and different branch networks process different types of data within the work data. Different branch networks collect different features for different types of work data. For example, for sales users, numerical features related to phone records can be extracted, such as call time, number of calls, and call duration; sales figures from the business system can be extracted; and textual features of communication content from social media can be extracted, such as specified semantic features. After the branch networks process different types of data, features are merged for the same job type to obtain the specified features.

[0029] While performance forecasting for a single user position can achieve the goal of intelligent forecasting, the performance of a user in a particular position should not only consider their own data. For example, when evaluating the execution efficiency of a sales position, in addition to the order processing speed data of the sales position, the task allocation response time of the management position should also be taken into consideration.

[0030] Therefore, in this embodiment, for a specific job type, a target job type associated with that job type is pre-determined. This target job type, such as sales or management, is then associated with it. During prediction, when the work data of the user to be predicted is input into the prediction model, the data of the target job type associated with that job type is also input into the prediction model. For a sales user, the user data of the associated management user is determined. The job data of both users are input into the prediction model, and different branch networks are activated based on different job types. Through the branch networks, specified features for different job types can be obtained. These features are aligned and concatenated into a job type embedding vector, which is then input into the shared encoding network. The general feature vector output from the shared encoding layer is used to represent common performance indicator features across job types.

[0031] After extracting different features for different job types, the branch network continues to extract general features, such as features related to the execution efficiency of sales positions. The input data for different positions are processed through independent feature extraction branches to generate job-specific features.

[0032] In step 103, the general features and the job-specific features are concatenated and input to the output header. The output header outputs the performance prediction results of the user to be predicted. For example, for a sales position, it could be the prediction of sales revenue, the prediction of data reflecting execution efficiency, etc. The data can be used to assign scores, and the comprehensive score is determined by assigning different weights to the assigned scores.

[0033] As an optional implementation of this embodiment, the output specified features, after processing, include: processing the job type to obtain a job type embedding vector, where e j =E·OneHot(j), where E is the job embedding matrix, j∈{1,2,...,N} j} represents the job type label; the specified feature h branch The concatenated feature h is obtained by concatenating the embedding vector of the job type with the concatenated feature h. cond =[h branch ;e j The spliced ​​features are input into a shared coding network to extract general features.

[0034] As an optional implementation of this embodiment, outputting the prediction results for different performance indicators for each position based on the general features through the output header includes: concatenating the specified features with the general features to obtain the target concatenated features; and outputting the prediction results based on the target concatenated features. Where h is the target feature, and σ = ReLU is the activation function. W is the bias term; W is the dynamic weight, W = W base +ΔW·e j Among them, W base The basic weights of the shared encoding network are represented by ΔW, which is a dynamically adjusted tensor.

[0035] In the above-mentioned optional implementation methods, job embedding plays a guiding role, directing the output head to focus on job-related features. (Job embedding matrix) d e For the embedding dimension, N j This represents the total number of positions. The position embedding vector is e. j Through tensor multiplication ΔW·e j Generate parameter offsets so that the weights within the same layer are dynamically adjusted according to job type, i.e., dynamic weight adjustment, such as the weight of a technical job. j It increases the weight of neurons related to collaboration ability and suppresses the weight of neurons related to sales volume. baseThe output header contains the initial weights of the first layer, which are general parameters independent of job positions; the ΔW three-dimensional tensor generates weight offsets through the job position embedding vector, dynamically adjusting the basic weights.

[0036] The above-mentioned optional implementation methods combine shared features with job embedding through feature concatenation and dynamic weight generation mechanisms to achieve differentiated output (different jobs use independent W and b). j (Ensuring output space isolation) and enabling conditional computation.

[0037] For example, if the input data belongs to a developer, a core maintainer of a code module, and the obtained commit history shows that 15 defects were fixed in the past 30 days, with an average fix time of 1.8 hours, the model extracts commit history features (low average fix time, high code quality), shares the collaborative capabilities and execution efficiency of the coding layer, and generates general features; dynamic weights can enhance the contribution of collaborative capabilities to the fix speed, and the output predicted value is 1.5 hours / defect.

[0038] Among the above-mentioned optional implementation methods, the chained processing of dynamic routing, feature fusion, and conditional weight generation can accurately adapt to the needs of technical positions and extract key drivers of performance indicators from heterogeneous data. It maintains cross-position generalization, and common features support multi-task joint optimization. Efficient inference meets real-time prediction requirements.

[0039] As an optional implementation of this embodiment, the method for training the prediction model includes: collecting historical multi-source heterogeneous data of users under different job types, wherein different multi-source heterogeneous data are associated with job types, and different multi-source heterogeneous data are labeled with different performance indicator information as training samples; determining the branch network corresponding to the data based on the job type, and using the training samples associated with the job type as the input of the corresponding branch network, and using the corresponding performance indicator information as the output to train the prediction model; wherein, during training, the parameters of each branch network are updated based on the loss function, and the parameters of the shared coding network are updated based on the loss function.

[0040] In this optional implementation, different job types use different data types, such as structured data like sales revenue and code commit volume, and unstructured data like text (e.g., weekly reports). This data can be used as input to the model. Different job types have different performance indicators and output targets. For example, a sales position's prediction might include sales revenue forecasting, using a regression task; a development position's prediction might be code quality level forecasting, using a classification task to output a probability distribution. Different types of data from multiple job types (e.g., sales and related management positions) are input into different branch networks, and labeled performance indicators are used as output to train the prediction model.

[0041] During training, different input data and output targets can be used to train different branches of the network. Input data consists of user multimodal data (numerical, text, etc.). Each sample is labeled with its job type and performance indicators; for example, sales positions are labeled with sales revenue (regression task, labels are continuous values). Technical positions are labeled with code quality level (classification task, labels are category numbers). Management positions are labeled with team effectiveness score (regression) + risk level (classification). The error between the predicted value and the true label is calculated using a loss function (such as MSE, cross-entropy), and the model parameters are updated via backpropagation.

[0042] For example, the training process includes data preparation and batch processing, grouping data by job type to ensure a balanced sample size for different job positions in each training batch (e.g., each batch contains 10 sales positions and 10 management positions). During forward propagation, branch network activation is performed, selecting branch networks based on job labels (e.g., activating a fully connected network for sales positions and a GNN for technical positions), followed by feature extraction. The branch features are concatenated with the job embeddings and output through an output header; for example, for sales positions, the regression header outputs the predicted sales amount. For technical positions, the output header outputs the code quality level (logits).

[0043] As an optional implementation in this embodiment, when determining the loss function, the task-specific loss corresponding to different job types j includes classification loss. Regression loss The total loss of the task is α and β are task weights; the regularization loss λ||θ|| is determined based on the parameters of the shared coding network. 2 Based on the total task loss and the regularization loss, determine the loss function:

[0044] In this optional implementation, the loss function is determined by the task objective type. Taking a development role as an example, code quality grading corresponds to a classification task, using cross-entropy loss; sales forecasting for a sales role could be a regression task, using mean squared error loss. Each role's output head selects the loss function based on the objective type; for example, for a sales role, the numerical branch would use a regression task head with mean squared error loss. In other words, the branch network processes different types of data, and the output head determines the type of loss function. The model can flexibly adapt to multiple roles and task scenarios while maintaining the clarity and rationality of the training objectives.

[0045] Furthermore, we distinguish between shared parameters and dynamic parameters, and apply targeted regularization. Specifically, we apply regularization to shared parameters (general feature extraction layer parameters θ). shared θ, the parameter of the dynamic weight generation layer dynamic Apply L2 regularization This loss function is used to jointly update the shared coding network layer and the dynamic weight layer. θ shared The purpose of setting a regularization loss is to limit the complexity of general features and prevent them from becoming overly dependent on noise or local features. θ dynamic The purpose of setting regularization loss is to constrain the fluctuation range of dynamic weights and avoid model instability due to over-adjustment.

[0046] In the early stages of training, set a small λ (e.g., 0.01) to avoid interfering with the model's rapid convergence. In the later stages of training, gradually increase λ (e.g., 0.1) to strengthen the constraint on complex parameters. Adjust λ according to the training stage to achieve a balance between fitting and generalization.

[0047] As an optional implementation in this embodiment, the prediction model activates a branch network based on the job type corresponding to the user to be predicted. The specified features output by the branch network include: determining the data types contained in the training samples corresponding to different job types; if the training samples corresponding to the same job type contain both numerical and text types, then the numerical and text features extracted by each branch network are fused to obtain the specified features: h fused =w num ·h num +w text ·h text h num For numerical characteristics, h text For text features; w num w text The weights corresponding to different job types, w num w represents the attention weights corresponding to the numerical features. text For the attention weights corresponding to the text features; where, during training, the attention weights are based on the job type embedding vector e. j Adjust modal fusion weights: w num w text =MLP(e j Based on the modal fusion weights, the fusion features, i.e. the specified features, are determined.

[0048] In this optional implementation, the work data for a job type may include multiple types of data, and the different types of data are input into different branch networks and then merged.

[0049] As an optional implementation of this embodiment, the prediction model activates a branch network based on the job type corresponding to the user to be predicted. The specified features output by the branch network include: if the training samples corresponding to the same job type are numerical or textual, then for training samples of numerical type, the first branch network is activated, wherein the training samples are normalized by the first branch network and then input to the fully connected layer to extract numerical features; or, for training samples of text type, the second branch network is activated, wherein the CLS vector is extracted by the second branch network and input to the fully connected layer for dimensionality reduction to obtain textual features.

[0050] In this optional implementation, feature extraction is performed through different branch networks based on the data type corresponding to the job type.

[0051] As an optional implementation of this embodiment, the method further includes: randomly generating an initial development path based on the development actions of each performance indicator in the development action library, wherein the development action includes the development action name, required time, required cost, improvement of any performance indicator after completing the development action, and coverage of the development action to various needs of the enterprise; filtering the initial development path based on the objective function to obtain development paths corresponding to different positions and different users, wherein each development path includes three stages: short-term, medium-term and long-term, and each stage includes m development actions.

[0052] In this optional implementation, based on performance prediction results, the system generates a personalized career development path for each user. By analyzing the user's strengths and weaknesses, and combining company business needs with the user's personal career goals, the system provides customized training suggestions, skills enhancement plans, and career advancement paths to help users achieve mutual development with the organization.

[0053] Specifically, after performance prediction, a user's performance score can be obtained. Based on the scores of each indicator and the total score, the user's weaknesses can be identified, such as insufficient execution efficiency. Furthermore, because companies have different strategic needs, objective functions related to both can be established and optimized to obtain different development paths for different users.

[0054] The development path planning needs to simultaneously optimize user skill improvement (goal 1), company demand coverage (goal 2), and cost control (goal 3). There may be conflicts between the three goals (such as high costs required for rapid skill improvement), so an optimal solution is needed.

[0055] First, an action library containing all possible development actions needs to be built. Each action includes an action name (e.g., Advanced Python Training), time cost (e.g., 2 months), expense cost, associated skills (e.g., Python), skill improvement margin (e.g., 0.2), and the company's business needs covered by completing the action (e.g., "Cloud Computing": 0.0, "AI": 0.3). Each path is divided into three phases: short-term (<6 months), medium-term (6-18 months), and long-term (>18 months). Each phase contains at least one action (the number is random). Generating the initial development path includes initializing an empty path; filling each phase with actions by randomly sampling from the action library without replacement and assigning them to each phase, while also satisfying phase time limits; deleting duplicate actions (e.g., selecting the same training multiple times in the same phase); and regenerating if the total time or expense exceeds the limit (e.g., total time > 24 months, total expense > 20,000 yuan) (this condition also serves as a constraint on the optimization objective function). Total time refers to the total time cost required for the user to complete all actions in the development path; and the direct financial cost paid by the company for the user's development path.

[0056] Furthermore, the initial development path is selected based on the constructed objective function.

[0057] As an optional implementation in this embodiment, the method for constructing the objective function includes: determining the score of any user in any position under different performance indicators, and the difference between the score and the target threshold of each performance indicator corresponding to that position; and calculating the corresponding weakness weight based on the difference. in, Based on the weight of the weakest link And the increase Δs of any performance indicator after the completion of the development action. i Determine the first objective function Based on the set weights of each enterprise's needs j The degree to which development actions in the development path cover firm demand j (c) j Determine the second objective function The sum of the contributions of all development actions in the development path to firm demand j is the coverage c. j The third objective function is determined based on the total time cost T required for the user to complete all development actions in the development path and the direct capital cost K paid by the company for the user's development path, f3=-(α·T+β·k), where α and β are cost coefficients; the objective function composed of the first objective function, the second objective function and the third objective function is optimized based on the preset total time and total cost as constraints.

[0058] In this optional implementation, the speed of improving users' skill gaps is maximized.

[0059] The weight of improving the i-th skill is determined by the gap in weaknesses. Δs i The improvement margin of the i-th skill, such as "Python training" with an improvement margin of 0.2, is determined by the effect of the development action.

[0060] Maximize the skill match with the company's business needs. c j The coverage of the j-th company's requirement is the sum of the weights of the related actions in the path, c. j =∑(Action demand weight × Whether the action is selected). Manually set. For example, c j The sum of the contributions of all actions in the development path to the company's demand for item j reflects the degree to which the path aligns with the company's strategy. For example, if the company's future demand is to focus on developing cloud computing (demand weights)... The value is 0.9; Actions in the user development path:

[0061] Action 1: Attend AWS certification training (related to "cloud computing", requirement coverage = 0.6).

[0062] Action 2: Participate in cloud migration projects (related to "cloud computing", requirement coverage = 0.4).

[0063] Coverage calculation: c 云计算 =0.6+0.4=1.0 This path fully covers the cloud computing needs, contributing 100%.

[0064] Furthermore, minimizing time and capital costs, α and β, the cost coefficients of time and expenses, and controlling the penalty intensity of costs in the overall objective can be set as needed.

[0065] Furthermore, the constraints can be total time ≤ 24 months (long-term path upper limit) and total cost ≤ company budget (e.g., 20,000 yuan / year).

[0066] For example, refer to Figure 2 For the technical user, the skill gaps are in "cloud computing" (currently 0.5, target 0.8) and "collaboration ability" (currently 0.6, target 0.7). The company requires cloud computing (weight 0.9) and AI (weight 0.5); optional development actions could include AWS certification training, collaboration skills workshops, and AI algorithm project practice.

[0067] Path A: AWS Certification Training → AI Algorithm Project Practice, Skills Enhancement, Cloud Computing: 0.5 → 0.8 (Δs = 0.3); Demand Coverage, Cloud Computing: 0.6, AI: 0.5, Total Coverage: c 云计算

[0068] +c AI=0.6+0.5=1.1.

[0069] Path B: Collaboration Skills Workshop → AWS Certification Training, Skills Enhancement, Collaboration Skills: 0.6 → 0.8 (Δs = 0.2), Cloud Computing: 0.5 → 0.8 (Δs = 0.3); Requirements Coverage, Cloud Computing: 0.6, Total Coverage c 云计算 =0.6.

[0070] Optimize your choices. If your company's current priorities are cloud computing > AI > collaboration capabilities: Path A is better (covering higher priority needs). If you need to address weaknesses simultaneously: Path B is better (improving both collaboration capabilities and cloud computing).

[0071] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0072] According to an embodiment of the present invention, a user performance intelligent prediction device is also provided, including a data acquisition unit for extracting work data of the user to be predicted from different data sources in response to a user performance prediction request; a prediction unit for inputting the work data of the user to be predicted into a prediction model, wherein, in the prediction model, a corresponding branch network is activated based on the job type corresponding to the user to be predicted; and a specified feature is determined by the branch network; the output specified feature is processed and input into a shared coding network to extract a general feature; and a prediction result under different performance indicators for each job is output through an output head based on the general feature; and an output unit for determining the score of the user under different performance indicators for different jobs based on the prediction result, and determining the comprehensive performance value based on the score.

[0073] According to embodiments of the present invention, the present invention also provides an electronic device, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to implement the methods described in any of the above embodiments.

[0074] According to embodiments of the present invention, the present invention also provides a readable storage medium storing computer instructions that enable a computer to perform the methods described in any of the above embodiments when executed.

[0075] According to embodiments of the present invention, the present invention also provides a computer program product that, when executed by a processor, can implement the methods described in any of the above embodiments.

[0076] Figure 3 A schematic block diagram of an example electronic device 300 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices.

[0077] like Figure 3 As shown, the electronic device 300 includes a computing unit 301, which can perform various appropriate actions and processes based on a computer program stored in a read-only memory (ROM) 302 or a computer program loaded from a storage unit 308 into a random access memory (RAM) 303. The RAM 303 may also store various programs and data required for the operation of the electronic device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0078] Multiple components in electronic device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows electronic device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0079] The computing unit 301 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as the object matching method. For example, in some embodiments, the object matching method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of the methods described above may be performed.

[0080] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0081] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0082] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

Claims

1. An information generation method, characterized in that, include: Extract work data of the users to be predicted from different data sources; The work data of the user to be predicted is input into the prediction model, wherein, in the prediction model, the corresponding branch network is activated based on the job type of the user to be predicted; and the branch network determines the specified features; the output specified features are processed and input into the shared coding network to extract the general features; and the prediction results under different performance indicators for each job are output through the output head based on the general features. Based on the prediction results, the scores for different performance indicators for users in different positions are determined, and the overall performance value is determined based on the scores. When the specified output features are processed, the method includes: processing the job type to obtain a job type embedding vector, wherein, e j =E·OneHot(j) , E Embed the matrix for job positions. As a job type label; the specified features h branch The concatenated feature is obtained by concatenating the embedding vector of the job type. ; The output head outputs prediction results for different performance indicators for each position based on the general features, including: concatenating the concatenated features with the general features to obtain target concatenated features; and outputting prediction results based on the target concatenated features. ,in, To splice features for the target, For activation function, For bias terms; For dynamic weights, ,in, To share the basic weights of the coding network, To dynamically adjust the tensor.

2. The information generation method according to claim 1, characterized in that, When training the prediction model, the method includes: collecting historical multi-source heterogeneous data of users under different job types, wherein different multi-source heterogeneous data are associated with job types, and different multi-source heterogeneous data are labeled with different performance indicator information as training samples; The corresponding branch network is determined based on the job type, and the training samples associated with the job type are used as the input of the corresponding branch network, while the corresponding performance indicator information is used as the output to train the prediction model. Specifically, during training, the parameters of each branch network are updated based on the loss function, and the parameters of the shared coding network are updated based on the loss function.

3. The information generation method according to claim 2, characterized in that, When determining the loss function, the task-specific loss corresponding to different job types j includes classification loss. Regression loss ; Based on the parameters of the shared coding network, the regularization loss is determined; Based on the task-specific loss and the regularization loss, determine the loss function: ,in, , These are the weights for the regression task and the classification task, respectively.

4. The information generation method according to claim 3, characterized in that, The prediction model activates a branch network based on the job type corresponding to the user to be predicted, and the branch network outputs specified features including: determining the data types contained in the training samples corresponding to different job types; If the training samples corresponding to the same job type contain both numerical and textual data, then the numerical and textual features extracted from each branch network are fused to obtain the specified features: , For numerical features, h text Text features; , The weights corresponding to different job types These are the attention weights corresponding to the numerical features. These are the attention weights corresponding to the text features; During training, embedding vectors based on job type are used. Adjust modal fusion weights: ; The fusion feature, i.e. the specified feature, is determined based on the modality fusion weights.

5. The information generation method according to claim 4, characterized in that, The prediction model activates a branch network based on the job type corresponding to the user to be predicted, and the branch network outputs specified features including: If the training samples corresponding to the same job type are of numerical or text type, then for training samples of numerical type, the first branch network is activated. The training samples are normalized by the first branch network and then input into the fully connected layer to extract numerical features. Alternatively, for training samples of text type, the second branch network is activated. The training samples are extracted using the second branch network and then input into the fully connected layer for dimensionality reduction to obtain text features.

6. The information generation method according to claim 1, characterized in that, The method further includes: An initial development path is randomly generated based on the development actions of each performance indicator in the development action library. The development action includes the name of the development action, the time required, the cost required, the improvement of any performance indicator after the completion of the development action, and the coverage of the development action to various needs of the enterprise. The initial development path is selected based on the objective function to obtain the development path corresponding to different positions and users. Each development path includes three stages: short-term, medium-term and long-term, and each stage includes m development actions.

7. The information generation method according to claim 6, characterized in that, When constructing the objective function, the method includes: determining the score of any user in any position under different performance indicators, and the difference between the score and the target threshold of each performance indicator corresponding to that position; and calculating the corresponding weakness weight based on the difference. ,in, = Difference between each performance indicator / Total difference between all performance indicators; Based on the weight of the weakest link and the extent of improvement in any performance indicator after the completion of development actions. Determine the first objective function ; Based on the set weights of each enterprise's needs j The degree to which development actions in the development path cover firm demand j Determine the second objective function The coverage is the sum of the contributions of all development actions in the development path to firm demand j. ; The third objective function is determined based on the total time cost T required for a user to complete all development actions in the development path, and the direct financial cost K paid by the company for the user's development path. ,in, , Cost coefficient; The objective function, composed of the first objective function, the second objective function, and the third objective function, is optimized based on the preset total time and total cost as constraints.

8. A user performance intelligent prediction device, characterized in that, include: The data acquisition unit extracts the work data of the user to be predicted from different data sources; The prediction unit inputs the work data of the user to be predicted into the prediction model. In the prediction model, the corresponding branch network is activated based on the job type of the user to be predicted; the branch network determines the specified features; the output specified features are processed and input into the shared coding network to extract the general features; and the prediction results for different performance indicators of each job are output through the output head based on the general features. The output unit is used to determine the scores for different performance indicators for users in different positions based on the prediction results, and to determine the comprehensive performance value based on the scores; When the specified output features are processed, the method includes: processing the job type to obtain a job type embedding vector, wherein, e j =E·OneHot(j) , E Embed the matrix for job positions. As a job type label; the specified features h branch The concatenated feature is obtained by concatenating the embedding vector of the job type. ; The output head outputs prediction results for different performance indicators for each position based on the general features, including: concatenating the concatenated features with the general features to obtain target concatenated features; and outputting prediction results based on the target concatenated features. ,in, To splice features for the target, For activation function, For bias terms; For dynamic weights, ,in, To share the basic weights of the coding network, To dynamically adjust the tensor.

Images