CN-121998281-A - Human work auxiliary method and system based on artificial intelligence

CN121998281ACN 121998281 ACN121998281 ACN 121998281ACN-121998281-A

Abstract

The invention relates to an artificial intelligence-based human work auxiliary method and system, wherein the method comprises the following steps of obtaining multi-mode data input by a user, preprocessing, carrying out preliminary task disassembly, calculating logic dependency intensity, granularity adaptation degree and disassembly priority, carrying out task disassembly optimization, generating a subtask list, respectively generating a human capability vector and an artificial intelligence capability vector based on capability portraits, generating a subtask skill requirement vector based on the subtask list, calculating an allocation utility function constructed by skill adaptation degree, load penalty items and risk rewarding items, carrying out human-machine task allocation on each subtask in the subtask list under the constraint of allocation decision rule, and dynamically adjusting the human-machine task allocation by monitoring the execution progress and execution quality of the tasks in real time in the auxiliary work task execution process. Compared with the prior art, the invention has the advantages of realizing effective disassembly and distribution of complex tasks, further improving the auxiliary efficiency of artificial intelligence and the like.

Inventors

WU JIAHUI
YIN CHENGLIANG
HUANG HAOFENG
GAO LIANGQUAN

Assignees

上海智能网联汽车技术中心有限公司

Dates

Publication Date: 20260508
Application Date: 20251205

Claims (10)

1. A human work assisting method based on artificial intelligence, the method comprising the steps of: Acquiring multi-mode data input by a user, preprocessing the multi-mode data, and generating structured task data; Performing primary task disassembly on the structured task data to generate an initial subtask set, calculating the logic dependency strength, granularity adaptation degree and disassembly priority of each task in the initial subtask set, performing task disassembly optimization, and generating a subtask list; performing human capability image quantization, respectively generating a human capability vector and an artificial intelligence capability vector, generating a subtask skill demand vector based on a subtask list, calculating skill adaptation degree based on the matching degree between the capability vector and the subtask skill demand vector, constructing an allocation utility function by combining the skill adaptation degree, a load penalty term and a risk rewarding term, and performing human-machine task allocation on each subtask in the subtask list under the constraint of meeting allocation decision rules based on the allocation utility function; In the process of executing the auxiliary work task, the execution progress and the execution quality of the task are monitored in real time, and the man-machine task distribution is dynamically adjusted.
2. The artificial intelligence based human work support method according to claim 1, wherein the calculation of the logic dependency strength comprises the steps of: For any of the initial subtasks in the initial subtask set Extracting its direct pre-subtasks from the initial subtask set and determining the number of pre-subtasks ; Constructing a Bayesian network, calculating subtasks according to the front subtasks based on dependency execution data of similar historical tasks Is dependent on the trigger probability of (2) ; Based on the number of the front-end subtasks Relying on trigger probability Calculating the logic dependency strength: , Wherein, the For subtasks Is a function of the strength of the logical dependence of (a), Is the total number of subtasks in the initial subtask set.
3. The artificial intelligence based human work support method according to claim 1, wherein the calculation of the granularity adaptation degree comprises the steps of: based on historical task data, K-means clustering is adopted to obtain a time-consuming interval capable of being efficiently completed in a single role, and the median of the time-consuming interval is taken as an optimal time-consuming threshold value ; Calculating subtask estimated time consumption based on historical execution time data of similar tasks ; Calculating granularity adaptation degree based on subtask estimated time consumption and optimal time consumption threshold values: , Wherein, the For subtasks Granularity adaptation of (c).
4. The human work assisting method based on artificial intelligence according to claim 1, wherein the calculation mode of the disassembly priority is as follows: , Wherein, the 、 As the weight of the material to be weighed, For subtasks Is a function of the strength of the logical dependence of (a), For subtasks Is used for the degree of granularity adaptation of the (a), For subtasks Is to be determined.
5. The human work assisting method based on artificial intelligence according to claim 1, wherein the task disassembly optimization specifically comprises the following steps: When the disassembly priority is greater than or equal to a preset priority threshold, not adjusting the subtasks corresponding to the disassembly priority; When the disassembly priority is smaller than a preset priority threshold, further judging the relation between the logic dependency strength and the corresponding threshold of the granularity adaptation degree: if the logic dependency strength is smaller than the first preset threshold value but the granularity adaptation degree is larger than or equal to the second preset threshold value, merging similar subtasks, recalculating the granularity adaptation degree and the disassembly priority, or adjusting the task dependency relationship, and recalculating the logic dependency strength and the disassembly priority; If the logic dependency strength is greater than or equal to a first preset threshold value and the granularity adaptation degree is smaller than a second preset threshold value, performing granularity calibration, performing subtask secondary splitting when granularity is too coarse, merging similar subtasks when granularity is too fine, and recalculating the logic dependency strength, the granularity adaptation degree and the disassembly priority; If the logic dependency strength is smaller than the first preset threshold value and the granularity adaptation degree is smaller than the second preset threshold value, reconstructing the subtask, deleting the redundant subtask, and recalculating the logic dependency strength, the granularity adaptation degree and the disassembly priority; If the logic dependency strength is greater than or equal to a first preset threshold value and the granularity adaptation degree is greater than or equal to a second preset threshold value, the corresponding subtasks are not adjusted.
6. The artificial intelligence based human work support method according to claim 1, wherein the calculation of skill fitness comprises the steps of: performing human capability portrait quantization to respectively generate a human capability vector and an artificial intelligence capability vector, wherein each element in the human capability vector and the artificial intelligence capability vector respectively represents a skill value of human and artificial intelligence corresponding to a skill; Marking a core skill requirement for each subtask in a subtask list, and generating a subtask skill requirement vector, wherein each element in the subtask skill requirement vector is a skill requirement indication value, the skill requirement indication value is 1 to indicate that the subtask needs the skill, and the skill requirement indication value is 0 to indicate that the subtask does not need the skill; After carrying out normalization processing based on skill types on the human ability vector and the artificial intelligence ability vector, calculating skill probabilities corresponding to each type of skill; Calculating skill importance weights based on the skill probabilities and information entropy; Multiplying the skill importance weight with a skill demand indication value of the corresponding skill in the subtask skill demand vector to obtain a normalized skill importance weight; And multiplying and summing the normalized skill importance weight with corresponding skills in the human ability vector and the artificial intelligence ability vector respectively to obtain the human skill adaptation degree and the artificial intelligence skill adaptation degree.
7. The artificial intelligence based human work support method according to claim 1, wherein the distribution utility function is expressed as: , Wherein, the Representing the execution subject of the task, which is human or artificial intelligence; representing subtasks The execution subject of (2) is Degree of skill adaptation at the time; 、 representing the adjustment coefficient; Representing execution subject Is used for the load of the (c), Penalty term for load; For subtasks The execution subject of (2) is The risk coefficient is determined according to the corresponding risk level, Items are awarded for risk.
8. The artificial intelligence based human work support method according to claim 1, wherein the allocation decision rule comprises: a main body with the largest allocation utility function value is selected as an execution role; and constraint rules, namely when the risk level is greater than a preset level, if the main body with the largest assigned utility function value is artificial intelligence, forcibly converting the main body into human beings.
9. The human work assisting method based on artificial intelligence according to claim 1, wherein the real-time monitoring of the execution progress and execution quality of the task dynamically adjusts the human-machine task allocation, comprising the steps of: in the task execution process, calculating a progress index and a quality index in real time, and calculating a progress-quality coupling coefficient based on the progress index and the quality index; And when the progress-quality coupling coefficient is greater than or equal to an adjustment threshold, updating the human capacity vector and the artificial intelligence capacity vector, and carrying out human-computer task allocation again, wherein the adjustment threshold is determined based on the risk level.
10. An artificial intelligence based human work assistance system, the system comprising: The data acquisition and preprocessing module is used for acquiring multi-mode data input by a user, preprocessing the multi-mode data and generating structured task data; the task disassembly module is used for carrying out primary task disassembly on the structured task data to generate an initial subtask set, calculating the logic dependency strength, granularity adaptation degree and disassembly priority of each task in the initial subtask set, carrying out task disassembly optimization, and generating a subtask list; The task allocation module is used for carrying out human capability image quantization, respectively generating a human capability vector and an artificial intelligence capability vector, generating a subtask skill demand vector based on a subtask list, calculating skill adaptation degree based on the matching degree between the capability vector and the subtask skill demand vector, constructing an allocation utility function by combining the skill adaptation degree, a load penalty item and a risk rewarding item, and carrying out human-machine task allocation on each subtask in the subtask list under the constraint of meeting allocation decision rules based on the allocation utility function; And the real-time adjustment module is used for monitoring the execution progress and the execution quality of the task in real time in the execution process of the auxiliary work task and dynamically adjusting the human-machine task allocation.

Description

Human work auxiliary method and system based on artificial intelligence Technical Field The invention relates to the technical field of AI auxiliary application, in particular to a human work auxiliary method and system based on artificial intelligence. Background Currently, a plurality of artificial intelligence auxiliary tools exist at home and abroad and cover a plurality of fields of office work, medical treatment, education, industrial manufacturing and the like. For example, an intelligent document processing tool (such as an AI text recognition and editing function of Adobe Acrobat) in an office scene can automatically extract document key information, an AI auxiliary diagnosis system in the medical field (such as a hundred-degree medical brain) can primarily recognize focuses based on medical images, AI quality inspection equipment in industrial manufacturing can detect product surface defects through image recognition, and AI coaching software in the education field (such as a simian coaching AI answering system) can provide question analysis and knowledge point explanation for students. In addition, general AI assistants (e.g., chatGPT, confucius) can provide basic assistance services such as text generation, information query, logical reasoning, etc. However, the existing tool can only process single-link and low-logic-associated work tasks, and cannot cope with multi-link and high-logic-complexity work flows. For example, in a project management scene, the existing AI tool can remind the task deadline, but can not automatically adjust the task allocation scheme according to the workload of team members and the task dependency relationship (if the task A is not completed and the task B is not started), and the AI text tool in the legal field can search legal strips but can not combine the evidence chain and court trial disputed focus of a specific case to generate a targeted dialect and thinking frame. This is essentially because AI is difficult to effectively disassemble complex tasks and thus also difficult to effectively handle. Chinese patent CN120996543a discloses a man-machine co-creation enterprise business process management system, which is used for solving the problems of unreasonable task allocation, insufficient process monitoring and difficult knowledge multiplexing of the traditional system. The system comprises six modules, namely task triggering and analyzing, disassembling and planning, workflow generating and optimizing. The task triggering and analyzing module is used for acquiring and analyzing the task, the task is disassembled by utilizing various strategies, the workflow is determined through multi-objective optimization, the man-machine execution of the subtasks is coordinated, the flow is monitored in real time and dynamically adjusted, and meanwhile, the AI capacity of the business data optimization is collected and the enterprise knowledge is accumulated. However, the task disassembly machine of the method lacks quantization logic, is only distributed through the efficiency index of a single dimension, and cannot be accurately adapted to the task with high logic complexity. Disclosure of Invention The invention aims to overcome the defects of the prior art and provide an artificial intelligence-based human work assisting method and system. The aim of the invention can be achieved by the following technical scheme: according to a first aspect of the present invention, there is provided an artificial intelligence based human work assistance method comprising the steps of: Acquiring multi-mode data input by a user, preprocessing the multi-mode data, and generating structured task data; Performing primary task disassembly on the structured task data to generate an initial subtask set, calculating the logic dependency strength, granularity adaptation degree and disassembly priority of each task in the initial subtask set, performing task disassembly optimization, and generating a subtask list; performing human capability image quantization, respectively generating a human capability vector and an artificial intelligence capability vector, generating a subtask skill demand vector based on a subtask list, calculating skill adaptation degree based on the matching degree between the capability vector and the subtask skill demand vector, constructing an allocation utility function by combining the skill adaptation degree, a load penalty term and a risk rewarding term, and performing human-machine task allocation on each subtask in the subtask list under the constraint of meeting allocation decision rules based on the allocation utility function; In the process of executing the auxiliary work task, the execution progress and the execution quality of the task are monitored in real time, and the man-machine task distribution is dynamically adjusted. The calculation of the logic dependency strength comprises the following steps: For any of the initial subtasks in the initial subtask