Calculation co-processing hardware system and method based on user personality modeling

Abstract

The invention relates to the field of cognitive computing, and discloses a computing co-processing hardware system and method based on user personality modeling, which comprises the following steps: aggregating records by taking task instances as boundaries, and executing time sequence correction on the aggregated records according to an operation triggering precedence relationship and a context dependency relationship; splitting the behavior data set according to the task stage switching points, and mapping the split behavior data to an operation type space; user cognitive feature description data is injected into a configuration entry of an information processing flow, and calling conditions, execution sequences and parallel relationships of information analysis nodes, result splicing nodes and intermediate data resident nodes are combined and set; continuously recording confirmation, revision and cross-stage calling behaviors of a user on a stage result; and according to the adjusted information processing node participation set, performing structured splitting and sequential recombination on the information data flow entering the system. The method has the advantage of improving the information matching accuracy.

Description

Technical Field

[0001] This invention relates to the field of cognitive computing, specifically to a hardware system and method for collaborative computing based on user personality modeling. Background Technology

[0002] In applications such as intelligent customer service, decision support, and human-computer collaboration, cognitive computing systems typically need to understand, reason about, and respond to user input. Existing systems often employ a uniform cognitive processing flow, lacking fine-grained differentiation of different users' behavioral preferences, understanding habits, and decision-making styles, usually only performing simple adaptations at the interface or rule level. When the same business system is used by multiple users at different work stages and under varying cognitive loads, the system struggles to dynamically adjust the information processing order and computational focus based on individual user differences, easily leading to issues such as response redundancy and insufficient matching of key information. Taking enterprise internal knowledge retrieval and task assistance scenarios as an example, users in different roles exhibit significant differences in the dimensions of information they focus on, the acceptable information density, and the pace of interaction when facing the same business problem. However, existing cognitive computing systems, when processing such requests, typically treat user input as isolated events, lacking joint modeling of users' long-term behavioral characteristics and immediate cognitive states. This makes it difficult to coordinate and optimize computational resource allocation and cognitive processing results, impacting overall system efficiency and user experience. Therefore, designing a cognitive computing collaborative processing system and method based on user-personality modeling to improve information matching accuracy is essential. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a computational collaborative processing hardware system and method based on user personality modeling, which has the advantage of improving the accuracy of information matching and solves the problems mentioned in the background technology.

[0004] To achieve the aforementioned goal of improving information matching accuracy, this invention provides the following technical solution: a computational collaborative processing hardware method based on user personality modeling, comprising the following steps:

[0005] Aggregate records by task instance, and perform timing correction on the aggregated records according to the order of operation triggering and context dependency to construct a behavioral data set that reflects the user's task progression.

[0006] The behavioral data set is split according to the task stage switching point. The split behavioral data is mapped to the operation type space. The information access trigger interval, backtracking operation span and interruption position distribution are segmented and modeled. Repeated operation structures and isolated operation structures are identified in each segment to form user authentication feature description data.

[0007] User cognitive feature description data is injected into the configuration entry of the information processing flow. The calling conditions, execution order and parallel relationship of information parsing nodes, result splicing nodes and intermediate data resident nodes are combined and set to generate a processing flow definition corresponding to user behavior features.

[0008] During the process of defining and running the workflow, the user's confirmation, revision and cross-stage call behavior of the phase results are continuously recorded, and the participation set and processing level of the information processing nodes that have not been executed are dynamically adjusted.

[0009] Based on the adjusted set of information processing nodes, the system performs structured splitting and sequential reorganization of the information data stream entering the system, and completes information output according to the defined processing flow.

[0010] Preferably, the process of constructing a behavioral data set reflecting the user's task progression method is as follows:

[0011] Task instance identifiers are bound to the operation logs, information access records, and system feedback records generated by users under the same target, forming a set of original operation events;

[0012] Based on the trigger timestamps, preceding and following references, and data dependency markers of each operation event, the original set of operation events is initially sorted to construct the operation trigger sequence;

[0013] Introduce context dependency constraints to correct the timing of operation events that have asynchronous returns, cache hits, or lazy loading, thereby eliminating the discrepancy between apparent time and actual perceived order.

[0014] The corrected operation events are merged according to task instances to generate a set of behavioral data reflecting the user's decision-making order and operation rhythm during task progress.

[0015] Preferably, the process of segmenting and modeling the information access trigger interval, backtracking operation span, and interruption location distribution is as follows:

[0016] Based on changes in operational objectives, resource switching, and output format changes in the behavioral data set, identify the stage switching points in the task advancement process;

[0017] Using the phase switching point as the boundary, the behavioral data set is divided into several phased behavioral subsequences. According to the pre-established operation type mapping rules, the operation events in each phased behavioral subsequence are mapped to a unified operation type space.

[0018] Within the operation type space, the distribution of information access trigger intervals, the range of backtracking operations, and the frequency of operation interruption locations are statistically analyzed to form a segmented description model of behavior within a stage.

[0019] Preferably, the process of forming user cognitive feature description data is as follows:

[0020] In the behavior segmentation description model at each stage, a structured scan is performed on the operation type sequence to extract the operation combination patterns that appear continuously or at intervals.

[0021] Operation combinations that occur more frequently than a set threshold are marked as stable operation structures, representing the user's normal cognitive path;

[0022] Combinations of operations that occur only under specific conditions or in a single context are marked as isolated operation structures, representing the user's exploratory behavior;

[0023] By combining the distribution characteristics of stable and isolated operation structures, user cognitive feature description data reflecting user operation preferences, fault tolerance habits, and stage control methods is generated.

[0024] Preferably, the process of injecting user cognitive feature description data into the configuration entry point of the information processing flow is as follows:

[0025] Set up a process configuration entry point associated with user tasks in the information processing system to receive external cognitive control parameters;

[0026] The user cognitive feature description data is parsed into process constraint parameters, node preference parameters, and parallel control parameters;

[0027] The triggering conditions and participation methods of adjustable nodes in the information processing flow are initially bound to complete the mapping and injection between user cognitive characteristics and information processing flow.

[0028] Preferably, the process of generating a processing flow definition corresponding to user behavior characteristics is as follows:

[0029] Based on the node preference parameters injected in the process configuration entry, the trigger priority and parsing depth of the information parsing node are set;

[0030] Based on users' habits of viewing and revising results in historical tasks, the merging rules and output granularity of result splicing nodes are adjusted;

[0031] Configure the retention time and call visibility of intermediate data nodes based on the user's dependence on intermediate results;

[0032] The execution order and parallel relationships of various nodes are combined and arranged to form a processing flow definition that matches the user's behavior characteristics.

[0033] Preferably, the process of continuously recording users' confirmation, revision, and cross-stage call behavior of interim results is as follows:

[0034] When the process is defined and executed, the stage results output by each information processing node are identified and the corresponding process stage number and node identifier are recorded.

[0035] When a user performs a confirmation operation, a content revision operation, or calls the results of other stages across the current stage on a stage result, the operation type, the trigger time, and the associated stage result identifier are recorded respectively.

[0036] The recorded operations are associated with the current execution stage, the set of completed nodes, and the set of nodes to be executed in the current processing flow to construct a runtime behavior record that reflects the user's intervention position and intervention method during the process.

[0037] Preferably, the process of dynamically adjusting the participation set and processing level of unexecuted information processing nodes is as follows:

[0038] Based on runtime behavior records, the information processing nodes that are actually triggered, confirmed, or called across stages by the user in the current task instance are counted to determine the subset of nodes participating in the execution of the current process;

[0039] For information processing nodes that are not triggered in multiple consecutive task phases, lower their triggering priority in the processing flow; for information processing nodes that are called multiple times across phases in the runtime behavior record, increase their processing level in the processing flow.

[0040] Write the adjustment results of node trigger priority, processing level and node execution order into the process status description corresponding to the current task instance, and update the information processing node participation set.

[0041] Preferably, the process of completing information output according to the processing flow definition is as follows:

[0042] Based on the set of information processing nodes, the input information data stream is split at the field level and semantic level.

[0043] According to the updated processing flow definition, the split information data units are sequentially reorganized and distributed to nodes.

[0044] After each information processing node completes its corresponding processing, the intermediate results are aggregated and their consistency is verified. Then, the information output results are generated according to the output rules defined in the processing flow.

[0045] A hardware system for collaborative computing based on user personality modeling, comprising:

[0046] Task aggregation module: It aggregates user operation records by task instance and corrects the execution sequence of the records according to the order of operation triggering and contextual dependencies to form a behavioral data set;

[0047] Stage modeling module: The behavioral data set is split according to the task stage switching points, and the split behavioral data is mapped to the operation type space to perform segmented modeling of the operation characteristics within each stage;

[0048] Process configuration module: Receives user cognitive feature description data, combines and sets the calling conditions, execution order and parallel relationship of each processing node in the information processing process, and generates the processing process definition;

[0049] Process adjustment module: Records user actions on phased results during the execution of the processing flow definition, and adjusts the set and processing level of unexecuted processing nodes;

[0050] Data Reassembly Module: Based on the adjusted set of processing nodes, the module splits and reassembles the incoming data stream into the system, and completes the information output according to the defined processing flow.

[0051] Compared with existing technologies, this invention provides a hardware system and method for collaborative computing based on user personality modeling, which has the following beneficial effects:

[0052] This invention injects user cognitive feature description data into the information processing flow configuration entry point, enabling fine-grained adjustment of the calling conditions, execution order, and parallel relationships of key processing nodes such as information parsing, result splicing, and intermediate data retention. This transforms the information processing flow from a fixed preset mode to a collaborative processing mode that can dynamically adapt to user behavior. During the flow's execution, it continuously records user confirmation, revision, and cross-stage calling behaviors for phased results, and dynamically adjusts the participation set and processing level of unexecuted nodes. This effectively reduces the ineffective participation of redundant processing nodes, lowers overall computational and waiting overhead, and improves the execution efficiency and result relevance of the processing flow. By performing structured decomposition and sequential reorganization of the information data stream entering the system, the output results are more in line with users' actual usage habits in terms of content organization, granularity, and presentation, thereby improving the understandability, reusability, and user satisfaction of the information processing results. Overall, it enhances the adaptability and collaborative processing effect of the cognitive computing system in complex task scenarios. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the method of the present invention;

[0054] Figure 2 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Example 1: Please refer to Figure 1 As shown in the figure, a hardware method for collaborative computing based on user personality modeling in an embodiment of the present invention includes the following steps:

[0057] S1: Aggregate records by task instance, and perform timing correction on the aggregated records according to the order of operation triggering and contextual dependency to construct a behavioral data set that reflects the user's task progression.

[0058] The process of constructing a behavioral data set reflecting the user's task progression in S1 is as follows:

[0059] The system binds operation logs, information access records, and system feedback records generated by users under the same target to form a set of original operation events. It continuously collects operation logs, information access records, and system feedback records generated by users under the same target, and determines the task instance of the records by identifying the consistency of the task identifier explicitly specified by the user, the source of the operation entry, the processed information object, and the time continuity. For records that meet the same target determination conditions, a unified task instance identifier is assigned to them, and the operation type, triggering object, return result, and timestamp information involved in the records are structured and encapsulated to form a set of original operation events containing multiple types of operation events.

[0060] Based on the trigger timestamps, preceding and following references, and data dependency markers of each operation event, the original set of operation events is initially sorted to construct an operation trigger sequence. Then, the operation events are arranged in chronological order according to their trigger timestamps. Following and following references and data dependency markers are introduced to associate and correct operation events with clear input-output relationships. For example, analysis operations that depend on the results of preceding queries are bound to their reference sources. For operation events with the same timestamp or very short time intervals, their order is determined based on the calling relationship and resource consumption relationship between the events. In this way, while ensuring time consistency, an operation trigger sequence that reflects the logical relationship of the operations is constructed.

[0061] Context dependency constraints are introduced to correct the timing of operation events involving asynchronous returns, cache hits, or lazy loading, eliminating the discrepancy between apparent time and actual perceived order. Asynchronous flags, cache read flags, and lazy response flags in operation events are identified, and combined with the context state associated with the event, the actual effective time of the relevant operation event is re-estimated. For asynchronous return result events, their perceived order is adjusted to after the corresponding triggering operation. For cache hits, their apparent response time is corrected so that it does not affect the user's judgment of the actual decision order, eliminating the discrepancy between apparent time in the operation record and the user's actual perceived order.

[0062] The corrected operation events are merged according to task instances to generate a set of behavioral data reflecting the user's decision-making order and operation rhythm during task progression. After completing the timing correction, the corrected operation events are merged according to the task instance identifier, and operation events belonging to the same task instance are integrated into a continuous sequence of behaviors. During the merging process, the sequential relationship of operation events, the distribution of time intervals, and the switching characteristics of operation types are comprehensively considered to standardize the representation of the behavior sequence. The generated set of behavioral data not only retains the user's decision-making order during task progression but also depicts the operation rhythm, pause frequency, and backtracking characteristics, thus forming a set of behavioral data that can truly reflect the user's task progression method.

[0063] S2: The behavioral data set is split according to the task stage switching point. The split behavioral data is mapped to the operation type space. The information access trigger interval, backtracking operation span and interruption position distribution are segmented and modeled. The recurring operation structure and the isolated operation structure are identified in each segment to form user cognitive feature description data.

[0064] The process of segmenting and modeling the information access trigger interval, backtracking operation span, and interruption location distribution in S2 is as follows:

[0065] Based on changes in operational goals, resource switching, and output format changes in the behavioral data set, the stage switching points in the task progression process are identified; the changes in key attributes in the operational event sequence are continuously monitored, including changes in operational goals, switching of accessed resource types, and changes in the presentation format of output results. When it is detected that a user switches from one type of information object to another or from browsing to editing / analysis, it is marked as a potential stage change signal. Combined with the abrupt changes in the time interval between adjacent operations, the continuously appearing explicit change signals are aggregated and judged to avoid misjudgment caused by occasional operations, and to determine the switching points of each stage in the task progression process.

[0066] Using the stage switching point as the boundary, the behavior data set is divided into several stage behavior subsequences. According to the pre-established operation type mapping rules, the operation events in each stage behavior subsequence are mapped to a unified operation type space. After identifying the stage switching point, the complete behavior data set is divided into multiple stage behavior subsequences using the switching point as the boundary. For each stage behavior subsequence, the operation events contained therein are uniformly encoded according to the pre-established operation type mapping rules. The mapping rules group operations from different sources and triggered by different interfaces but with the same functional semantics into the same operation type, thereby eliminating the impact of interface differences and implementation details. Through this mapping process, each stage behavior subsequence is converted into an operation sequence in a unified operation type space.

[0067] Within the operation type space, the distribution of information access trigger intervals, the span of backtracking operations, and the frequency of operation interruption locations are statistically analyzed to form a segmented behavioral description model within the stage. After completing the operation type space mapping, for each stage of behavioral subsequence, user behavior characteristics are statistically modeled within a unified operation type space. The time interval distribution between adjacent information access operations is calculated to reflect the user's information acquisition rhythm in that stage. The number of operations or time span spanned by the user when performing backtracking operations is statistically analyzed to characterize the intensity of their correction and backtracking behavior. The location and frequency of interruption behaviors in the operation sequence are recorded to reflect cognitive pauses or decision switching within the stage. Based on the statistical results, a segmented behavioral description model describing the behavioral structure characteristics of that stage is formed.

[0068] The process of forming user cognitive feature description data in S2 is as follows:

[0069] In the behavior segmentation description model of each stage, a structured scan is performed on the operation type sequence to extract the operation combination patterns that appear continuously or at intervals. A scanning window combining fixed length and variable length is used to traverse adjacent operation combinations and interval operation combinations in the operation type sequence, and the occurrence position, occurrence interval and occurrence order of each combination are recorded. For combinations that have time intervals but continuous operation semantics, they are identified as the same operation combination pattern by setting the maximum allowable interval threshold. This can completely extract the operation combination patterns that appear continuously or at intervals within the stage.

[0070] Operation combinations that occur more frequently than a set threshold are marked as stable operation structures, representing the user's normal cognitive path. After extracting the operation combination pattern, the frequency of each operation combination in the corresponding stage is statistically analyzed and normalized in combination with the stage length. Operation combinations that occur more frequently than a preset threshold after normalization are marked as stable operation structures. The threshold is set according to historical task data and stage complexity to ensure that the stable operation structure can reflect the operation path repeatedly used by the user in the stage. Operation combinations marked as stable operation structures are used to represent the normal cognitive path and fixed operation habits formed by the user in the process of task advancement.

[0071] Operation combinations that only occur under specific conditions or in a single context are marked as isolated operation structures, representing the user's exploratory behavior. The context in which the operation combination occurs is analyzed, including the types of operations before and after, triggering resources, and result status. When it is confirmed that the operation combination does not constitute a repeated path within the stage and has no stable dependency on the main operation flow of the stage, it is marked as an isolated operation structure. Isolated operation structures are used to represent exploratory, corrective, or tentative operations performed by the user in a specific context. They do not participate in the construction of normal paths, but provide a reference for characterizing the flexibility of user behavior.

[0072] By combining the distribution characteristics of stable and isolated operation structures, user cognitive feature description data reflecting user operation preferences, fault tolerance habits, and stage control methods is generated. The coverage ratio, repetition intensity, and cross-stage consistency of stable operation structures in each stage are statistically analyzed to reflect user operation preferences and process dependence. The frequency of occurrence, triggering conditions, and stage position of isolated operation structures are analyzed to characterize user fault tolerance habits and exploratory tendencies. Based on the analysis results, structured user cognitive feature description data is generated, which can reflect users' personalized cognitive characteristics from the dimensions of stage control methods, operation stability, and behavioral flexibility.

[0073] S3: Inject user cognitive feature description data into the configuration entry point of the information processing flow, and combine and set the calling conditions, execution order and parallel relationship of information parsing nodes, result splicing nodes and intermediate data resident nodes to generate a processing flow definition corresponding to user behavior features.

[0074] The process of injecting user cognitive feature description data into the configuration entry point of the information processing flow in S3 is as follows:

[0075] In the information processing system, a process configuration entry point associated with user tasks is set up to receive external cognitive control parameters. The process configuration entry point can exist in the form of a parameter interface or a configuration description unit and is bound to the task instance identifier. When the user starts a task or the system detects a task switch, the process configuration entry point is activated to receive control data output from the user's personalized modeling module. Through this configuration entry point, personalized configurations associated with specific user tasks can be introduced without modifying the basic structure of the information processing process.

[0076] User cognitive feature description data is parsed into process constraint parameters, node preference parameters, and parallel control parameters; features reflecting user operation stability and stage control methods are parsed into process constraint parameters to limit the execution boundaries of each stage in the process; features reflecting user operation preferences and normal paths are parsed into node preference parameters to indicate the priority and triggering tendency of information processing nodes; features reflecting user exploration behavior and parallel usage habits are parsed into parallel control parameters to adjust the concurrent execution relationship of multiple nodes. Through the parsing process, abstract user cognitive features are transformed into a set of control parameters that can be directly identified and used by the process engine.

[0077] The triggering conditions and participation methods of adjustable nodes in the information processing flow are initially bound to complete the mapping and injection between user cognitive characteristics and the information processing flow. The scope of stages that nodes can participate in is limited according to the flow constraint parameters, the default trigger priority of nodes is set according to the node preference parameters, and the concurrency or serial relationship between nodes is adjusted according to the parallel control parameters. Through the binding process, user cognitive characteristics are mapped and injected into the execution configuration of the information processing flow, thereby realizing the adaptive response of the information processing flow to user behavior characteristics without changing the functional logic of the flow.

[0078] The process of generating the processing flow definition corresponding to user behavior characteristics in S3 is as follows:

[0079] Based on the node preference parameters injected in the process configuration entry, the trigger priority and parsing depth of the information parsing nodes are set; based on the frequency of user access to different information sources and the duration of stay in historical tasks, the trigger priority of the corresponding information parsing nodes is assigned to determine their response order when the process starts or data arrives; based on the degree of user viewing of information details and the frequency of revision, the parsing depth of the information parsing nodes is set, such as limiting the range of parsed fields or expanding the semantic parsing level, so that the information parsing nodes are more in line with the user's actual information processing habits while meeting the process functional requirements;

[0080] Based on users' viewing and revision habits in historical tasks, the merging rules and output granularity of the result splicing nodes are adjusted. If users tend to view intermediate results frequently, the result splicing nodes are configured to segmented output mode to reduce the amount of data spliced ​​in a single session. If users are more concerned about the final summary results, the splicing level is increased to centrally merge results from multiple stages. The granularity level of the result output is adjusted according to the frequency of user modifications to the result content, so that the output results can support both quick browsing and allow users to make fine revisions when needed.

[0081] Based on the user's dependence on intermediate results, the retention time and visibility of intermediate data nodes are configured. For task scenarios where users frequently backtrack or call intermediate results across stages, the retention time of intermediate data is extended and it is set to be visible to subsequent processing nodes and user interface. For scenarios where users use intermediate results less often, the retention time is shortened or it is only temporarily visible within the process. This ensures processing efficiency while taking into account the user's actual needs for using intermediate results.

[0082] The execution order and parallel relationships of various nodes are combined and arranged to form a processing flow definition that matches user behavior characteristics. The execution order of nodes is sorted according to the node trigger priority, and the set of nodes that can be executed in parallel is determined by combining parallel control parameters. For nodes with clear dependencies, their serial execution order is maintained. For nodes that are independent of each other or that users are used to viewing in parallel, they are configured to be executed in parallel. Through combination and arrangement, a processing flow definition that matches user behavior characteristics is generated, so that the information processing process is consistent with the user's cognitive mode in structure.

[0083] S4: During the process of defining and running the process flow, continuously record the user's confirmation, revision and cross-stage call behavior of the phase results, and dynamically adjust the participation set and processing level of the information processing nodes that have not been executed.

[0084] The process of continuously recording user confirmation, revision, and cross-stage call behavior for interim results in S4 is as follows:

[0085] When the processing flow is defined and executed, the stage results output by each information processing node are identified and the corresponding process stage number and node identifier are recorded. When any information processing node completes its processing and generates a stage result, a unique result identifier is assigned to the result and the result identifier is bound to the current process stage number, node identifier and generation timestamp. The result identifier is used to distinguish intermediate results or stage outputs generated by different stages and different nodes, thereby ensuring that user operations can accurately point to the corresponding processing results and provide clear association objects for behavior records.

[0086] When a user performs a confirmation operation, a content revision operation, or calls another stage's result across the current stage, the operation type, trigger time, and associated stage result identifier are recorded respectively. After the stage result is generated and presented to the user, the user's operation behavior in the interactive interface is continuously monitored. When it is detected that a user performs a confirmation operation, a content revision operation, or calls another stage's result across the current stage, the operation is classified and identified, and the operation type, the time point of the operation, and the stage result identifier of the operation are recorded. For cross-stage call behavior, the process stage number to which the called result belongs is recorded to clarify the user's access relationship to different stage results during the process, thereby forming a structured record of user operation behavior.

[0087] The recorded operational behaviors are associated with the current execution stage, the set of completed nodes, and the set of nodes to be executed in the current processing flow to construct a runtime behavior record reflecting the user's intervention position and intervention method during the process. The execution stage, the set of completed information processing nodes, and the set of information processing nodes to be executed when the operation occurs are obtained and bound to the corresponding user operation behaviors. Through association, a runtime behavior record reflecting the user's intervention position, intervention frequency, and intervention method during the process is constructed. The runtime behavior record is continuously updated as the processing flow continues to execute, and is used to characterize the user's real-time decision-making intervention.

[0088] The process of dynamically adjusting the participation set and processing level of unexecuted information processing nodes in S4 is as follows:

[0089] Based on runtime behavior records, the system counts the information processing nodes that are actually triggered, confirmed, or called across stages in the current task instance, and determines the subset of nodes participating in the current process execution. The system also counts the information processing nodes that are actually triggered by the user, the nodes whose corresponding stage results are confirmed by the user, and the nodes whose output results are called across stages by the user. For any information processing node that meets any of these conditions, it is marked as an active node in the current task instance. All active nodes are then aggregated to form a subset of nodes participating in the current process execution, which can distinguish between nodes that the user actually participates in and nodes that have not yet participated or have been ignored.

[0090] For information processing nodes that are not triggered in multiple consecutive task phases, their triggering priority in the processing flow is reduced. For information processing nodes that are called multiple times across phases in the runtime behavior record, their processing level in the processing flow is increased. After determining the node subset, the usage of the nodes is evaluated in stages based on the runtime behavior record. For information processing nodes that are not triggered in multiple consecutive task phases and whose output results are not confirmed or called by the user, their triggering priority in the processing flow is gradually reduced to reduce their execution probability in subsequent processes. For information processing nodes that are called multiple times across phases in the runtime behavior record or whose results are frequently used in subsequent processing, their processing level is increased so that they are in a more prominent or core position in the process execution structure. The adjustment is completed through parameterization, without changing the functional logic of the nodes, only affecting their participation level and execution order.

[0091] The adjustment results of node trigger priority, processing level, and node execution order are written into the process status description corresponding to the current task instance, and the information processing node participation set is updated. The process status description is used to record the trigger priority, processing level, and relative position of the node in the node execution sequence. Based on the updated process status description, the participation set of the information processing node is updated synchronously to clarify the scope of nodes that should participate in the execution in the current process. By continuously updating the process status description and the node participation set, the information processing process can dynamically respond to changes in user behavior without rebuilding the process definition.

[0092] S5: Based on the adjusted set of information processing nodes, perform structured splitting and sequential reorganization of the information data stream entering the system, and complete the information output according to the processing flow definition.

[0093] The process for completing information output according to the processing flow definition in S5 is as follows:

[0094] Based on the set of participating information processing nodes, the input information data stream is split at the field level and semantic level. According to the predefined data structure template, the information data stream is split at the field level, and the data is divided into basic field units. Combining the data source identifier and content description information, the split field units are semantically identified and divided into data units of different semantic categories. Through the two-level splitting at the field level and semantic level, the complex information data stream is transformed into a set of data units that can be processed independently by different information processing nodes.

[0095] According to the updated processing flow definition, the split information data units are reordered and distributed to nodes. Based on the node trigger priority and execution order recorded in the process status description, the data units are sorted according to the processing order. According to the processing level and participation status of the nodes, the corresponding data units are distributed to the matching information processing nodes. For nodes configured for parallel execution, the relevant data units are distributed at the same time. For nodes that need to be processed serially, they are distributed sequentially to ensure that the information data units are correctly scheduled according to the processing logic corresponding to the user behavior characteristics during the process execution.

[0096] After each information processing node completes its corresponding processing, the intermediate results are aggregated and their consistency is verified. The information output results are generated according to the output rules defined in the processing flow. According to the result splicing rules in the processing flow definition, the intermediate results from different nodes are structurally aligned and their contents are merged. The merged results are then subjected to consistency verification to avoid result conflicts caused by the execution order or parallel processing of nodes. When the consistency verification passes, the final information output results are generated according to the output format, output granularity and presentation method set in the processing flow definition.

[0097] Example 2: As Figure 2 As shown, a computational collaborative processing hardware system based on user personality modeling includes:

[0098] Task aggregation module: Collects user operation logs through high-speed caching, with timestamp accuracy down to the nanosecond level.

[0099] Stage Modeling Module: Utilizes embedded AI coprocessors to analyze user behavior characteristics, identify task stages, and generate models.

[0100] Process configuration module: Dynamically loads node configuration bitstreams via FPGA to achieve millisecond-level reconfiguration.

[0101] Process regulation module: Provides real-time feedback on the activation status of the controller monitoring nodes and dynamically adjusts the FPGA routing table.

[0102] Data Reassembly Module: The dual-channel DMA engine performs field-level splitting / reassembly of input data, directly connecting to host memory and local storage.

[0103] Example 3: A computing co-processing hardware system based on user-personalized modeling also includes a main control SoC, a cache, an embedded AI coprocessor, a reconfigurable logic unit, a real-time feedback controller, a dual-channel DMA engine, local storage, and a security unit; the local storage includes a data memory pool and a knowledge memory pool, with the data memory pool using a 2TB NVMe SSD and the knowledge memory pool using a 512GB 3D XPoint; the security unit includes a hardware encryption engine, a secure boot mechanism, and a data isolation mechanism, with the hardware encryption engine supporting SM4 / SM9 / AES-256 full pipeline encryption, and a unique device key based on PUF ensuring firmware integrity.

[0104] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0105] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hardware method for collaborative computing based on user personality modeling, characterized in that, Includes the following steps: Aggregate records by task instance, and perform timing correction on the aggregated records according to the order of operation triggering and context dependency to construct a behavioral data set that reflects the user's task progression. The behavioral data set is split according to the task stage switching point. The split behavioral data is mapped to the operation type space. The information access trigger interval, backtracking operation span and interruption position distribution are segmented and modeled. The recurring operation structure and the isolated operation structure are identified in each segment to form user cognitive feature description data. User cognitive feature description data is injected into the configuration entry of the information processing flow. The calling conditions, execution order and parallel relationship of information parsing nodes, result splicing nodes and intermediate data resident nodes are combined and set to generate a processing flow definition corresponding to user behavior features. During the process of defining and running the workflow, the user's confirmation, revision and cross-stage call behavior of the phase results are continuously recorded, and the participation set and processing level of the information processing nodes that have not been executed are dynamically adjusted. Based on the adjusted set of information processing nodes, the system performs structured splitting and sequential reorganization of the information data stream entering the system, and completes information output according to the defined processing flow.

2. The computational collaborative processing hardware method based on user personality modeling according to claim 1, characterized in that, The process of constructing a behavioral data set that reflects how users advance tasks is as follows: Task instance identifiers are bound to the operation logs, information access records, and system feedback records generated by users under the same target, forming a set of original operation events; Based on the trigger timestamps, preceding and following references, and data dependency markers of each operation event, the original set of operation events is initially sorted to construct the operation trigger sequence; By introducing context dependency constraints, timing corrections are made for operation events that involve asynchronous returns, cache hits, or lazy loading, eliminating the discrepancy between apparent time and actual perceived order. The corrected operation events are merged according to task instances to generate a set of behavioral data reflecting the user's decision-making order and operation rhythm during task progress.

3. The computational collaborative processing hardware method based on user personality modeling according to claim 2, characterized in that, The process of segmenting and modeling the information access trigger interval, backtracking operation span, and interruption location distribution is as follows: Based on changes in operational objectives, resource switching, and output format changes in the behavioral data set, identify the stage switching points in the task advancement process; Using the phase switching point as the boundary, the behavioral data set is divided into several phased behavioral subsequences. According to the pre-established operation type mapping rules, the operation events in each phased behavioral subsequence are mapped to a unified operation type space. Within the operation type space, the distribution of information access trigger intervals, the range of backtracking operations, and the frequency of operation interruption locations are statistically analyzed to form a segmented description model of behavior within a stage.

4. The computational collaborative processing hardware method based on user personality modeling according to claim 3, characterized in that, The process of forming user cognitive feature description data is as follows: In the behavior segmentation description model at each stage, a structured scan is performed on the operation type sequence to extract the operation combination patterns that appear continuously or at intervals. Operation combinations that occur more frequently than a set threshold are marked as stable operation structures, representing the user's normal cognitive path; Combinations of operations that occur only under specific conditions or in a single context are marked as isolated operation structures, representing the user's exploratory behavior; By combining the distribution characteristics of stable and isolated operation structures, user cognitive feature description data reflecting user operation preferences, fault tolerance habits, and stage control methods is generated.

5. A computational collaborative processing hardware method based on user personality modeling according to claim 4, characterized in that, The process of injecting user cognitive feature description data into the configuration entry point of the information processing flow is as follows: Set up a process configuration entry point associated with user tasks in the information processing system to receive external cognitive control parameters; The user cognitive feature description data is parsed into process constraint parameters, node preference parameters, and parallel control parameters; The triggering conditions and participation methods of adjustable nodes in the information processing flow are initially bound to complete the mapping and injection between user cognitive characteristics and information processing flow.

6. The computational collaborative processing hardware method based on user personality modeling according to claim 5, characterized in that, The process of generating the processing flow definition corresponding to user behavior characteristics is as follows: Based on the node preference parameters injected in the process configuration entry, the trigger priority and parsing depth of the information parsing node are set; Based on users' habits of viewing and revising results in historical tasks, the merging rules and output granularity of result splicing nodes are adjusted; Configure the retention time and call visibility of intermediate data nodes based on the user's dependence on intermediate results; The execution order and parallel relationships of various nodes are combined and arranged to form a processing flow definition that matches the user's behavior characteristics.

7. A hardware method for collaborative computing based on user personality modeling according to claim 6, characterized in that, The process of continuously recording users' confirmation, revision, and cross-stage call behavior of interim results is as follows: When the process is defined and executed, the stage results output by each information processing node are identified and the corresponding process stage number and node identifier are recorded. When a user performs a confirmation operation, a content revision operation, or calls the results of other stages across the current stage on a stage result, the operation type, the trigger time, and the associated stage result identifier are recorded respectively. The recorded operations are associated with the current execution stage, the set of completed nodes, and the set of nodes to be executed in the current processing flow, to construct a runtime behavior record that reflects the user's intervention position and intervention method during the process.

8. A computational collaborative processing hardware method based on user personality modeling according to claim 7, characterized in that, The process of dynamically adjusting the participation set and processing level of unexecuted information processing nodes is as follows: Based on runtime behavior records, the information processing nodes that are actually triggered, confirmed, or called across stages by the user in the current task instance are counted to determine the subset of nodes participating in the execution of the current process; For information processing nodes that are not triggered in multiple consecutive task phases, lower their trigger priority in the processing flow; for information processing nodes that are called multiple times across phases in the runtime behavior record, increase their processing level in the processing flow. Write the adjustment results of node trigger priority, processing level and node execution order into the process status description corresponding to the current task instance, and update the information processing node participation set.

9. A hardware method for collaborative computing based on user personality modeling according to claim 8, characterized in that, According to the processing flow definition, the process of completing information output is as follows: Based on the set of information processing nodes, the input information data stream is split at the field level and semantic level. According to the updated processing flow definition, the split information data units are sequentially reorganized and distributed to nodes. After each information processing node completes its corresponding processing, the intermediate results are aggregated and their consistency is verified. Then, the information output results are generated according to the output rules defined in the processing flow.

10. A computational collaborative processing hardware system based on user personality modeling, applied to the method described in any one of claims 1-9, characterized in that, include: Task aggregation module: It aggregates user operation records by task instance and corrects the execution sequence of the records according to the order of operation triggering and contextual dependencies to form a behavioral data set; Stage modeling module: The behavioral data set is split according to the task stage switching points, and the split behavioral data is mapped to the operation type space to perform segmented modeling of the operation characteristics within each stage; Process configuration module: Receives user cognitive feature description data, combines and sets the calling conditions, execution order and parallel relationship of each processing node in the information processing process, and generates the processing process definition; Process adjustment module: Records user actions on phased results during the execution of the processing flow definition, and adjusts the set and processing level of unexecuted processing nodes; Data Reassembly Module: Based on the adjusted set of processing nodes, the module splits and reassembles the incoming data stream into the system, and completes the information output according to the defined processing flow.

Images