Task bidirectional matching method and device, equipment and storage medium

Abstract

The invention provides a task bidirectional matching method and device, equipment and a storage medium, the task bidirectional matching method comprises the following steps: obtaining a task description text and a capability semantic vector of a task member, and constructing a task semantic vector of a to-be-matched task according to the task description text; on the basis of the capability semantic vector and the task semantic vector, respectively determining candidate task members of the task description text and candidate to-be-matched tasks of the task members; and performing task matching according to the candidate task members and the candidate tasks to be matched. According to the method, two-way closed-loop adaptation of tasks is realized through two adaptation paths of task driving and member driving.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a task two-way matching method, device, equipment and storage medium. Background Art

[0002] With the flattening of corporate organizational structures, the prevalence of remote collaboration, and the normalization of multi-project parallelism, achieving efficient and accurate task allocation within teams has become a core issue impacting collaborative efficiency. Traditional task allocation methods rely primarily on managers' subjective judgment or static role divisions, making it difficult to dynamically adapt to task semantics and individual capabilities. This leads to frequent problems such as low task allocation efficiency, uneven team load, and collaboration mismatches. Summary of the Invention

[0003] The present invention provides a task bidirectional matching method, device, equipment and storage medium to solve the defects in the prior art.

[0004] The present invention provides a task bidirectional matching method, comprising: Acquire a task description text and a task member's capability semantic vector, and construct a task semantic vector for the task to be matched based on the task description text; Based on the capability semantic vector and the task semantic vector, determining candidate task members of the task description text and candidate tasks to be matched for the task members respectively; Task matching is performed based on the candidate task members and the candidate tasks to be matched.

[0005] According to a task bidirectional matching method provided by the present invention, constructing a task semantic vector of the task to be matched based on the task description text includes: Extracting multiple task elements from the task description text; Encode each of the task elements respectively to obtain a corresponding task code; According to each of the task codes, the task semantic vector is obtained.

[0006] According to a task bidirectional matching method provided by the present invention, determining candidate task members of the task description text and candidate matching tasks for the task members based on the capability semantic vector and the task semantic vector, respectively, includes: Calculating, based on the capability semantic vector and the task semantic vector, a first semantic matching degree, a collaboration compatibility degree, a skill fit degree, and a workload index between the task to be matched and each of the task members; Calculating a first target matching value between the task to be matched and each of the task members according to the semantic matching degree, the collaborative compatibility, the skill compatibility and the workload index; Determining the candidate task member for the task to be matched according to the first target matching value; Calculating, based on the capability semantic vector and the task semantic vector, a second semantic matching degree, a preference matching degree, and a task rhythm adaptability between the task member and each of the tasks to be matched; Calculating a second target matching value between the task member and each of the tasks to be matched according to the second semantic matching degree, the preference matching degree, and the task rhythm adaptability; The candidate to-be-matched tasks of the task member are determined according to the second target matching value.

[0007] According to a task bidirectional matching method provided by the present invention, after determining the candidate to-be-matched task of the task member according to the second target matching value, the method further includes: generating a first matching reason summary for each of the candidate task members based on the first target matching value; Based on the second target matching value, a second matching reason summary is generated for each of the candidate tasks to be matched.

[0008] According to a task bidirectional matching method provided by the present invention, before obtaining the task description text and the task member's capability semantic vector, the method further includes: Acquiring task behavior data of the task members, and encoding the task behavior data respectively to obtain corresponding behavior codes; The capability semantic vector is obtained according to the behavior coding.

[0009] According to a two-way task matching method provided by the present invention, task matching is performed based on the candidate task members and the candidate tasks to be matched, and the method further includes: Determining a target task member for the task to be matched based on the candidate task members and the candidate tasks to be matched; The execution risk of the target task member for the task to be matched is calculated, and the target task member and the execution risk are displayed on a preset display page.

[0010] According to a task bidirectional matching method provided by the present invention, after performing task matching based on the candidate task members and the candidate tasks to be matched, the method further includes: Acquiring task execution information after executing the task to be matched; wherein the task execution information includes member feedback information, collaboration score and task completion status; The matching weight function and the semantic vector bias are adjusted based on the task execution information feedback.

[0011] The present invention also provides a task bidirectional matching device, comprising the following modules: A construction module is configured to obtain a task description text and a capability semantic vector of a task member, and construct a task semantic vector of the task to be matched based on the task description text; a determination module configured to determine candidate task members of the task description text and candidate tasks to be matched for the task members based on the capability semantic vector and the task semantic vector; The task matching module is configured to perform task matching based on the candidate task members and the candidate tasks to be matched.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the two-way task matching method as described above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described task bidirectional matching methods.

[0014] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned task bidirectional matching methods.

[0015] The present invention provides a bidirectional task matching method, apparatus, device and storage medium, which obtains a task description text and a capability semantic vector of a task member, constructs a task semantic vector of the task to be matched based on the task description text, and determines candidate task members of the task description text and candidate tasks to be matched of the task member based on the capability semantic vector and the task semantic vector, respectively, and performs task matching based on the candidate task members and the candidate tasks to be matched. The present invention realizes bidirectional closed-loop adaptation of tasks through two adaptation paths, namely task-driven and member-driven, thereby improving the flexibility and generalization capability of task matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 It is a flowchart of the task bidirectional matching method provided by the present invention.

[0018] Figure 2It is a structural diagram of the task bidirectional matching device provided by the present invention.

[0019] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0020] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0021] Figure 1 FIG. 1 is a flowchart of a task bidirectional matching method according to an exemplary embodiment. Figure 1 As shown, in an exemplary embodiment, the task bidirectional matching method includes steps 110 to 130, which are described in detail as follows.

[0022] Step 110 : Obtain a task description text and a capability semantic vector of a task member, and construct a task semantic vector of the task to be matched based on the task description text.

[0023] In an embodiment of the present invention, a natural language task description text is received from a project platform or task management system. The task description text includes an unstructured description of the task objectives, background, execution requirements, time limits, delivery methods, etc. A task semantic vector for the task to be matched is constructed based on the task description text.

[0024] Step 120 : Based on the capability semantic vector and the task semantic vector, candidate task members of the task description text and candidate tasks to be matched with the task members are determined respectively.

[0025] In the embodiment of the present invention, based on the capability semantic vector and the task semantic vector, candidate task members of the task description text and candidate matching tasks of the task members are determined respectively. Candidate task members are task members that match the matching tasks, and candidate matching tasks are matching tasks that match the task members.

[0026] Step 130: Perform task matching based on the candidate task members and the candidate tasks to be matched.

[0027] In the embodiment of the present invention, task matching is performed based on candidate task members and candidate tasks to be matched. Two adaptation paths, task-driven and member-driven, are used to achieve bidirectional closed-loop adaptation of tasks, thereby improving the flexibility and generalization capability of task matching.

[0028] In an exemplary embodiment of the present invention, constructing a task semantic vector of the task to be matched based on the task description text includes: Extracting multiple task elements from the task description text; Encode each of the task elements respectively to obtain a corresponding task code; According to each of the task codes, the task semantic vector is obtained.

[0029] In this embodiment of the present invention, a large language model (LLM) is used to perform deep semantic understanding of task description text, extracting task elements of the task to be matched, including semantic factors such as skill requirements, domain, task complexity, task timeliness, and collaboration methods. After extraction, these task elements are uniformly encoded to generate a corresponding task encoding, which in turn yields a task semantic vector. This task semantic vector not only includes a semantic representation of the content but also incorporates task structure features and contextual preference information. This ensures that the task and member vectors are in a unified semantic embedding space, supports cross-task comparison and semantic reasoning, and provides a standardized, high-resolution task representation for subsequent matching.

[0030] Specifically, when constructing the task semantic vector, a multi-factor fusion task encoder is constructed, including an input layer, a first feature embedding layer, a first fusion layer and an output layer.

[0031] The input layer receives the task description text and uses a large language model (such as BERT, GPT, Deepseek-R1) to perform semantic understanding and feature extraction to obtain multiple task elements; The first feature embedding layer encodes the extracted semantic elements into vectors: BERT is used to convert skill requirements into corresponding encoding vectors; a trainable embedding matrix is ​​used to convert the corresponding domain into a corresponding encoding vector; discrete features such as task timeliness and task complexity are converted into one-hot encoding or embedding vectors; and assistance methods are converted into rule encoding or semantic vectors.

[0032] The first fusion layer is composed of a multilayer perceptron (MLP) or a transformer. After splicing the above embeddings, it is sent to a multilayer perceptron or a lightweight transformer to fuse semantic information of different dimensions and output a unified task semantic vector.

[0033] The output layer generates task semantic vectors (e.g., 768 dimensions), which is used to calculate semantic similarity with the capability semantic vector.

[0034] In the embodiment of the present invention, a large language model is used to perform semantic analysis and structural extraction on the task description text, and standardized semantic embedding is constructed by combining multi-dimensional task elements, breaking through the limitations of traditional task labeling processing.

[0035] In an exemplary embodiment of the present invention, before obtaining the task description text and the capability semantic vectors of the task members, the method further includes: Acquiring task behavior data of the task members, and encoding the task behavior data respectively to obtain corresponding behavior codes; The capability semantic vector is obtained according to the behavior coding.

[0036] In this embodiment of the present invention, a member capability modeling module is used to model the capabilities, collaboration styles, and accumulated experience of task members. This process is based on the collection and integration of multi-source heterogeneous data, including task member performance records, submitted outputs, discussion content, usage frequency, response time habits, and other task behavior data within the project system.

[0037] Using feature fusion and representation learning techniques, task members' multi-dimensional task behavior data is encoded to generate corresponding behavior codes, which in turn generate capability semantic vectors. The capability semantic vectors include elements such as knowledge domain proficiency, processing style preferences, task response cycles, and communication habits. Ultimately, these vectors are mapped into capability semantic representations, which reside in the same embedding space as the task semantic vectors to support subsequent semantic similarity calculations.

[0038] When constructing the capability semantic vector, a multi-source behavior representation fusion model (Behavior-Aware Embedding Model) is constructed. The multi-source behavior representation fusion model includes a data preprocessing layer, a second feature embedding layer, and a second fusion layer.

[0039] The data preprocessing layer normalizes and encodes the multi-source task behavior data collected from task members by dimension.

[0040] The feature embedding layer encodes various types of task behavior data into vectors in different ways: historical task categories are converted into task label embeddings; response time and response frequency are numerically normalized and converted into vector embeddings; text content (such as discussions) is converted into BERT encoding representations; and time distribution and style preferences are converted into one-hot or learnable embeddings.

[0041] The second fusion layer is composed of MLP or Transformer, which concatenates the above embeddings and sends them to the fusion network (such as multi-layer perceptron) to uniformly generate the ability semantic vector ,The capability semantic vector describes the member ’s capability, style, processing preference and load status. This vector not only reflects the ,static capability of the member but also includes its dynamic performance and ,collaboration habits.

[0042] The present invention constructs a multi-dimensional dynamic capability semantic vector by fusion modeling of the task behavior data of task members in an actual collaborative system, thereby realizing the evolution of capability representation from static role labels to dynamic semantic vectors.

[0043] In an exemplary embodiment of the present invention, determining candidate task members of the task description text and candidate tasks to be matched for the task members based on the capability semantic vector and the task semantic vector, respectively, includes: Calculating, based on the capability semantic vector and the task semantic vector, a first semantic matching degree, a collaboration compatibility degree, a skill fit degree, and a workload index between the task to be matched and each of the task members; Calculating a first target matching value between the task to be matched and each of the task members according to the semantic matching degree, the collaborative compatibility, the skill compatibility and the workload index; Determining the candidate task member for the task to be matched according to the first target matching value; Based on the capability semantic vector and the task semantic vector, calculating the second semantic matching degree, preference matching degree and task rhythm adaptability between the task member and each of the tasks to be matched; Calculating a second target matching value between the task member and each of the tasks to be matched according to the second semantic matching degree, the preference matching degree, and the task rhythm adaptability; The candidate to-be-matched tasks of the task member are determined according to the second target matching value.

[0044] In the embodiment of the present invention, after obtaining the task semantic vector and the capability semantic vector, a semantic matching engine is called to perform a bidirectional matching calculation, where the bidirectional matching includes a task-driven path and a member-driven path.

[0045] For the task-driven path, the task to be matched is used as the query center. The first semantic match is calculated with the capability semantic vectors of all task members in sequence. The first target match value is then calculated by combining weighted factors such as collaboration compatibility, skill fit, and workload indicators. This is then used to determine candidate task members. Collaboration compatibility refers to the degree of compatibility between task members in collaboratively completing the task to be matched. Skill fit refers to the degree to which task members can use their skills to complete the task to be matched. The workload indicator refers to the workload that task members are expected to bear for the matching task.

[0046] Specifically, the encoded task semantic vector and capability semantic vector set { , ,…, } n capability semantic vectors in Perform matching calculations.

[0047] Use cosine similarity or vector distance function (such as dot-product, Euclidean) to calculate the first semantic matching degree, collaborative compatibility, skill fit and workload index of the task to be matched and each task member, and integrate multiple weight factors for weighted summation to obtain the first target matching value Score i , such as the first target matching value Score i It can be expressed as: Score i =cos( )+α×collaboration compatibility+β×skill fit+γ×workload index; among them, cos( ) represents the first semantic matching degree between the task to be matched and the task member i calculated based on cosine similarity.

[0048] After obtaining the first target matching value of the task to be matched and each task member, the top-K task members are selected as candidate task members for the task to be matched.

[0049] For the member-driven path, with the task members as the center, identify candidate tasks to be matched that match the task members' current ability status, preference tags, and rhythm matching.

[0050] Capability semantic vector of task members With the task semantic vector set { , ,…, }m task semantic vectors in Perform matching calculations.

[0051] Specifically, the cosine similarity method is also used to calculate the second semantic matching degree, preference matching degree and task rhythm adaptation degree between the task member and each task to be matched. The preference matching degree refers to the matching degree between the task preference of the task member and the task to be matched. The task rhythm adaptation degree refers to the matching degree between the task rhythm of the task member and the task rhythm of the task to be matched. The second target matching value Score is obtained by integrating multiple weight factors for weighted summation. j , such as the second target matching value Score j It can be expressed as: Score j =cos( )+δ×preference matching+ε×task rhythm adaptation; among them, cos( ) represents the second semantic matching degree between the task member and the task to be matched j calculated based on cosine similarity.

[0052] After obtaining the second target matching value of the task member and each task to be matched, the top-K tasks to be matched are selected as candidate task members of the task member.

[0053] In an exemplary embodiment of the present invention, after determining the candidate to-be-matched tasks of the task member according to the second target matching value, the method further includes: generating a first matching reason summary for each of the candidate task members based on the first target matching value; Based on the second target matching value, a second matching reason summary is generated for each of the candidate tasks to be matched.

[0054] In this embodiment of the present invention, a first matching reason summary is generated for each candidate task member based on the first target matching value. The first matching reason summary is a matching explanation for each candidate task member. A second matching reason summary is generated for each candidate task to be matched based on the second target matching value. The second matching reason summary is a matching explanation for each candidate task to be matched, such as Task A involves a client system you are familiar with, and the work pace is adapted to the current schedule.

[0055] In an exemplary embodiment of the present invention, the method of performing task matching based on the candidate task members and the candidate tasks to be matched further includes: Determining a target task member for the task to be matched based on the candidate task members and the candidate tasks to be matched; The execution risk of the target task member for the task to be matched is calculated, and the target task member and the execution risk are displayed on a preset display page.

[0056] In an embodiment of the present invention, target task members for the task to be matched are determined based on candidate task members and candidate tasks to be matched. The target task members are the task members who will execute the task to be matched. The execution risk of the target task members for the task to be matched is calculated, and the target task members and the execution risk are displayed on a preset display page. The execution risk is determined by quantitatively modeling and weighted scoring key risk factors in the matching process (such as member load, skill gaps, conflicting collaborative styles, and similarity of historical failures). A comprehensive assessment of the execution risk that may arise from the matching results is then made and output in the form of "risk level + explanation of reasons," assisting project managers in making more reliable task allocation decisions.

[0057] The matching calculation results are structured and output, including: a list of candidate task members / candidate matching tasks, a first target matching value / second target matching value, a summary of the first matching reason / second matching reason, a matching explanation summary (e.g., "Matching Skills: Report Analysis; Matching Style: Document-Driven"), and execution risks. The results can be used directly by project managers or automatically synchronized to the task platform for automated assignment.

[0058] In addition, it can also support interface docking with third-party systems such as schedule management, communication tools, and work order platforms to achieve full-process automation from matching to task issuance, notification, and collaboration initiation, and build an intelligent, human-machine collaborative modern team task management system.

[0059] The matching results of the present invention include not only ranking, but also explanation of semantic matching factors, risk warnings and recommendation reasons. It supports docking with third-party systems to achieve automatic task allocation and semantic transparent collaboration.

[0060] In an exemplary embodiment of the present invention, after performing task matching based on the candidate task members and the candidate tasks to be matched, the method further includes: Obtaining task execution information after executing the task to be matched; wherein the task execution information includes member feedback information, collaboration score and task completion status; The matching weight function and the semantic vector bias are adjusted based on the task execution information feedback.

[0061] In an embodiment of the present invention, in order to improve the robustness and adaptability during long-term use, the matching engine includes a strategy adjustment and self-learning mechanism. After each round of matching, the matching execution results, member feedback information, collaboration scores, task completion status and other information will be recorded to form a feedback loop. In subsequent matching, the matching weight function and semantic vector bias can be dynamically adjusted based on the above-mentioned feedback task execution information, such as improving the matching priority of a task member in a specific field task, or adjusting the task co-occurrence probability between members with incompatible styles. The strategy adjustment module supports user preference guidance and organizational strategy injection, so that matching has controllability and evolution capabilities on an intelligent basis.

[0062] The matching weight function refers to the weighted proportion of each influencing factor (such as semantic matching, collaborative compatibility, skill fit, etc.) when calculating the matching degree between the task to be matched and the task members. These weights can be automatically adjusted according to the feedback task execution information to optimize the final target matching value.

[0063] Semantic vector bias refers to fine-tuning or offsetting the vectors based on historical performance when generating or comparing task semantic vectors and ability semantic vectors. For example, the representation vector of a task member on a certain type of task can be fine-tuned to improve its matching priority or reduce the impact of style conflicts, thereby better aligning with the actual collaboration effect.

[0064] Task co-occurrence probability refers to the historical probability of multiple tasks to be matched being assigned and successfully completed simultaneously to the same task member or member combination. It essentially reflects the feasibility and compatibility of collaborative completion between tasks in collaborative practice. Optimizing batch task allocation strategies based on task co-linearity probability can prioritize assigning tasks to be matched with high task co-occurrence probability to the same task member or member combination, improving execution efficiency and reducing context switching costs. Avoiding style conflicts and rhythm mismatches based on task co-linearity probability: if a task member has historically failed to handle two types of tasks simultaneously, a low task co-occurrence probability can be used as "negative feedback" to avoid re-combination. Task co-linearity probability supports task combination recommendations and scheduling decisions. It not only assigns single tasks but also recommends "task packages" or multi-task flows, enhancing scheduling intelligence and collaborative optimization.

[0065] The present invention supports the introduction of collaborative preference factors and member style information as matching auxiliary items, and dynamically adjusts the matching function parameters based on historical completion quality, feedback satisfaction and other data, and has the ability of continuous learning and personalized adaptation.

[0066] Based on the existing task matching, the present invention comprehensively enhances the semantic understanding and intelligent scheduling capabilities. Compared with the traditional method based on role assignment or static label matching, the present invention understands the task semantics through a large model and combines the dynamic capability semantic vector for deep semantic matching, which significantly improves the accuracy of task-member matching and the actual execution effect.

[0067] The present invention can dynamically adjust the allocation results according to the workload, style adaptability and execution risk estimation of task members, avoid problems such as task backlog and resource waste, and achieve an overall improvement in team efficiency.

[0068] The present invention provides structured matching output and matching reason explanation, which helps project managers understand the recommendation mechanism and improve users' trust and adoption rate of artificial intelligence matching results.

[0069] The task-member bidirectional semantic matching mechanism constructed by the present invention has strong versatility and is suitable for various task-intensive collaboration scenarios such as enterprise project management, human resources systems, education and teaching division of labor platforms, open source community collaboration tools, etc., and has broad promotion value.

[0070] The following describes the task bidirectional matching device provided by the present invention. The task bidirectional matching device described below can be referenced in conjunction with the task bidirectional matching method described above. It should be noted that the device provided in the following embodiments and the method provided in the above embodiments are based on the same concept. The specific manner in which each module and unit performs its operations has been described in detail in the method embodiments and will not be repeated here.

[0071] In an exemplary embodiment of the present invention, see Figure 2 , Figure 2 A task bidirectional matching device according to an exemplary embodiment includes the following modules.

[0072] A construction module 210 is configured to obtain a task description text and a capability semantic vector of a task member, and construct a task semantic vector of the task to be matched based on the task description text; A determination module 220 is configured to determine candidate task members of the task description text and candidate tasks to be matched with the task members based on the capability semantic vector and the task semantic vector, respectively; The task matching module 230 is configured to perform task matching based on the candidate task members and the candidate tasks to be matched.

[0073] In an exemplary embodiment of the present invention, the building block 210 includes: an extraction submodule configured to extract a plurality of task elements from the task description text; The encoding submodule is configured to encode each of the task elements to obtain a corresponding task code; The semantic vector submodule is configured to obtain the task semantic vector according to each task encoding.

[0074] In an exemplary embodiment of the present invention, the determination module 220 includes: A first calculation submodule is configured to calculate a first semantic matching degree, a collaboration compatibility degree, a skill fit degree, and a workload index of the task to be matched and each of the task members based on the capability semantic vector and the task semantic vector; A second calculation submodule is configured to calculate a first target matching value between the task to be matched and each of the task members according to the semantic matching degree, the collaboration compatibility, the skill fit and the workload index; A first determining submodule is configured to determine the candidate task member of the task to be matched according to the first target matching value; A third calculation submodule is configured to calculate the second semantic matching degree, preference matching degree and task rhythm adaptability of the task member and each of the tasks to be matched based on the ability semantic vector and the task semantic vector; a fourth calculation submodule, configured to calculate a second target matching value between the task member and each of the tasks to be matched according to the second semantic matching degree, the preference matching degree, and the task rhythm adaptability; The second determining submodule is configured to determine the candidate to-be-matched tasks of the task member according to the second target matching value.

[0075] In an exemplary embodiment of the present invention, the task bidirectional matching device further includes: a first generating module configured to generate a first matching reason summary for each candidate task member based on the first target matching value; The second generating module is configured to generate a second matching reason summary for each of the candidate tasks to be matched based on the second target matching value.

[0076] In an exemplary embodiment of the present invention, the task bidirectional matching device further includes: an encoding module configured to obtain task behavior data of the task members and encode the task behavior data respectively to obtain corresponding behavior codes; The semantic vector module is configured to obtain the capability semantic vector according to the behavior coding.

[0077] In an exemplary embodiment of the present invention, the task matching module 230 includes: A third determining submodule is configured to determine a target task member of the task to be matched based on the candidate task members and the candidate tasks to be matched; The fifth calculation submodule is configured to calculate the execution risk of the target task member for the task to be matched, and display the target task member and the execution risk on a preset display page.

[0078] In an exemplary embodiment of the present invention, the task bidirectional matching device further includes: An acquisition module configured to acquire task execution information after executing the task to be matched; wherein the task execution information includes member feedback information, collaboration score and task completion status; An adjustment module is configured to adjust a matching weight function and a semantic vector bias based on the task execution information feedback.

[0079] Figure 3 An example of a physical structure diagram of an electronic device is shown below. Figure 3As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other via the communication bus 340. The processor 310 may call logic instructions in the memory 330 to execute a task bidirectional matching method, which includes: obtaining a task description text and a capability semantic vector of a task member, and constructing a task semantic vector for the task to be matched based on the task description text; Based on the capability semantic vector and the task semantic vector, determining candidate task members of the task description text and candidate tasks to be matched for the task members respectively; Task matching is performed based on the candidate task members and the candidate tasks to be matched.

[0080] Furthermore, the logic instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0081] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being storable on a non-transitory computer-readable storage medium, and the computer being capable of executing the task bidirectional matching method provided by the above methods when the computer program is executed by a processor, the method including: obtaining a task description text and a capability semantic vector of a task member, and constructing a task semantic vector of the task to be matched based on the task description text; Based on the capability semantic vector and the task semantic vector, determining candidate task members of the task description text and candidate tasks to be matched for the task members respectively; Task matching is performed based on the candidate task members and the candidate tasks to be matched.

[0082] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the task bidirectional matching method provided by the above methods, the method comprising: obtaining a task description text and a capability semantic vector of a task member, and constructing a task semantic vector of the task to be matched based on the task description text; Based on the capability semantic vector and the task semantic vector, determining candidate task members of the task description text and candidate tasks to be matched for the task members respectively; Task matching is performed based on the candidate task members and the candidate tasks to be matched.

[0083] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0084] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A task two-way matching method, characterized in that: include: Acquire a task description text and a task member's capability semantic vector, and construct a task semantic vector for the task to be matched based on the task description text; Based on the capability semantic vector and the task semantic vector, determining candidate task members of the task description text and candidate tasks to be matched for the task members respectively; Task matching is performed based on the candidate task members and the candidate tasks to be matched.

2. The task two-way matching method according to claim 1, characterized in that: The step of constructing a task semantic vector for the task to be matched based on the task description text includes: Extracting multiple task elements from the task description text; Encode each of the task elements respectively to obtain a corresponding task code; According to each of the task codes, the task semantic vector is obtained.

3. The task two-way matching method according to claim 1, characterized in that: The determining, based on the capability semantic vector and the task semantic vector, candidate task members of the task description text and candidate tasks to be matched for the task members, respectively, includes: Calculating, based on the capability semantic vector and the task semantic vector, a first semantic matching degree, a collaboration compatibility degree, a skill fit degree, and a workload index between the task to be matched and each of the task members; Calculating a first target matching value between the task to be matched and each of the task members according to the semantic matching degree, the collaborative compatibility, the skill compatibility and the workload index; Determining the candidate task member for the task to be matched according to the first target matching value; Based on the capability semantic vector and the task semantic vector, calculating the second semantic matching degree, preference matching degree and task rhythm adaptability between the task member and each of the tasks to be matched; Calculating a second target matching value between the task member and each of the tasks to be matched according to the second semantic matching degree, the preference matching degree, and the task rhythm adaptability; The candidate to-be-matched tasks of the task member are determined according to the second target matching value.

4. The task two-way matching method according to claim 3, characterized in that: After determining the candidate to-be-matched tasks of the task member according to the second target matching value, the method further includes: generating a first matching reason summary for each of the candidate task members based on the first target matching value; Based on the second target matching value, a second matching reason summary is generated for each of the candidate tasks to be matched.

5. The task two-way matching method according to claim 1, characterized in that: Before obtaining the task description text and the task member's capability semantic vector, the method further includes: Acquiring task behavior data of the task members, and encoding the task behavior data respectively to obtain corresponding behavior codes; The capability semantic vector is obtained according to the behavior coding.

6. The task two-way matching method according to claim 1, characterized in that: The performing task matching according to the candidate task members and the candidate tasks to be matched includes: Determining a target task member for the task to be matched based on the candidate task members and the candidate tasks to be matched; The execution risk of the target task member for the task to be matched is calculated, and the target task member and the execution risk are displayed on a preset display page.

7. The task bidirectional matching method according to any one of claims 1 to 6, characterized in that: After performing task matching based on the candidate task members and the candidate tasks to be matched, the method further includes: Obtaining task execution information after executing the task to be matched; wherein the task execution information includes member feedback information, collaboration score and task completion status; The matching weight function and the semantic vector bias are adjusted based on the task execution information feedback.

8. A task two-way matching device, characterized in that: include: A construction module is configured to obtain a task description text and a capability semantic vector of a task member, and construct a task semantic vector of the task to be matched based on the task description text; a determination module configured to determine candidate task members of the task description text and candidate tasks to be matched for the task members based on the capability semantic vector and the task semantic vector; The task matching module is configured to perform task matching based on the candidate task members and the candidate tasks to be matched.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the task bidirectional matching method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the task bidirectional matching method according to any one of claims 1 to 7 is implemented.

Images