68 results about "Outcomes evaluation" patented technology

The invention discloses a large language model agent system and method and related equipment. The large language model agent system comprises an input analyzer module, a strategy generation module, an execution controller module, a task result evaluation module, an reflection module and a strategy rewriting module, and a self-adaptive optimization closed loop for task execution is formed through the reflection module and the strategy rewriting module. Wherein the reflection module can generate a report containing failure reasons, and the problem of failure diagnosis is solved; the strategy rewriting module optimizes a plan according to a report, adjusts tool calling, and realizes self-learning and strategy reconstruction; the closed-loop design forms a complete cognitive process, and an input analyzer guarantees the stability of prompt. According to the system, the task completion rate, the execution stability and the scene adaptability are remarkably improved, the robustness and the error tolerance in a multi-step complex task are particularly enhanced, and the key blank of an existing language model agent system in the aspects of self-reflection and self-optimization is filled.

The invention discloses a project similarity evaluation method based on semantic comprehension and multi-dimensional feature comparison, and the method comprises the steps: obtaining the project text data of a current declaration project, carrying out the semantic analysis and semantic mode understanding processing, extracting the text semantic features according to the semantic mode, and obtaining the project semantic features; historical declaration project data is obtained to carry out key text feature extraction processing, a project management model is constructed according to key text features and corresponding project declaration stages, and project semantic features are input into the model to carry out multi-dimensional semantic feature comparison processing. And evaluating the item similarity according to a multi-dimensional feature comparison result to obtain a repeated item approval evaluation result of the item, performing feature independent comparison processing on the item semantic feature of each sub-item of the current declaration item, and performing independent evaluation on the sub-item similarity to obtain a sub-item similarity evaluation result of the current declaration item. The method and the device have the effects of effectively evaluating the item similarity and improving the accuracy of item similarity evaluation.

The invention provides a fine tuning method of a preset model, and a question answering method and device based on the preset model, and relates to the technical field of artificial intelligence. The fine tuning method comprises the following steps: rewriting a sample problem by utilizing a preset model to obtain a retrieval problem; performing retrieval based on the retrieval problem to obtain a retrieval result for the retrieval problem; performing reflection on the retrieval result by utilizing a preset model to obtain a reflection result aiming at the retrieval result; under the condition that the reflection result represents that the retrieval result meets a preset answer generation condition, generating an answer of the sample question based on the retrieval result by utilizing a preset model; respectively evaluating the retrieval result and the answer to obtain a process evaluation score for the retrieval result and a result evaluation score for the answer; and performing fine tuning on parameters of a preset model based on a training sample constructed from the sample question, the answer, the retrieval result, the process evaluation score and the result evaluation score.

The invention relates to a clearing result evaluation method and device for an electric power spot market, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of collecting electric power index data of an electric power spot market; establishing an index system of the to-be-evaluated object based on the power index data; according to the index system, obtaining the connection degree between each evaluation index and the state level by using the clearing result index layer evaluation model; according to the connection degree between each evaluation index and the state level, utilizing a clearing result target layer evaluation model to obtain a comprehensive basic probability distribution value of each evaluation index under different state levels; and according to the comprehensive basic probability distribution value, performing evaluation operation to obtain a state level of a clearing result of the to-be-evaluated object of the electric power spot market. By adopting the method, the evaluation weight can be adaptively adjusted, and an evaluation system of a multi-dimensional index network is fused, so that the transparency of a clearing result is improved.

The invention provides an intelligent reading control system, which is applied to the technical field of intelligent reading and comprises an acquisition module used for acquiring multi-modal original data in a reading process of a user in real time; the extraction module is used for extracting depth features of the multi-modal original data, generating behavior cognition feature vectors and calculating feature fusion coefficients; the evaluation module is used for outputting an evaluation result and evaluation confidence; the adaptation module is used for performing deep semantic analysis on the reading material, constructing a knowledge graph and calculating semantic association strength; based on the evaluation result, the evaluation confidence and the semantic association strength, generating multi-type and multi-level adaptive difficulty questions; the recommendation module is used for constructing a user ability portrait based on an evaluation result, calculating an interest matching degree and recommending personalized books, learning tasks and auxiliary materials; the intervention module is used for judging whether intervention is needed or not based on a preset intervention triggering threshold value, so that the learning efficiency of the user is improved.

This invention discloses a method and system for automatically extrapolating dynamic test schemes based on one-click planning rules. In the test planning and design phase, based on preset one-click planning rules, a dynamic test scheme including test scenarios, test factors, and test samples is automatically generated, and automated rules and equipment action modes are configured for the simulation extrapolation process. In the test execution phase, according to the generated dynamic test scheme, simulation resources are automatically scheduled to execute the simulation extrapolation without human intervention. During the extrapolation process, based on the automated rules and equipment action modes, command issuance, autonomous equipment behavior response, data collection, and result evaluation are automatically completed, forming a fully automated simulation extrapolation closed loop. This invention enables a top-down automated process, forming a complete solution for the entire process of automated simulation extrapolation in the "human-in-the-loop" environment, helping users solve various problems encountered in the construction of current "human-in-the-loop" simulation models.

The invention discloses an energy-based economic key element linkage prediction method and system, and relates to the technical field of economics and prediction modeling, and the method comprises the steps: determining and collecting a data source, and carrying out the preprocessing of the obtained data; exporting key business data based on the multi-dimensional data model, and constructing an energy consumption data and economic growth prediction model; training an energy consumption data and economic growth prediction model, and performing result evaluation and model optimization; and economic growth and energy prediction are carried out by using the optimized model, and analysis is carried out according to a prediction result and a corresponding decision is provided. By constructing the data cube model, data from different fields can be efficiently integrated, multi-dimensional comprehensive analysis is realized, and the comprehensiveness of the prediction model is improved; according to the method, the optimized model is utilized, a more flexible nonlinear modeling method is adopted, the complex relation between economic growth and energy consumption is better captured, and the prediction accuracy is improved.

The invention provides a model evaluation method and device. The method comprises the steps that a scoring system determines a first score of a first reply and a first explanation of the first score according to the first evaluation question, a first reference answer of the first evaluation question and the first reply of the first evaluation question, and the first reply is the first reply of a to-be-evaluated model to the first evaluation question; the first reference answer is a preset correct answer for the first evaluation question; the first explanation is used for representing a reason for giving the first score; the scoring system outputs the first score and the first explanation to an evaluation person or a receiving module; the first score and the first explanation are used for determining an evaluation result, and the evaluation result is used for evaluating the accuracy of replying the first evaluation question by the to-be-evaluated model. By adopting the method, the scoring reason can be explained while the reply to the to-be-evaluated model is scored, so that the efficiency of generating a complete evaluation result is improved.

This invention provides a resource scheduling method, apparatus, device, storage medium, and computer program product, relating to the field of communication technology. The method includes: responding to a task offloading request sent by a terminal device, publishing the task to be offloaded carried in the task offloading request to a nearby MEC device; receiving an evaluation result of the task to be offloaded sent by the nearby MEC device; the evaluation result is used to characterize whether the nearby MEC device has the capability to process the task to be offloaded; if, based on the evaluation result, it is determined that the task to be offloaded needs to be migrated, at least one target MEC device is determined according to the energy consumption information and latency information of the MEC device, and the task to be offloaded is scheduled to the target MEC device. In this invention, tasks do not need to be offloaded from a cloud central node to an edge device; instead, the local edge device accepts the task request, employs a distributed computing strategy, and autonomously completes resource allocation, achieving decentralized resource scheduling.

The invention discloses a water pumping test intelligent analysis system and method, relates to the technical field of hydrogeological exploration, and aims to solve the problems of low analysis efficiency and poor accuracy of a water pumping test analysis system in the prior art. According to the system, a test configuration module is used for determining target test configuration, and a data input module is used for data input and data analysis to obtain target original data; the intelligent preprocessing module is used for performing data preprocessing on the target original data to obtain preprocessed data; the multi-algorithm calculation engine is used for calculating the preprocessed data by adopting multiple calculation methods to obtain target hydrogeological parameters; the visual analysis module is used for generating a plurality of target charts based on the target hydrogeological parameters; the result evaluation module is used for generating a result comparison table for multiple calculation methods and determining the permeability grade for the target original test data; therefore, the analysis efficiency and accuracy of the water pumping test are improved.

The invention relates to a method for realizing a resource adaptive pressure test based on load scaling. The method comprises the following steps: load modeling; constructing a distributed test cluster; service simulation and Mock construction are carried out; load prediction; self-adaptive pressurization is carried out; and evaluating a result. The invention also relates to a system and a device for realizing the resource adaptive pressure test based on load scaling, a processor and a computer readable storage medium thereof. By adopting the method, the system and the device for realizing the resource adaptive pressure test based on the load scaling, the processor and the computer readable storage medium thereof, the resource utilization rate is greatly improved, the test efficiency is greatly improved, manual intervention is reduced through an automatic test process, and the test period is shortened. The environment adaptability is enhanced, and the method has good universality for different types of service systems. The cost effectiveness is outstanding, the dependence on a special high-performance test environment is greatly reduced, the existing low-configuration equipment is fully utilized, and the hardware investment and the maintenance cost are saved.

The application discloses a kind of based on temperature control curve's forging heat treatment process optimization method, it is related to data processing technical field, including, based on process task data packet, temperature control performance prediction model is constructed and reverse solution is executed, generates core temperature control reference curve;To core temperature control reference curve executes heat response mapping, obtains furnace temperature control trajectory, and obtains online core tracking control scheme by heat treatment risk timing evaluation;Online core tracking control scheme is executed, process data is recorded synchronously, and combined furnace temperature control trajectory is passed through heat history inversion operation, obtains execution temperature control curve;Based on execution temperature control curve, temperature control performance prediction model is optimized, and obtains process quality evaluation result by performance deviation evaluation.The application is by constructing temperature control performance prediction model and executing reverse solution and feedback optimization, improves the calculation consistency of process parameter design and the quantifiable and iterative optimization ability of result evaluation.

The invention relates to the technical field of intelligent monitoring, and discloses a system operation state monitoring method based on multi-source information mining and organic fusion, which comprises the following steps: acquiring internal operation state data and external operation state data, calibrating the internal operation state data and the external operation state data by using a sensor calibration algorithm based on artificial intelligence; the method comprises the following steps: extracting internal operation state data and external operation state data, performing organic fusion to form a comprehensive data set, performing correlation analysis on the comprehensive data set, mining a correlation relationship between the internal operation state data and the external operation state data, and deriving key variables influencing the operation state of a monitoring target system; and comparing the real-time data of the key variables with a preset normal operation range to obtain a comparison result, evaluating the current operation state of the monitoring target system based on the comparison result, obtaining an evaluation result, and when the evaluation result exceeds a preset standard value, triggering an early warning mechanism to prompt a user to take corresponding measures. According to the invention, the monitoring precision of the monitoring target can be improved.

We provide the system. [Solution] A means of receiving student information and generating student avatars based on that information, A means of receiving information on the personnel needs of a company and generating interviewer avatars based on that information, A method for generating student avatars and interviewer avatars to conduct interviews with each other and evaluate the degree of matching based on the results, A means of notifying each user of the evaluation results, A system that includes this.

The invention discloses an enterprise human resource management evaluation method and system. The method comprises the following steps: acquiring, analyzing and evaluating management process data to obtain an index evaluation value of each preset monitoring index; determining a weight corresponding to each preset monitoring index, and performing calculation based on the weight and the index evaluation value to obtain an initial evaluation value of enterprise human resource management; obtaining management result data of the enterprise for analysis, and determining result parameter data related to enterprise human resource management; evaluating the enterprise human resource management result based on the result parameter data to obtain a management result evaluation value, and determining a correction coefficient based on the management result evaluation value; and correcting the initial evaluation value based on the correction coefficient to obtain a comprehensive management evaluation value. According to the invention, the transformation of human resource management from an empirical type and qualitative type to a data type and quantitative type is realized, an intelligent evaluation system capable of self-calibration and continuous optimization is established, and powerful technical support is provided for enterprises to improve the human resource management level.

The present disclosure provides a virtual world-based data processing method, system, device and electronic equipment, relates to the field of artificial intelligence, in particular to the technical field of augmented reality, virtual reality, computer vision, deep learning and the like, and can be applied to the scene of metaverse, virtual digital person and the like. The implementation scheme is as follows: obtaining a to-be-evaluated strategy and region data of a target region in which the to-be-evaluated strategy is executed, the region data at least including region deployment data and social activity data, the to-be-evaluated strategy being a strategy to be executed in the target region, and the to-be-evaluated strategy being used to balance economic development and resource and environment development in the target region; rendering a virtual world of the target region based on the region deployment data and the social activity data, to obtain a target virtual world of the target region; obtaining a strategy execution result of executing the to-be-evaluated strategy in the target virtual world; evaluating the to-be-evaluated strategy based on the strategy execution result, to obtain an evaluation result, the evaluation result representing an execution effect of executing the to-be-evaluated strategy in the target region.

An evaluation method, apparatus, medium, and electronic device for intelligent agents and large language models are disclosed, relating to the fields of large model technology and intelligent agent technology. The evaluation method includes: creating an evaluation task for evaluating at least one dialogue of multiple objects to be evaluated in a dialogue process from at least one evaluation dimension; executing the evaluation task to obtain evaluation results for comparing the dialogue capabilities of the multiple objects to be evaluated; displaying the evaluation results on a first interface and then displaying a second interface; and obtaining a configured evaluator in response to a configuration operation for an evaluator of a target type on the second interface. The configured evaluator is used to evaluate the dialogue from the corresponding evaluation dimension, facilitating user analysis of the building effects of intelligent agents with different configurations. Furthermore, custom configurations can be performed based on the configuration operation to obtain evaluators with different evaluation dimensions.

The embodiment of the invention relates to a method and device for determining an execution strategy, equipment, a storage medium and a program product. The method provided herein includes: acquiring a candidate strategy associated with a robot system, the candidate strategy indicating a sequence of candidate actions for the robot system to perform a task; based on the candidate action sequence, a risk threshold value corresponding to the candidate strategy is determined, and the risk threshold value indicates a risk upper limit allowed by the robot system to execute the candidate strategy; based on a plurality of risk components corresponding to the candidate action sequence in the plurality of dimensions and a plurality of component thresholds corresponding to the risk threshold in the plurality of dimensions, determining an evaluation result, the evaluation result indicating feasibility of the candidate strategy, and the plurality of dimensions indicating multiple constraints associated with the candidate strategy; and determining an execution strategy for the robot system to execute the task based on the evaluation result. In this way, according to the embodiment of the invention, the safety risk of the robot system in the task execution stage can be remarkably reduced.

The invention discloses an intelligent project auditing method and system based on multi-agent fusion, and the method comprises the steps: obtaining a project application text, carrying out the pre-auditing processing, analyzing the content logic relation of the project text after the pre-auditing, and obtaining project text auditing data, extracting text features of the project text review data, analyzing potential problems in the project declaration document in combination with the corresponding content logic relationship to obtain project problem analysis data, and performing multi-dimensional compiling processing on the project text review data and the project problem analysis data according to a preset project review model to obtain project text review data and project problem analysis data; and obtaining a project review analysis report conforming to the corresponding project review model, performing key data verification processing on the project review analysis report, evaluating the whole project declaration compliance according to a key data verification result, and feeding back and adjusting the whole project review process according to a compliance evaluation result. The method has the effects of improving the auditing accuracy of massive complex project declaration materials and improving the project material auditing efficiency.

The invention discloses a forging heat treatment process optimization method based on a temperature control curve, which relates to the technical field of data processing, and comprises the following steps: constructing a temperature control performance prediction model and executing reverse solution based on a process task data packet to generate a core temperature control reference curve; thermal response mapping is carried out on the center temperature control reference curve, a furnace temperature control track is obtained, and a line center tracking control scheme is obtained through thermal treatment risk time sequence evaluation; executing a centroid tracking control scheme, synchronously recording process data, and obtaining an execution temperature control curve through thermal history inversion operation in combination with a furnace temperature control track; and optimizing the temperature control performance prediction model based on the execution temperature control curve, and obtaining a process quality evaluation result through performance deviation evaluation. According to the method, by constructing the temperature control performance prediction model and executing reverse solution and feedback optimization, the calculation consistency of process parameter design and the quantifiable and iterative optimization capability of result evaluation are improved.

The invention provides an intelligent agent evaluation method and device and a storage medium, relates to the technical field of artificial intelligence, and can comprehensively evaluate an intelligent agent. The method comprises the following steps: acquiring trajectory data in a process of executing a preset evaluation task by an intelligent agent group; the agent group comprises at least one agent; based on the trajectory data, determining a score of each preset index in the plurality of preset indexes; the plurality of preset indexes comprise at least one of the following items: task execution rate, quality score, throughput, response delay, system resource consumption and information interaction times among the intelligent agents; and based on the score of each preset index, a preset evaluation task and an evaluation model, determining an evaluation result of the agent group, the evaluation model being used for determining the evaluation result of the agent group.

The invention discloses a reasoning enhancement method and device, equipment, a storage medium and a product. The method comprises the following steps: acquiring a debate theme, an initial parameter and a target model; according to the initial parameters and the target model, respectively generating an initial claim argument and an initial opposite argument aiming at the debate theme; iteratively generating a new claim argument and a new opposite argument based on the initial claim argument and the initial opposite argument until a set iteration termination condition is reached; and performing debate result evaluation according to the context data in the argument iteration process, and determining a target conclusion according to an evaluation result.

To easily provide a result based on more quantitative evaluation for a change in layout.SOLUTION: The evaluation system A1 includes an environmental information acquirer 11, a request acquirer 12, an evaluation processor 13, and an outputter 14. The environmental acquirer 11 acquires first environmental information S1 including one or more environmental elements in the target space B1 before the layout change and second environmental information B2 including one or more environmental elements after the layout change. The request acquirer 12 acquires the change request information C1. The evaluation processing unit 13 derives an evaluation result. The output unit 14 outputs the evaluation result. The evaluation processor 13 obtains differential information of the second environmental information B1 with respect to the first environmental information B2. The evaluation processing unit 13 determines thresholds in accordance with the change request information C1. The evaluation processing unit 13 derives an evaluation result based on a comparison result between the difference information and the threshold value.SELECTED DRAWING: Figure 1

The present disclosure relates to a task planning method and device based on a large language model, an electronic device, a storage medium and a computer program product. The method comprises: for each task in a plurality of robot tasks, randomly sampling target parameters from a parameter set corresponding to the task; generating a target task instruction in natural language form based on the target parameters; constructing a structured step true value sequence corresponding to the target task instruction; inputting the target task instruction into a pre-trained large language model; comparing a structured step prediction sequence with the structured step true value sequence; and based on the comparison result, evaluating the inference performance of the large language model. In this way, since the present disclosure can automatically and quantitatively evaluate the structured inference ability of the large language model under fixed task constraints, the adverse effects of human factors on the inference performance evaluation of the large language model caused by relying on artificial labeling or subjective judgment can be effectively avoided.

The invention discloses an island forest carbon flux monitoring and evaluation system and method based on multi-source data fusion. The system comprises a data set construction module, a collaborative relationship analysis module, a dynamic model construction module, an update mode classification module and a result evaluation module. According to the method, the multi-time scale change rule of the island forest carbon flux can be accurately monitored by integrating the multi-source data and constructing the dynamic model, and meanwhile, the island forest carbon flux is comprehensively evaluated according to the quantitative relation between the carbon flux and the carbon sink function in different updating modes. The method not only improves the understanding of the carbon balance of the island forest ecological system, but also provides a scientific basis for formulating a targeted optimization strategy, effectively enhances the stability of the carbon sink function of the island forest in the process of coping with climate change and post-disaster recovery, and provides powerful support for ecological protection and sustainable development.

The invention discloses a social media data visual subsystem based on a parameter insight associated agent. The system comprises a dialogue module which displays interaction between a user and the agent and receives and analyzes a problem; the action analysis path tree module presents an intelligent agent exploration path in a tree structure, and displays an operation node through a node aggregation and hiding strategy of the tree structure; the mining parameter-result association analysis module supports interactive optimization by using a parallel coordinate graph visualization algorithm hyper-parameter and result evaluation index relationship; the insight visualization module displays insight discovery in an image-text form and supports linkage traceability with the action analysis path tree module; and the rear-end calculation module is responsible for data processing and interaction coordination among the modules. According to the method, full-link transparency in the LLM intelligent agent exploration process is achieved, users are supported to intervene in parameter tuning and path selection in real time, automatic analysis and user control are balanced through deep fusion of multi-modal interaction and automatic visual generation, and the interpretability and credibility of social media data insight are improved.

The model training method comprises the following steps: constructing a training data set required by the round of iterative training, and training a first model based on the training data set to obtain a second model; constructing a regression test set required by the current round of iterative evaluation, and evaluating the second model based on the regression test set to obtain a first evaluation result; and on the basis of the first evaluation result and a second evaluation result, the correctness of the second model in the current iterative optimization direction is evaluated, and the second evaluation result is obtained by evaluating the second model on the basis of the general test set. Therefore, in the model training process, two-stage testing is carried out based on a regression test set constructed by current network data of each round of iterative sampling and a general evaluation set containing various basic capability evaluations to fully evaluate the correctness of the optimization direction and guide the next step of planning, so that the optimization direction can be prevented from deviating and the model effect can be prevented from being degraded.

The invention discloses a clustering and multi-classification model-based power anomaly detection method. The method comprises the following steps of: preprocessing data; constructing an abnormal sample; feature selection; performing clustering analysis; training and selecting a classification model; selecting evaluation indexes; compared with a traditional single machine training model detection method, the clustering and multi-classification model-based power anomaly detection method provided by the invention aims to perform personalized selection for power consumption behaviors of different customer groups through refined modeling and find out the most suitable detection model for different customer groups, so that the accuracy of power anomaly detection is improved. According to the method, the optimal model can be selected, more accurate electricity stealing detection capability can be provided under different types of electricity consumption behavior modes according to different characteristics of customer groups, the accuracy and the adaptability of electricity stealing detection are remarkably improved through the detailed modeling strategy, and particularly, the accuracy and the adaptability of electricity stealing detection are improved in the case of electricity consumption fluctuation of different customer groups. The electricity stealing behavior can be effectively detected, and the electricity stealing behavior can be quickly and accurately identified.

The invention discloses a test case evaluation method and computing equipment, and relates to the technical field of computing equipment. The method comprises the steps of obtaining evaluation data of a target test case in multiple dimensions; wherein the multiple dimensions are at least two of the following dimensions: a static code dimension, a dynamic execution dimension or a community feedback dimension; the evaluation data under the static code dimension is feature data of codes of the target test case, the evaluation data under the dynamic execution dimension is operation data of the target test case, and the evaluation data under the community feedback dimension is use feedback data of the target test case. The target test case is evaluated according to the evaluation data of the target test case in the multiple dimensions, an evaluation result of the target test case is determined, the evaluation result comprises the quality grade and / or the quality score of the target test case, and the quality score is in positive correlation with the quality grade. The method is used for quality evaluation of the test case.

This application provides a method and related apparatus for assessing students' programming abilities based on multimodal data. The method includes: firstly, systematically collecting multimodal data generated by students throughout the entire programming training process, such as interactive behaviors, source code and modification sequences, running results and performance indicators, and extracting them into three types of feature vectors representing process, quality and effectiveness. These features are then input into a pre-trained multidimensional ability assessment model, which maps the above features to professional ability standard items, outputs a fine-grained ability mastery score and identifies ability gaps. Subsequently, based on these gaps, a personalized sequence of subsequent training tasks is dynamically generated. By executing tasks and collecting new data, a closed-loop tracking and continuous updating of the ability status is achieved. Finally, all information is integrated to generate a visualized ability assessment report, which can improve the support effectiveness and accuracy of the assessment results for personalized teaching guidance, and effectively help programming education transform from outcome evaluation to process empowerment.