Remote guidance and real-time feedback system for continued education of resident physicians

Abstract

The invention relates to the technical field of medical education, and discloses a remote guidance and real-time feedback system for continued education of an inpatient physician, and the system comprises a learning task management module which is used for managing and recommending learning tasks of the inpatient physician, and comprises the steps: dynamically adjusting the priority and difficulty of the learning tasks; the real-time feedback and evaluation module is used for evaluating the learning performance, the operation precision and the clinical decision of the resident physician in real time and providing immediate feedback; and the multi-objective optimization and path adjustment module is used for optimizing the learning path according to the learning objective and progress of the physician and dynamically adjusting the task sequence and difficulty. Through reinforcement learning, adaptive path adjustment and game theory optimization, learning paths, task recommendation and teaching strategies of resident doctors are dynamically adjusted, personalized education, intelligent feedback and accurate optimization are realized, and the learning efficiency and the education effect are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of medical education technology, and in particular to a remote guidance and real-time feedback system for continuing education of resident physicians. Background Art

[0002] With the continuous progress and development of the medical field, the training and continuing education of resident doctors have become increasingly important. Resident doctors not only need to master a wealth of medical knowledge, but also must have superb clinical operation skills and keen decision-making ability. Therefore, how to effectively improve the comprehensive ability of resident doctors has become a challenge that needs to be solved urgently.

[0003] In the existing technology, many resident education systems adopt a fixed mode based on task push and performance evaluation. These systems can perform basic task management and learning progress tracking by evaluating and recording the learning data of residents. For example, the task management module can provide a series of task recommendations based on the learning progress of residents, and the learning management system can also record and summarize the performance of residents. Such systems have certain advantages in basic education, and can effectively arrange tasks and record basic data, helping residents to complete their learning goals step by step.

[0004] Although existing technologies can provide certain educational support, there are still some shortcomings. First, traditional education models usually ignore the learning differences of residents. Task push and recommendation mostly rely on preset rules, which are difficult to adjust according to the real-time performance and feedback of residents. Secondly, the static task recommendation model adopted by most existing technologies makes it impossible for doctors to get targeted learning tasks or feedback in time when they encounter bottlenecks. In addition, existing systems usually lack in-depth analysis of residents' behavior, cannot fully identify their weak links at different learning stages, and cannot achieve adaptive task difficulty adjustment. Finally, traditional systems also lack game optimization in the interaction between doctors and mentors and colleagues, resulting in the inability of teaching interaction and learning path selection to fully stimulate the learning potential of doctors. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a remote guidance and real-time feedback system for resident continuing education, which solves the problem of lack of personalization, dynamic adjustment of learning paths and task recommendations in the prior art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a remote guidance and real-time feedback system for continuing education of resident physicians, comprising: A learning task management module is used to manage and recommend learning tasks for residents, including dynamically adjusting the priority and difficulty of learning tasks; Real-time feedback and evaluation module, which is used to evaluate residents' learning performance, operation accuracy and clinical decision-making in real time and provide immediate feedback; Multi-objective optimization and path adjustment module, which is used to optimize the learning path according to the physician's learning goals and progress, and dynamically adjust the task sequence and difficulty; Game theory and strategy optimization module, which is used to simulate the decision-making interaction between residents and mentors, and between physicians, and optimize task allocation and learning strategies; Reinforcement learning and adaptive pathway adjustment module, used to dynamically adjust learning pathways and task allocation based on residents’ feedback and performance; The data collection and analysis module is used to collect and analyze physicians’ learning data to support the optimization and decision-making of other modules.

[0007] Preferably, the learning task management module includes: A task library that stores all available learning tasks and generates task recommendations based on residents’ progress and needs; Dynamic task push unit, used to adjust the priority and difficulty of tasks based on physician feedback and performance; The task generation unit is used to generate new tasks or adjust existing tasks based on the real-time performance of residents.

[0008] Preferably, the real-time feedback and evaluation module includes: A clinical decision-making assessment unit to assess residents’ decision-making process in complex cases and provide feedback; A feedback generation unit is used to generate feedback based on the evaluation results and provide feedback to the resident physician via text, voice or image.

[0009] Preferably, the multi-objective optimization and path adjustment module includes: The objective function calculation unit is used to calculate multiple objective functions of residents in terms of knowledge mastery, operational skills, and clinical decision-making ability; An optimization algorithm unit, used to optimize the learning path, task sequence and difficulty according to the objective function results; The path adjustment unit is used to dynamically adjust the learning path according to the results of the optimization algorithm.

[0010] Preferably, the game theory and strategy optimization module includes: A strategic decision-making unit, which is used to simulate the strategic interaction between residents and mentors and optimize task allocation; The task allocation unit is used to optimize task allocation through game theory models to make the interaction between residents and mentors more efficient; The strategy updating unit is used to update the task allocation strategy and education strategy according to the changes in the game process.

[0011] Preferably, the reinforcement learning and adaptive path adjustment module includes: A reinforcement learning engine is used to optimize the learning path based on the physicians’ learning feedback and gradually improve the learning efficiency; A task adjustment unit, which is used to dynamically adjust the difficulty and sequence of tasks according to the residents’ learning situation; The adaptive adjustment unit is used to adjust task allocation and learning progress according to real-time feedback.

[0012] Preferably, the data collection and analysis module includes: Data collection unit, used to collect residents’ learning progress, test scores, and operation data in real time; Data analysis engine, used to conduct in-depth analysis of collected learning data and discover trends and potential problems in learning; The report generation unit is used to generate personalized learning reports for doctors and feed back the analysis results to the learning task management module and optimization module.

[0013] Preferably, the real-time feedback and evaluation module further comprises: The Virtual Patient Simulation Unit is used to simulate virtual patients and provide realistic clinical feedback to help residents train their skills and decision-making; The simulation feedback generation unit is used to generate feedback and suggestions according to the physician's operations during the virtual patient simulation process.

[0014] Preferably, the learning task management module further includes: Task priority management unit, used to dynamically adjust the priority of tasks according to the importance of tasks, physicians’ learning progress and clinical needs; The Task Adaptability Assessment Unit is used to evaluate residents’ performance on different tasks, ensuring that the tasks are always adapted to the physician’s learning ability.

[0015] Preferably, the game theory and strategy optimization module further includes: Multi-physician task coordination unit, used to simulate cooperation and competition among multiple physicians and optimize task allocation strategy according to game results; The decision optimization engine is used to optimize the strategies between physicians and tutors in real time based on the results of game theory analysis, ensuring optimal task allocation and interaction during the education process.

[0016] The present invention provides a remote guidance and real-time feedback system for continuing education of resident physicians. It has the following beneficial effects: 1. Through reinforcement learning and adaptive path adjustment, the present invention can dynamically adjust the learning path and task difficulty according to the resident's real-time feedback and learning progress, ensuring that the resident is always in the most suitable learning state. This flexible way of adjusting tasks is much smarter than the static recommendation system in the prior art, avoiding the singleness of learning tasks and solving the problem of insufficient adjustment of personalized learning paths.

[0017] 2. Game theory and strategy optimization modules simulate the interaction between residents and mentors, optimize task allocation and teaching strategies, and make the teaching process more efficient and targeted. This interactive optimization method greatly improves the quality of cooperation between physicians and mentors, while the traditional one-way teaching model often ignores the dynamic nature of feedback and cannot achieve real-time adjustment.

[0018] 3. Through the data collection and analysis module, the learning data of residents is deeply analyzed, learning bottlenecks are identified in real time, and personalized optimization suggestions are provided based on the analysis results to help doctors quickly discover and make up for shortcomings in the learning process. This precise analysis based on big data breaks the limitations of insufficient learning data collection and feedback analysis in the traditional education model, and greatly improves learning efficiency.

[0019] 4. Through multi-objective optimization and path adjustment, the system can not only ensure the balanced development of physicians in terms of knowledge, operation and clinical decision-making, but also dynamically optimize according to different learning goals and progress. This multi-dimensional goal optimization strategy is more comprehensive than the traditional education model that only focuses on a single goal, and solves the problem of resident physicians' possible skill bias or neglect. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a system structure diagram of the present invention; Figure 2 The module architecture diagram of the data collection and analysis module of the present invention; Figure 3 This is a module architecture diagram of the learning task management module of the present invention; Figure 4 The module architecture diagram of the real-time feedback and evaluation module of the present invention; Figure 5 This is a module architecture diagram of the multi-objective optimization and path adjustment module of the present invention; Figure 6 This is a module architecture diagram of the game theory and strategy optimization module of the present invention; Figure 7 This is a module architecture diagram of the reinforcement learning and adaptive path adjustment module of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Please see attached Figure 1 -Attached Figure 7 The embodiment of the present invention provides a remote guidance and real-time feedback system for continuing education of resident physicians, including: A learning task management module is used to manage and recommend learning tasks for residents, including dynamically adjusting the priority and difficulty of learning tasks; The core purpose of this module design is to help residents efficiently master the required knowledge and skills during the teaching process through intelligent and dynamic task recommendations and priority adjustments, ensuring that the difficulty and sequence of learning tasks can change dynamically according to their real-time performance and needs, thereby greatly improving learning efficiency and quality. This module works closely with the aforementioned modules (such as the real-time feedback and evaluation module, the multi-objective optimization and path adjustment module, etc.) to ensure that residents are always on the most appropriate learning path and obtain tasks that match their current learning stage.

[0023] In this embodiment, the learning task management module includes a task library, a dynamic task pushing unit, a task generating unit, and a task priority management unit, which are described in detail as follows.

[0024] The task library is the basic part of the learning task management module. The task library contains all available learning tasks, which cover all aspects from basic theory to clinical practice, including medical theoretical knowledge, surgical operation skills, clinical diagnosis and decision-making, etc. Each task has a clear learning goal and difficulty level, and will be dynamically updated according to the physician's learning progress, mastery and clinical needs.

[0025] Specifically, tasks in the task library are classified and organized according to the following criteria: Task type: e.g. medical knowledge, operational skills, clinical decision making, etc.; Task difficulty: Each task has different difficulty levels, such as elementary, intermediate, and advanced, according to the depth of knowledge and complexity of operations involved; Task objectives: Each task has a clear learning goal, which may be the mastery of knowledge points, the improvement of skills, or the enhancement of clinical decision-making ability.

[0026] The task library will be updated regularly based on the residents' learning needs and progress to ensure that the task content is consistent with the latest medical advances and the residents' learning needs.

[0027] The dynamic task push unit dynamically adjusts the order and difficulty of task recommendations based on residents' real-time feedback, learning progress, historical performance and other data. The goal of this unit is to ensure that residents always receive tasks that are most suitable for their current learning stage and ability level.

[0028] During the task push process, the system automatically adjusts the task recommendations based on the residents’ learning goals and actual performance. Specifically: If a resident performs well on a task, the system will recommend a more difficult task to ensure that the resident continues to improve; If a resident performs poorly on a certain task, the system will automatically push relatively basic or related review tasks so that the physician can strengthen his or her deficiencies.

[0029] The intelligence and personalization of task push are particularly prominent in this module. It can accurately identify the needs of residents in the learning process and continuously optimize the task recommendation strategy through data analysis.

[0030] The task generation unit is an important part of this module, which is responsible for generating new tasks based on the residents' real-time learning performance. Task generation is not limited to the recommendation of basic tasks, but can dynamically adjust the task content and goals based on the physician's feedback and clinical needs.

[0031] Specifically, the task generation process is based on the following data sources: Residents’ learning feedback: such as feedback from physicians on existing tasks, performance in simulations, and quality of clinical decision making; The physician's progress: such as the difficulty of completed tasks, the quality of task completion, etc.

[0032] The generated new tasks will be closely related to the residents’ learning goals, and the content and form of the tasks will be automatically adjusted according to their actual performance. For example, if the resident performs well in a skill training simulation task, the system may generate a more difficult training task based on that task, or add a more challenging situational task to it.

[0033] The task priority management unit is used to dynamically adjust the priority of tasks based on the needs of residents, the importance of tasks, and the current learning status of the physician. Priority management ensures that residents can prioritize tasks that are critical to their current learning progress.

[0034] Priority adjustment takes into account the following factors: Urgency of tasks: Some tasks may be basic tasks necessary for residents’ clinical work, such as emergency department knowledge training, and these tasks have a higher priority; The current capabilities of the resident: By analyzing the resident’s performance on tasks, the system can adjust priorities to ensure that the resident’s learning path always matches their capabilities; Match between tasks and learning goals: If the task directly helps the resident's current learning goals, the system will give it higher priority.

[0035] The priority is calculated based on the following formula:

[0036] in: Indicates the task Priority; Indicates the task The urgency of the task, such as whether it is closely related to clinical work; A resident physician The level of competence in this task is usually determined by the physician’s learning history and feedback performance; Indicates the task Relevance to the resident’s current learning goals: if the task is highly relevant to the resident’s current learning goals, it will be given a higher priority; , , It is to adjust the weight coefficient of priority calculation to ensure that the impact of different factors on priority calculation is reasonably balanced.

[0037] Task recommendation and priority adjustment is not just a simple static management, but a dynamic optimization process. The learning progress and feedback of residents will continue to affect task push and priority adjustment to ensure that tasks always meet the current needs of doctors.

[0038] For example, if a resident has reached a high level of medical knowledge but still lacks operational skills, the system will automatically push more difficult operational training tasks instead of pushing too many internal medicine learning tasks. This intelligent adjustment is achieved through dynamic adjustment of task priorities and task push.

[0039] The task generation unit and the priority management unit work together to ensure that the task recommendations are both challenging and meet the personalized learning needs of residents. The cooperation between the two effectively avoids the monotony of task recommendations and the accumulation of overly complex tasks, greatly improving the pertinence and flexibility of the learning process.

[0040] Through the above functional modules and technical details, the learning task management module can intelligently generate, adjust and push tasks to ensure that residents can efficiently improve their medical knowledge, operational skills and clinical decision-making capabilities during the learning process.

[0041] Real-time feedback and evaluation module, which is used to evaluate residents' learning performance, operation accuracy and clinical decision-making in real time and provide immediate feedback; The core function of this module is to provide accurate and timely feedback based on the real-time performance of residents in the learning process, and to conduct a comprehensive assessment of their operational accuracy, clinical decision-making ability, etc. This module works closely with the learning task management module, game theory and strategy optimization module, etc., to ensure the continuous optimization of residents' learning path through a closed-loop feedback mechanism, thereby improving their clinical skills and decision-making ability.

[0042] The implementation of this module not only relies on advanced computer vision, image processing, deep learning and other technologies, but also combines intelligent feedback from clinical decision-making evaluation and simulation training. It can provide residents with real-time evaluation reports on operations and decisions, and then adjust their learning tasks or training content.

[0043] In this embodiment, the real-time feedback and evaluation module includes an operation evaluation unit, a clinical decision evaluation unit and a feedback generation unit.

[0044] The Operation Evaluation Unit uses high-precision image recognition technology and motion recognition algorithms to monitor the accuracy of residents in performing medical operations in real time. The goal of this unit is to accurately analyze the residents' operation steps, identify whether there are any operational deviations in real time, and generate improvement suggestions based on the analysis results.

[0045] Specifically, the Operational Evaluation Unit can analyze the following key indicators: Motion accuracy: Monitor the accuracy of residents during surgery or operations through image recognition and motion capture; Movement Fluency: Assess whether the resident's movements are smooth when performing the procedure, avoiding unnecessary pauses or errors.

[0046] The formula used for operational assessment is:

[0047] in: Indicates that residents are on task Operational evaluation scores in; is the evaluation indicator The weight of indicates the importance of the indicator in the evaluation; Is a resident doctor on duty Middle The error metric of each action measures the deviation in operation accuracy or fluency; Is a resident doctor on duty Action data in is the total number of action evaluations.

[0048] The function of this formula is to calculate the operation accuracy of the resident physician based on the operation data, and finally obtain the overall operation performance evaluation score by weighted synthesis of various error indicators. The system will provide feedback to the physician's operation based on the evaluation results, such as "the operation accuracy is too low, please adjust the scalpel angle" and other suggestions.

[0049] The task of the Clinical Decision Assessment Unit is to evaluate the decision-making ability of residents when faced with complex clinical situations, especially the quality of decision-making in diagnosis, treatment selection, etc. This unit is based on real or simulated clinical cases to analyze the decision-making path and rationality of residents when dealing with complex cases.

[0050] Specifically, clinical decision-making assessments consider the following aspects: Rationality of decision making: whether the resident selected an appropriate diagnosis and treatment plan; Timeliness of decision making: whether the resident made the decision at the appropriate time; Accuracy of the plan: Whether the resident selected the correct treatment method based on the condition and medical standards.

[0051] The evaluation formula for clinical decision making is:

[0052] in: Indicates that residents are on task Clinical decision making assessment scores in ; It is the decision path The importance coefficient of , which indicates the contribution of each decision path to the evaluation; Is a resident doctor on duty Middle The evaluation score of each decision path reflects the rationality of its decision; Is a resident doctor on duty Decision-making data in is the total number of decision paths.

[0053] The formula conducts a weighted comprehensive assessment of residents' decision paths to determine whether their decisions are appropriate when faced with complex clinical situations, and provides feedback to residents in the form of scores to help them improve their decision-making process.

[0054] The feedback generation unit generates real-time feedback based on the results of the operation evaluation unit and the clinical decision evaluation unit. Feedback generation is not limited to simple evaluation of operation accuracy and clinical decision-making, but also includes personalized suggestions for residents to help them better master skills and improve decision-making ability.

[0055] The functions of the feedback generation unit include: Multimodal feedback generation: Based on the evaluation results, the system will generate feedback in a form suitable for residents. The feedback can be text, voice, image or video, ensuring that residents can fully understand the feedback content and make effective improvements based on the feedback.

[0056] Personalized feedback: Based on the resident’s learning progress, feedback history, and task performance, the feedback generation unit adjusts the complexity and depth of the feedback content. For example, for a resident who performs well, the system may only give a brief confirmation feedback; for a resident who performs poorly in an operation or decision, detailed improvement suggestions will be given.

[0057] The optimization formula generated by feedback is:

[0058] in: Indicates the comprehensive score of the generated feedback content; is the weight of each feedback item, indicating the importance of the feedback item to physician learning; Is a feedback item The degree of helpfulness to the residents reflects the effectiveness of each feedback item; Is a resident doctor on duty Feedback data received in is the total number of feedback items.

[0059] Through this formula, the feedback generation unit can weight each feedback item to ensure that the generated feedback content meets the learning needs of residents and can effectively promote their skill improvement.

[0060] The real-time feedback and evaluation module works closely with other modules (such as the learning task management module and the multi-objective optimization and path adjustment module). After the task management module recommends a task, the real-time feedback and evaluation module will evaluate the resident's task performance in real time and pass the feedback results to the learning task management module to adjust the difficulty and priority of subsequent tasks.

[0061] Specifically, the system monitors the performance of residents during the task execution. If the physician's operation or decision-making evaluation in a task is low, the feedback generation unit will adjust the task recommendation in real time and provide appropriate training content. This feedback and adjustment mechanism forms a closed loop, ensuring that residents always stay on the learning path that best suits their current status during the learning process.

[0062] The real-time feedback and evaluation module combines computer vision, motion recognition, decision analysis and other advanced technologies to accurately evaluate residents' operation and decision-making performance and provide personalized instant feedback. Through this feedback and evaluation mechanism, residents can continuously improve their operation and decision-making capabilities during the education process, thereby improving their overall quality and meeting the high standards required in medical practice.

[0063] Multi-objective optimization and path adjustment module, which is used to optimize the learning path according to the physician's learning goals and progress, and dynamically adjust the task sequence and difficulty; The main function of this module is to dynamically optimize the learning path based on the resident's personalized learning goals, real-time feedback, and historical performance, to ensure that the resident can strike a balance between multiple learning goals (such as knowledge mastery, operational skills, clinical decision-making ability, etc.), and to adjust the order and difficulty of learning tasks in a timely manner. Through this optimization and adjustment, the system can ensure that the resident's learning process is efficient and targeted, and can comprehensively improve their comprehensive capabilities in the clinical and knowledge fields.

[0064] The multi-objective optimization and path adjustment module works closely with the aforementioned modules (such as the learning task management module, the real-time feedback and evaluation module, etc.) to ensure that residents can make balanced progress on multiple learning objectives when completing each learning task, while adapting to changes in their personal abilities and progress.

[0065] In this embodiment, the multi-objective optimization and path adjustment module includes an objective function calculation unit, an optimization algorithm unit and a path adjustment unit.

[0066] The objective function calculation unit calculates the evaluation results of different objective functions based on the residents’ learning goals and task performance. The main purpose of the objective function calculation is to measure the residents’ learning progress and ability in multiple dimensions (such as knowledge mastery, operational skills, clinical decision-making ability, etc.) and provide necessary data for subsequent optimization algorithms.

[0067] In this embodiment, each learning goal corresponds to an objective function, and the learning dimensions involved in these objective functions may include: Knowledge mastery objective function: evaluates residents’ knowledge mastery in various medical fields (such as internal medicine, surgery, etc.).

[0068] Operational skills objective function: evaluates the performance of residents in actual operations such as surgical operations and medical examinations.

[0069] Clinical decision-making objective function: Evaluate residents' decision-making ability when faced with complex conditions.

[0070] Each objective function can be calculated using the following formula:

[0071] in: Indicates that the resident physician The score on the objective function; is the first The weight coefficient of a factor indicates the influence of the factor on the calculation of the objective function; Is a resident physician in Learning tasks performance ratings in (e.g., operational accuracy, knowledge mastery, etc.); is the number of factors to be evaluated in the objective function.

[0072] This formula helps the system quantify residents’ progress on each learning objective and combines the calculated results of all objectives to guide subsequent learning task recommendations.

[0073] Optimization Algorithm Unit The optimization algorithm unit is responsible for calculating the optimal learning path for residents based on the calculation results of the objective function by applying intelligent optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.). This process not only ensures that the recommended order of tasks meets the learning needs of residents, but also ensures that the difficulty of the tasks is adapted to the current learning status of residents.

[0074] The core functions of the optimization algorithm unit include: Task sequence optimization: The system automatically optimizes the recommended order of tasks based on the resident’s learning progress and the objective function evaluation results. Sequence optimization ensures that residents can maximize their learning efficiency when completing each task and avoid repeated training.

[0075] Task difficulty adjustment: Automatically adjust the difficulty of subsequent tasks based on the resident’s performance in the task, ensuring that the task is challenging but within the physician’s capabilities.

[0076] Goal balance optimization: The optimization algorithm can achieve a reasonable balance between multiple learning goals based on the performance of residents, ensuring that each goal (such as knowledge, operational skills, clinical decision-making, etc.) is given corresponding attention and improvement.

[0077] During the optimization process, the system integrates multiple objective functions through weighted optimization methods and dynamically adjusts the order and difficulty of task recommendations. This process can be optimized using the following formula:

[0078] in: Indicates The evaluation value of the objective function; It is The weight coefficient of an objective function indicates the influence of the objective function on the optimization of the learning path; It is the learning status data of residents, including the difficulty, time, learning effect and other information of the currently completed tasks; is the total number of objective functions, representing the number of learning objectives that the system needs to optimize (e.g., knowledge mastery, skill improvement, clinical decision making, etc.).

[0079] The optimization algorithm uses these weight coefficients to ensure a reasonable balance between multiple objectives, thereby obtaining the optimal learning path for residents.

[0080] The task of the path adjustment unit is to adjust the residents' learning path according to the optimal learning path provided by the optimization algorithm to ensure that their learning tasks and goals are always consistent with their learning progress and feedback.

[0081] The functions of the path adjustment unit include: Dynamic task recommendation: This unit dynamically adjusts the recommended content of subsequent tasks based on the performance of the residents in the real-time feedback and evaluation module. The order and difficulty of tasks in the learning path are adjusted according to the performance of the residents in the current task to ensure that the residents can challenge the appropriate tasks at the appropriate time.

[0082] Task priority adjustment: The pathway adjustment unit adjusts task priorities to ensure that residents can complete the most critical tasks first and fill in the weak links in the learning process in a timely manner.

[0083] The Multi-Objective Optimization and Path Adjustment Module is closely integrated with other modules (such as the Learning Task Management Module, Real-Time Feedback and Evaluation Module, etc.). In the Learning Task Management Module, the system recommends tasks based on the resident's learning goals and progress. When the resident completes the task, the Real-Time Feedback and Evaluation Module will evaluate his or her performance and feed the evaluation results back to the Multi-Objective Optimization and Path Adjustment Module. This module will optimize the order and difficulty of tasks based on the feedback data and the calculation results of the objective function to ensure that the resident can make reasonable progress on each learning goal.

[0084] Specifically, when the resident completes the task, the real-time feedback and evaluation module evaluates his / her performance, and the generated feedback will directly affect the task recommendation strategy of the multi-objective optimization and path adjustment module. In this way, the entire system forms a closed loop, ensuring that the resident's learning path is always optimized, thereby improving his / her clinical skills and decision-making ability.

[0085] The multi-objective optimization and path adjustment module achieves dynamic optimization of residents' learning paths through intelligent algorithms, objective function calculations, task priority adjustment and other technical means.

[0086] Game theory and strategy optimization module, which is used to simulate the decision-making interaction between residents and mentors, and between physicians, and optimize task allocation and learning strategies; This module mainly simulates and optimizes the decision-making process between residents and mentors, and between residents and colleagues through game theory models to improve residents' learning efficiency and the effectiveness of educational interaction. Through game theory and strategy optimization, residents can efficiently achieve their educational goals in the continuously optimized learning path and task recommendation, while maintaining a good state of collaboration and competition.

[0087] This module works closely with the aforementioned learning task management module, real-time feedback and evaluation module, multi-objective optimization and path adjustment module, etc., to ensure that residents not only receive personalized task recommendations during the learning process, but also receive more accurate learning feedback and optimized task recommendations through teaching interactions with mentors and colleagues. The goal of the game theory and strategy optimization module is to ensure that residents always stay on the most appropriate learning path through strategy adjustment, task allocation optimization, interactive decision-making, etc.

[0088] In this embodiment, the game theory and strategy optimization module includes a strategy decision unit, a task allocation unit and a strategy update unit.

[0089] The main task of the strategy decision unit is to simulate the decision-making interaction process between residents and mentors, and between residents and other residents. During the education process, there is a game relationship between residents and mentors. The learning progress and feedback of residents will affect the teaching strategy of mentors, and the strategy of mentors will in turn affect the learning performance of residents. Therefore, the strategy decision unit optimizes the learning strategies and task recommendations of both parties based on the interaction between residents and mentors through a game theory model.

[0090] Specifically, the working principle of the policy decision unit is based on the following technologies: Application of game theory models: By simulating the decision-making interaction between residents and mentors, game theory models (such as Nash equilibrium, Stackelberg game, etc.) are used to calculate the optimal strategies of both parties; Optimize learning paths: Adjust teaching strategies based on residents’ performance and feedback, optimize task recommendations, and ensure that residents’ learning paths are maximized.

[0091] The interaction between residents and mentors can be modeled through the following game model to optimize the decision-making of both parties:

[0092] in: is the tutor's strategy, which indicates the strategies adopted by the tutor during the teaching process (e.g., task recommendation, teaching methods, etc.); is the resident's strategy, which indicates the response made by the resident during the learning process (e.g., task selection, learning speed, etc.); is the payoff function of the game, which represents the total utility of the resident and the mentor in the interaction; It is the learning status data of residents, including learning progress, feedback, etc.

[0093] This game model achieves the optimal strategy by maximizing the teaching effect of the tutor and the learning effect of the resident, thereby improving the collaboration efficiency of both parties.

[0094] The task of the task allocation unit is to reasonably allocate tasks according to the ability of the resident, the urgency of the task, the feedback of the resident and the learning goal, so as to ensure the efficient completion of the task. Through the optimization algorithm of game theory, the task allocation unit can dynamically adjust the priority and difficulty of the task according to the learning status of the resident, so as to ensure that the resident can obtain the most appropriate learning task.

[0095] Specifically, the functions of the task allocation unit include: Assignment of Task Priorities: Assign priorities to tasks based on their urgency, the resident’s current capabilities, and progress; Collaboration and competition among tasks: Optimize the distribution of tasks between residents and other colleagues to ensure that each resident can learn as needed and improve their clinical capabilities; Task difficulty adaptation: Dynamically adjust the difficulty of tasks based on the residents’ learning progress to ensure that the tasks are challenging but not beyond their capabilities.

[0096] The optimization of task allocation can be performed by the following formula:

[0097] in: Representation Task Assigned to resident physician This indicates that the task is in the resident physician Position in the learning path; Still indicates the task The urgency or importance of the task, reflecting the criticality of the task to the physician’s learning path; Indicates resident physician On Task This item represents the physician's ability to complete the task; It's a task With resident physicians relevance to current learning goals: the higher the relevance of the task to the physician's goals, the higher the priority assigned; A resident physician On Task This is used to optimize the distribution of tasks among multiple residents and ensure fairness and efficiency in task distribution; , , , are the weight coefficients for adjusting the priority, which respectively represent the impact of task urgency, resident competence, task relevance, and collaboration on the priority; This formula ensures that residents complete the most relevant and challenging tasks at the appropriate time, while dynamically adjusting the order of tasks based on their progress and capabilities.

[0098] The task of the strategy updating unit is to update the learning strategies between residents and mentors, and between residents and physicians in real time based on the feedback obtained during the game. During the interaction between residents and mentors, strategies may change, especially when the performance of residents improves or declines, the mentor's teaching strategy may be adjusted, and the resident's learning strategy will also change accordingly. Therefore, the strategy updating unit can adjust the strategy in real time based on the execution of the task and the feedback from the physician to ensure that the educational interaction remains efficient.

[0099] Specifically, the functions of the policy update unit include: Dynamic adjustment of strategies: Based on task feedback, residents’ learning progress and performance, the learning strategies between residents and mentors are adjusted in real time; Learning rate adjustment: Dynamically adjust the step size of strategy updates based on residents’ learning results and feedback.

[0100] The optimization of the strategy update can be performed by the following formula:

[0101] in: represents the update amount of the tutor's strategy; is the learning rate, which controls the step size of the policy update; is the mentor strategy relative to the payoff function The gradient of , indicates the optimization direction of the strategy.

[0102] Through this formula, the strategy updating unit can continuously adjust and optimize the strategy during the interaction between residents and mentors to achieve the best learning effect.

[0103] The game theory and strategy optimization module is closely integrated with other modules in the system (such as the learning task management module, real-time feedback and evaluation module, etc.). After the learning task management module recommends tasks, the real-time feedback and evaluation module will evaluate the performance of the residents, and the game theory and strategy optimization module will optimize task allocation and strategy adjustment based on the evaluation results. The strategic interaction between residents and mentors will affect the order, difficulty and learning path of subsequent task recommendations, thus forming an intelligent closed-loop feedback mechanism.

[0104] For example, when a resident performs poorly in a task, the real-time feedback and evaluation module generates feedback and passes it to the game theory and strategy optimization module. This module optimizes the interaction between residents and mentors through game models, adjusts mentors' teaching strategies and task allocation, and thus provides targeted learning path optimization. This synergy ensures that residents receive the most appropriate teaching support in continuous interaction, improving their learning outcomes.

[0105] The game theory and strategy optimization module optimizes task allocation and strategy adjustment by simulating the interaction between residents and mentors, and between residents and physicians. This module not only ensures the optimal cooperation between residents and mentors, but also improves the learning efficiency and clinical ability of residents through intelligent adjustment of task recommendations.

[0106] Reinforcement learning and adaptive pathway adjustment module, used to dynamically adjust learning pathways and task allocation based on residents’ feedback and performance; The goal of this module is to intelligently optimize the learning path of residents based on their real-time performance and learning feedback using reinforcement learning algorithms, and dynamically adjust the order and difficulty of tasks according to their learning progress and needs. Through this adaptive adjustment, residents can maintain the most appropriate learning state at different learning stages and different tasks, and gradually improve their clinical skills, knowledge mastery, and decision-making ability.

[0107] The reinforcement learning and adaptive path adjustment module is closely connected with the aforementioned modules (such as the learning task management module, the real-time feedback and evaluation module, the multi-objective optimization and path adjustment module, etc.). Specifically, the system evaluates the performance of residents in the task through real-time feedback, and passes the evaluation results to the reinforcement learning and adaptive path adjustment module, so that the learning path, task sequence and difficulty can be adjusted based on the actual situation of the residents, ensuring that they are always in the best learning state.

[0108] In this embodiment, the reinforcement learning and adaptive path adjustment module includes a reinforcement learning engine, a task adjustment unit and an adaptive adjustment unit.

[0109] The reinforcement learning engine is the core part of this module, which is responsible for gradually optimizing the learning path by learning the residents' performance and feedback in the tasks. The engine is based on the theoretical framework of reinforcement learning and uses a reward and punishment mechanism to guide the optimization of the learning path. During the learning process, the system adjusts its strategy based on the feedback data when the residents complete the tasks (such as the speed, quality, and operation accuracy of the tasks) to optimize the selection and sequence of subsequent tasks.

[0110] Specifically, the functions of the reinforcement learning engine include: State space definition: The learning state of residents (such as knowledge mastery, operational skill level, clinical decision-making ability, etc.) is regarded as a state space. Through feedback on the state, the reinforcement learning engine can dynamically adjust the learning path; Action selection mechanism: In each learning state, the system selects the most appropriate learning task (action) according to the current state and calculates the reward based on the effect of task completion; Reward and Penalty Mechanism: Through positive rewards and negative penalties, the reinforcement learning engine gradually optimizes the task selection strategy, ensuring that residents gradually improve their abilities throughout the learning process.

[0111] The core formula of a reinforcement learning engine usually includes the following mathematical expression that represents the learning reward mechanism:

[0112] in: Is a resident physician in state Next select action The current value of Is to perform an action awards received by post-resident physicians; is the discount factor, which is used to measure the impact of long-term rewards; is the learning rate, which controls the magnitude of each update; is the maximum Q-value in the next state, representing the maximum possible reward for the resident after transferring from the current state to the next state.

[0113] Through the above formula, the reinforcement learning engine can dynamically adjust the task recommendation strategy according to the real-time performance of the residents, thereby helping the residents maximize their learning effects.

[0114] The task adjustment unit ensures that the residents' learning path always adapts to their current learning progress and performance by dynamically adjusting the order and difficulty of tasks. Based on the feedback from the reinforcement learning engine, the task adjustment unit can automatically adjust the recommended order of tasks, increase or decrease the difficulty of tasks, and ensure that residents can learn efficiently with appropriate challenges.

[0115] Specifically, the functions of the task adjustment unit include: Task sequence optimization: The system dynamically optimizes the order of tasks based on the resident's performance to ensure that their learning process remains consistent and challenging. For example, when a resident performs poorly in a certain area, the system will prioritize basic tasks in related areas; Task difficulty adjustment: The task adjustment unit automatically increases or decreases the difficulty of tasks based on the resident's ability, ensuring that the tasks are challenging but not beyond their capabilities.

[0116] The function of the adaptive adjustment unit is to dynamically adjust the recommended strategy of the task based on the real-time feedback and performance of the residents to ensure that the learning path of the residents always matches their actual situation. Unlike the reinforcement learning engine, the adaptive adjustment unit focuses on real-time adjustments and quickly responds to changes in the residents during the learning process.

[0117] Specifically, the functions of the adaptive adjustment unit include: Real-time adjustment of task difficulty: Automatically adjust the difficulty of the task based on the resident’s performance in the current task to prevent the task from exceeding the resident’s ability.

[0118] Dynamic optimization of learning paths: When residents encounter obvious bottlenecks in a certain area, the adaptive adjustment unit can adjust the learning path in a timely manner to help residents strengthen their skills in that area through repeated learning and feedback.

[0119] The adaptive adjustment unit is optimized by the following formula:

[0120] in: is the amount of change in the strategy; is the learning rate, which controls the step size of the adjustment; is the loss function, which represents the deviation between the current learning path and the target learning state; It is the gradient of the loss function with respect to the strategy, indicating the update direction of the optimal strategy.

[0121] This formula helps the adaptive adjustment unit to update the learning path strategy in real time during the ever-changing learning process to ensure that residents are always in the best learning state.

[0122] The reinforcement learning and adaptive path adjustment module is closely integrated with other modules in the system (such as the learning task management module, the real-time feedback and evaluation module, the game theory and strategy optimization module, etc.). Specifically, when the resident completes the task, the real-time feedback and evaluation module will evaluate the resident's performance and pass the feedback to the reinforcement learning and adaptive path adjustment module. The reinforcement learning engine optimizes the learning path based on the evaluation results, and the task adjustment unit and the adaptive adjustment unit adjust the order and difficulty of the tasks in real time to ensure that the resident learns the appropriate content at the right time.

[0123] The reinforcement learning and adaptive path adjustment module ensures that residents can continuously optimize their learning paths throughout the learning process through technical means such as reinforcement learning algorithms, dynamic task adjustment and adaptive adjustment. Through real-time feedback and strategy adjustment, the system can dynamically adapt to the learning status of residents, ensuring that learning tasks match the physician's ability level and learning goals, thereby improving learning efficiency.

[0124] Data collection and analysis module, used to collect and analyze physicians’ learning data to support optimization and decision-making of other modules; This module is one of the core components of the system. Its main function is to collect and analyze a large amount of data during the residents' learning process, help other modules (such as learning task management module, real-time feedback and evaluation module, multi-objective optimization and path adjustment module, etc.) to make optimization decisions, and support personalized education path adjustment. Through effective algorithms and data processing, the data collection and analysis module can deeply explore the learning trends, behavior patterns and potential bottlenecks of residents, and then provide decision support for subsequent learning, ensuring that residents can continuously improve their professional capabilities according to the most appropriate learning path.

[0125] This module works in collaboration with other modules to provide intelligent support for the system, especially in terms of learning task management and path adjustment. By analyzing a large amount of learning data, it can promptly identify residents' weak links and learning bottlenecks, ensuring that they receive the most appropriate tasks at the appropriate time, thereby improving their learning outcomes.

[0126] In this embodiment, the data collection and analysis module includes a data acquisition unit, a data analysis engine and a report generation unit.

[0127] The data collection unit plays an important role in this module, and is responsible for collecting residents' learning data from other modules in the system in real time. These data include task completion, operation accuracy, learning time, feedback scores, clinical decision performance, etc. The data collection unit works closely with the learning task management module, real-time feedback and evaluation module, etc. to ensure that the performance of residents in various learning tasks can be accurately collected and transmitted to the data analysis engine for processing in real time.

[0128] Specifically, the functions of the data acquisition unit include: Task completion data collection: record the completion time of each task, the difficulty of the task, the score of the task, etc. Operation accuracy data collection: Through computer vision and motion recognition technology, the accuracy data of residents in operation tasks is collected in real time; Feedback data collection: Collect residents’ scores, evaluation opinions, etc. from the real-time feedback and evaluation module; Clinical decision data: records the decision-making process of residents in simulated or actual cases, including the accuracy of the decision, decision time, etc.

[0129] All these data provide necessary input for subsequent data analysis, and the data collection unit needs to ensure the timeliness and accuracy of the data.

[0130] The data analysis engine is the core part of the data collection and analysis module, responsible for in-depth analysis of the collected data. By processing and analyzing a large amount of data during the resident's learning process, the data analysis engine can provide residents with personalized learning feedback and support other modules to optimize the learning path and task sequence. The data analysis engine uses a variety of data analysis methods, including statistical analysis, machine learning, deep learning and other technologies, to explore the behavior patterns and learning trends of residents, identify potential learning bottlenecks, and predict future learning progress.

[0131] Specifically, the functions of the data analysis engine include: Data cleaning and processing: pre-process the data collected from multiple data sources to remove outliers and redundant information; Learning trend analysis: Through regression analysis of residents’ learning data, their performance trends at different learning stages were identified; Bottleneck identification: By comparing the performance of tasks in different fields, the system can automatically identify the weak links of residents in certain tasks, helping the system to adjust task recommendations in a timely manner; Prediction and modeling: Based on the historical data of residents, predict their future learning progress and help the multi-objective optimization and path adjustment module adjust the task sequence and difficulty.

[0132] The data analysis engine performs forecasting and trend analysis using the following formula:

[0133] in: is the predicted output value, which indicates the resident's future learning progress or task performance; It is Input features, representing different variables in the data (such as operation accuracy, task difficulty, etc.); It is a feature The weight of the output value indicates the influence of the feature on the learning progress; is the bias term, used to adjust the prediction model; is the number of features, representing the dimension of the input data.

[0134] Through the above formula, the data analysis engine can perform regression analysis or prediction based on historical data and real-time data, helping the system to adjust the learning path in real time.

[0135] The report generation unit is responsible for generating personalized learning reports for residents based on the analysis results of the data analysis engine. The report content includes the summary and analysis of various data such as the resident's learning progress, task completion, operation skills, clinical decision-making ability, knowledge mastery, etc. The generated report not only provides feedback for residents, but also provides an evaluation basis for mentors and education managers to help them make corresponding adjustments based on the physician's learning progress.

[0136] Specifically, the functions of the report generation unit include: Personalized learning report: Generate personalized reports based on the residents’ performance at different stages of the learning process. The report content includes task completion, operation accuracy, clinical decision quality, etc.; Learning trend chart: Display residents’ learning trends and bottlenecks through charts and data visualization, helping doctors and mentors understand progress more clearly; Optimization suggestions: Based on the residents’ learning data and analysis results, the report will also provide personalized learning optimization suggestions, such as further training of a certain skill, review of a certain knowledge point, etc.

[0137] The report generation unit provides a detailed and easy-to-understand feedback through comprehensive analysis of the data, helping residents and educators understand areas of improvement in the learning process.

[0138] The generation of report content can be optimized by the following formula:

[0139] in: Indicates the overall score of the generated report, which determines the level of detail of the report; is the weight of each content in the report, indicating the contribution of the content to the resident's learning; Is a resident doctor on duty Feedback data in; is the total number of content items in the report.

[0140] This formula ensures that every item in the report provides valuable feedback to residents, helping them better understand their own learning and further improve their learning outcomes.

[0141] Identify the learning bottlenecks and potential problems of residents, and help adjust task recommendations and learning paths. The data collection and analysis module provides the basis for intelligent decision-making of the entire system, ensuring that residents receive the best task recommendations during the learning process, maximizing their learning effects and clinical ability improvement.

[0142] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote guidance and real-time feedback system for continuing education of resident physicians, characterized in that: include: A learning task management module is used to manage and recommend learning tasks for residents, including dynamically adjusting the priority and difficulty of learning tasks; Real-time feedback and evaluation module, which is used to evaluate residents' learning performance, operation accuracy and clinical decision-making in real time and provide immediate feedback; Multi-objective optimization and path adjustment module, which is used to optimize the learning path according to the physician's learning goals and progress, and dynamically adjust the task sequence and difficulty; Game theory and strategy optimization module, which is used to simulate the decision-making interaction between residents and mentors, and between physicians, and optimize task allocation and learning strategies; Reinforcement learning and adaptive pathway adjustment module, used to dynamically adjust learning pathways and task allocation based on residents’ feedback and performance; The data collection and analysis module is used to collect and analyze physicians’ learning data to support the optimization and decision-making of other modules.

2. A remote guidance and real-time feedback system for continuing education of resident physicians according to claim 1, characterized in that: The learning task management module includes: A task library that stores all available learning tasks and generates task recommendations based on residents’ progress and needs; Dynamic task push unit, used to adjust the priority and difficulty of tasks based on physician feedback and performance; The task generation unit is used to generate new tasks or adjust existing tasks based on the real-time performance of residents.

3. A remote guidance and real-time feedback system for continuing education of resident physicians according to claim 1, characterized in that: The real-time feedback and evaluation module includes: A clinical decision-making assessment unit to assess residents’ decision-making process in complex cases and provide feedback; A feedback generation unit is used to generate feedback based on the evaluation results and provide feedback to the resident physician via text, voice or image.

4. A resident physician continuing education remote guidance and real-time feedback system according to claim 1, characterized in that: The multi-objective optimization and path adjustment module includes: The objective function calculation unit is used to calculate multiple objective functions of residents in terms of knowledge mastery, operational skills, and clinical decision-making ability; An optimization algorithm unit, used to optimize the learning path, task sequence and difficulty according to the objective function results; The path adjustment unit is used to dynamically adjust the learning path according to the results of the optimization algorithm.

5. A resident physician continuing education remote guidance and real-time feedback system according to claim 1, characterized in that: The game theory and strategy optimization module includes: A strategic decision-making unit, which is used to simulate the strategic interaction between residents and mentors and optimize task allocation; The task allocation unit is used to optimize task allocation through game theory models to make the interaction between residents and mentors more efficient; The strategy updating unit is used to update the task allocation strategy and education strategy according to the changes in the game process.

6. A resident physician continuing education remote guidance and real-time feedback system according to claim 1, characterized in that: The reinforcement learning and adaptive path adjustment module includes: A reinforcement learning engine is used to optimize the learning path based on the physicians’ learning feedback and gradually improve the learning efficiency; A task adjustment unit, which is used to dynamically adjust the difficulty and sequence of tasks according to the residents’ learning situation; The adaptive adjustment unit is used to adjust task allocation and learning progress according to real-time feedback.

7. A resident physician continuing education remote guidance and real-time feedback system according to claim 1, characterized in that: The data collection and analysis module includes: Data collection unit, used to collect residents’ learning progress, test scores, and operation data in real time; Data analysis engine, used to conduct in-depth analysis of collected learning data and discover trends and potential problems in learning; The report generation unit is used to generate personalized learning reports for doctors and feed back the analysis results to the learning task management module and optimization module.

8. The system for remote guidance and real-time feedback of resident physician continuing education according to claim 3, characterized in that: The real-time feedback and evaluation module further comprises: The Virtual Patient Simulation Unit is used to simulate virtual patients and provide realistic clinical feedback to help residents train their skills and decision-making; The simulation feedback generation unit is used to generate feedback and suggestions according to the physician's operations during the virtual patient simulation process.

9. The system for remote guidance and real-time feedback of resident physician continuing education according to claim 2, characterized in that: The learning task management module further includes: Task priority management unit, used to dynamically adjust the priority of tasks according to the importance of tasks, physicians’ learning progress and clinical needs; The Task Adaptability Assessment Unit is used to evaluate residents’ performance on different tasks, ensuring that the tasks are always adapted to the physician’s learning ability.

10. The system for remote guidance and real-time feedback of continuing education for resident physicians according to claim 5, characterized in that: The game theory and strategy optimization module further includes: Multi-physician task coordination unit, used to simulate cooperation and competition among multiple physicians and optimize task allocation strategy according to game results; The decision optimization engine is used to optimize the strategies between physicians and tutors in real time based on the results of game theory analysis, ensuring optimal task allocation and interaction during the education process.

Images