21 results about "Position Descriptions" patented technology

The invention discloses a position information auditing method and device based on multi-modal information. Comprising the steps of converting a position description text and enterprise entity information into a predefined structured data format, and generating a system cue word matched with the position description text to be audited based on structured enterprise information and structured position information, the system cue word comprises an auditing rule used for performing association analysis between the structured position information and the structured enterprise information on N target position auditing dimensions; the N target position auditing dimensions at least comprise consistency verification between structured position information and enterprise qualification text information, enterprise credibility evaluation based on enterprise historical release position records, and compliance identification of position description texts; and inputting the system cue word into a pre-trained large language model to obtain an audit result for the position description text to be audited. According to the invention, the identification accuracy of false recruitment positions can be improved.

The invention discloses a person and post matching method, device and system and a storage medium. The person and post matching method comprises a first stage and a second stage, wherein the first stage is pre-matching, quantitative evaluation pre-matching is carried out based on structured information extracted from resumes and position description, and candidates basically meeting position conditions are screened out; the second stage is a model training and application stage, firstly, a mixed deep learning architecture of a double-tower model and a cross encoder is adopted to construct a resume coding tower, a position coding tower and a cross encoder model, so that deep semantic matching analysis and vector similarity matching are performed on candidate resumes passing pre-matching, and then the candidate resumes are obtained. The candidate resumes with the highest similarity of the TOPK are obtained, then the candidate resumes of the TOPK are finely arranged, and finally a matched result list is output.

The invention discloses an intelligent resume screening method, system and device based on a large language model and a medium, and the method specifically comprises the steps: based on an initial weight, analyzing position description through the large language model, and extracting key skill requirements; analyzing the matching degree of each resume, and generating a matching degree score through key skill requirements; determining a resume screening result according to the matching degree score, and collecting an adjustment operation and a labeling reason of the HR on the resume screening result; dynamically optimizing weight distribution through a reinforcement learning model according to the adjustment operation and the marking reason; automatically adjusting the skill priority according to a post demand prediction model by using the optimized weight; and based on the skill priority, identifying the deep matching relationship between the candidate skill and the post demand in the resume screening result, and evaluating the potential index of the candidate to adapt to the new post according to the deep matching relationship and the industry talent flow data. According to the invention, efficient, accurate and dynamically adaptive intelligent resume screening is realized, and the recruitment quality and efficiency are improved.

The invention relates to the technical field of medical intelligent informatization, and discloses a thyroid cancer electronic medical record processing method and system based on multi-modal data, and the method comprises the steps: obtaining a thyroid image and an associated electronic medical record text containing lesion position description; extracting a text orientation semantic vector representing a relative spatial relationship from the text, and extracting a focus morphology feature vector representing a visual morphology and an image spatial relationship vector representing a spatial position relative to a preset anatomical structure from the image; respectively calculating an orientation semantic consistency score and a morphological neighborhood consistency score, and fusing to obtain a comprehensive consistency score; and finally, comparing the comprehensive consistency score with a preset threshold value to determine whether the text description corresponds to the focus in the image. According to the method, cross-modal accurate matching is realized, the diagnosis and treatment data utilization rate and the judgment accuracy are improved, the high efficiency and accuracy requirements of clinical diagnosis and treatment are met, and meanwhile, the method is adaptive to a medical teaching scene.

The invention discloses a network clue data change analysis method and system based on a knowledge graph, particularly relates to the technical field of data processing, and is used for solving the problem that a clue coordinate mapping result does not conform to an actual geographical current situation due to dependence on a static geographical entity knowledge graph in the prior art. The method comprises the following steps: describing and identifying spatial differences by aggregating multi-platform clue positions, and generating a change pre-check instruction when the differences relate to geographic entities in a preset range; performing topological manifold embedding analysis on the text in combination with the initial coordinates to generate a manifold embedding distance, and detecting a conflict rate of a spatial reference system; when the distance or the conflict rate exceeds a threshold value, triggering knowledge graph updating; and finally, re-matching based on the updated map to obtain a target geographic entity and coordinates. Real-time perception of geographic entity change and dynamic map correction are realized, and the accuracy and reliability of natural resource clue space positioning are effectively improved.

The invention relates to the technical field of wireless communication networks, and discloses a method for quickly establishing an indoor positioning system, which comprises the following steps of: among wireless communication nodes without position calibration, dynamically establishing a topological logic relationship based on Pearson's correlation coefficients of entropy change among the nodes by periodically calculating and broadcasting entropy representing signal environment fluctuation; according to the method, the node proximity relation is verified through entropy change synchronism, dependence of traditional positioning on physical coordinate calibration is avoided, and instantaneous self-organization of the topological network is achieved; environmental disturbance is reversed into a natural beacon for verifying a topological relation, so that the system endogenously maintains structural consistency in dynamic change; the terminal location description directly anchors environmental physical nodes, providing intuitive and robust relative location awareness.

The invention discloses a position description generation method and device based on a large model and a storage medium. The method comprises the steps of receiving a recruitment demand text input by a user; inputting the recruitment demand text into a position description generation model to obtain a plurality of candidate position descriptions; scoring each candidate position description according to a preset scoring strategy to obtain a scoring result; and determining the candidate position description with the highest score as the optimal position description according to the scoring result, and outputting the optimal position description. According to the invention, when the position description is generated, the plurality of candidate position descriptions are firstly generated, and then the optimal position description is selected from the plurality of candidate position descriptions, so that the generation quality of the position description is prevented from being uncontrollable, the generation quality of the position description is improved, a recruiter does not need to carry out manual verification and modification, and the efficiency of the recruiter is improved. And the generation efficiency of the position description is improved.

The application discloses a data synchronization method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining position information; determining a first target model according to the difference between the position information and historical position information recorded in a feature database, and version update data of a position feature generation model used by the historical position information, wherein the historical position information comprises a historical position identifier and historical position description information corresponding to the historical position identifier; calling the first target model, and analyzing first field data in first position description information through the first target model to obtain target position feature information corresponding to the first position description information; storing the first position description information and the target position feature information in the feature database corresponding to the first position identifier, so as to provide consumption data to a data consumer through the feature database. The embodiment of the application can improve the synchronization efficiency of recruitment information.

The application provides a region sound field adjustment method based on an inductive directional sound source, a server and a medium, a position description of a target object is acquired by an inductor, and a directional sound source is controlled to output associated first audio content; in a playing process, semantic paragraph division mark points pre-implanted in the audio are parsed, a remaining output duration of a nearest semantic breakpoint is determined according to a current output progress, and second audio content is pre-fetched from a preset audio content library to an output cache in combination with the position description; meanwhile, a position description sequence of the target object is continuously acquired and position fluctuation analysis is performed, a description representing a residence tendency strength degree is generated; when the residence tendency reaches a preset reference and playing reaches any semantic breakpoint, the first audio content tail and the second audio content starting end are continuously spliced into an uninterrupted connection audio stream and output; then the second audio content is continuously cycled as new first audio content until the inductive signal disappears. The application can avoid abrupt sound jumps.

The application discloses a kind of map geographic position information processing method and system, receive the geographic position data to be handled that user uploads, the latitude and longitude information in it is converted into string, and according to string query geographic position cache service, if the query result is not existing position description information, latitude and longitude information is converted into position description information, and string, latitude and longitude information and position description information are written into geographic position cache service, if the query result is existing position description information, return position description information, to this will the position description information cached storage obtained by conversion, need not again through network call third party conversion interface, effectively reduce the request number of times through network with third party map interface, and avoid the problem that conversion efficiency is not high due to the instability that network transmission can exist, to improve map geographic position conversion efficiency, and avoid the waste of network transmission flow.

The invention discloses a training method of a semantic analysis model for recalled positions and resumes. The training method of the semantic analysis model used for recalled positions and resumes comprises the steps that delivery behavior data, user resume data and position description data of a user are obtained, and the delivery behavior data comprise positions where the user delivers the resumes; based on the positions of the resumes delivered by the user, constructing matching samples of the positions of the resumes; based on a pre-configured language analysis model, identifying the delivery behavior data, the user resume data and the position description data corresponding to the matching samples, and screening out the matching samples corresponding to the user resume data which does not meet the position constraint requirements to obtain positive matching samples; in the samples which do not meet the position constraint requirements, negative matching samples are extracted; and performing fine tuning training on a pre-constructed semantic analysis model based on the positive matching sample and the negative matching sample to generate a semantic analysis model of the position and the resume.

The invention provides a controllable multi-mode perception model for a space-time agent and the space-time agent, and relates to the technical field of agents. The multi-modal encoder encodes the multi-modal data and converts the multi-modal data into encoding features; the input projector projects the coding features of the multi-modal data to the same feature space, and the geographic space encoder aligns the coding features of the multi-modal data; the multi-layer Transform neural network carries out reasoning according to the aligned coding characteristics; the output projector converts the reasoning result into instruction features suitable for being used by the multi-mode generator; and the multi-modal generator executes different commands according to the instruction characteristics and outputs data of multiple modals. The problem of position description identification deviation is effectively solved, spatio-temporal data modes such as moving tracks and vector maps are supported, the time sequence prediction accuracy is improved, the sensing accuracy and reliability of geographic space tasks are improved, and multiple spatio-temporal data modes are supported.

The invention relates to the technical field of natural language processing, and discloses a recruitment information and resume matching method, which comprises the following steps: firstly, converting a position description text and a resume text into vector sequences, respectively inputting the vector sequences into a recurrent neural network for encoding, aggregating position side codes into position vectors, and extracting a hidden state sequence from a resume side; segmenting the hidden state sequence according to window parameters, and obtaining collapse measurement in each segment through state derivative calculation and matrix operation; generating a time weight according to collapse measurement and adjustment parameters, and performing weighted aggregation on the hidden state sequence to obtain a resume vector; and finally, calculating the similarity between the resume vector and the position vector to obtain a matching score, comparing the matching score with a threshold parameter, and outputting a resume matching or resume mismatching result to realize inhibition and high-precision semantic matching of the resistant resume.

The invention provides a resume processing method and system based on multi-modal semantic matching, and the method comprises the steps: obtaining a resume document of a candidate, and processing the resume document to obtain a candidate portrait; semantic analysis is carried out on the position description of the target position, and a position portrait is constructed; and performing semantic matching on the candidate portrait and the post portrait based on a double-tower deep neural network model to obtain a person-post matching degree score of the resume document and the target post, and generating a candidate recommendation report based on a large language model. According to the method, the resume screening efficiency is improved, and meanwhile, the resume analysis accuracy is improved, so that the man-post matching is more accurate.

Embodiments of the described technologies receive, via a user interface, an input associated with a first user of a user connection network. The input identifies first position data related to a position capable of being filled by a hiring of a person. In response to validating the first position data, second position data different from the first position data is extracted from the user connection network, based on the first position data. A first prompt is formulated based on the first position data and the second position data. The first prompt is sent to a generative language model. A first piece of writing is received from the generative language model. The first piece of writing includes a position description output by the generative language model based on the first prompt. The position description is sent to the user interface in response to the input.

The invention discloses an NLP-based recruitment behavior skill data encryption protection system and method, and relates to the technical field of data encryption protection, and the system comprises a skill extraction module which is used for carrying out the behavior skill extraction of obtained position description data through a natural language processing model, and obtaining a behavior skill index; the skill quantification module is used for performing skill effect quantity analysis on the acquired behavior data to obtain a behavior skill quantification benchmark; the encryption protection module is used for performing encryption processing on the obtained recruitment data to obtain behavior encryption data; and the skill matching module is used for carrying out behavior skill encryption analysis on the behavior encryption data to obtain a skill radar map, and carrying out matching analysis on the skill radar map and the behavior skill quantification reference to obtain a skill matching degree. According to the method, the quality and efficiency of enterprise recruitment decision making are remarkably improved, and the bottleneck of subjectivity and low efficiency of a traditional manual definition post standard is broken through.

This invention discloses a method and electronic device for generating four-dimensional scenes, relating to the field of 4D technology. In this method, a three-dimensional representation of a target static object, an initial three-dimensional representation of a target dynamic object, and a target three-dimensional deformation amount are generated based on text descriptions of static and dynamic objects, respectively. Path points are inserted according to the spatial position descriptions of the target dynamic object at any two time points, and the three-dimensional representations of the inserted path points are obtained. A four-dimensional scene is generated based on the combined three-dimensional representations at multiple time points. Furthermore, by splitting the scene into static and dynamic objects for separate processing and then combining the three-dimensional representations, the quality of four-dimensional scene generation is improved, and the difficulty of scene generation is reduced. In addition, only the three-dimensional representation at the first time point and the three-dimensional deformation amount at the second time point need to be calculated; the three-dimensional representations of other path points are determined by the initial three-dimensional representation and the target three-dimensional deformation amount. Therefore, the computational load is reduced, and the efficiency of obtaining multiple three-dimensional representations of dynamic objects is improved.

The application relates to an object evaluation method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises: obtaining professional constraint data of a position description text and multi-modal data of a target object; performing feature extraction on the multi-modal data according to a multi-modal large model to obtain multi-modal features, and performing feature fusion on the multi-modal features to generate comprehensive feature data of the target object; performing inference analysis on a semantic matching degree relationship between the comprehensive feature data and the professional constraint data based on an inference large model and a preset evaluation strategy to generate a first position matching degree; and determining a competency evaluation result of the target object according to the first position matching degree. The method can improve the accuracy of competency evaluation.

The invention relates to an object searching method and device, equipment and a storage medium, and the method comprises the steps: responding to a target searching instruction for a target object, analyzing the target searching instruction, and determining target semantic information corresponding to the target object; under the condition that the preset space model does not include the preset object node corresponding to the target object, determining position probability distribution information of the target object in the current space based on the preset space model and the target semantic information; determining a target scanning path corresponding to the target object based on the position probability distribution information; searching a target object based on the target scanning path, and determining target position information of the target object; and outputting position description information corresponding to the target object based on the target position information. The possible position distribution of the target object is predicted according to the semantic information, the optimal scanning path is generated for searching, the corresponding spatial relation description is output after the target object is found, and the object searching efficiency and the user interaction experience are improved.

The application discloses a method for adaptive matching of engineering file data based on character similarity, comprising the following steps: obtaining a target specification table name in an original process; extracting a to-be-compared string in the target specification table name; and matching a result table with the highest similarity to the to-be-compared string by using a preset algorithm. When an electronic signature is submitted, a signature omission item is searched, a process option string of the signature omission item is obtained, a result table with the highest similarity to the process option string is matched by using the preset algorithm, and a signature position description is arranged in the result table. In the preparation of construction engineering files, process data is generated and signature and process omissions are adapted by using the analysis of data in a specific field and the method for similarity matching.

The embodiment of the invention provides an interaction task processing method, a model training method and platform and computing equipment, and the method comprises the steps: responding to a target interaction task, and obtaining interaction data of a front-end interface of the target interaction task; a target action decision corresponding to the interaction data is determined through the target planning model, and the target action decision comprises target element position description of the to-be-executed action, a target action type and target action parameters; determining a target element position corresponding to the target element position description through a target positioning model based on the target element position description of the to-be-executed action and the visual data of the front-end interface; and generating a to-be-executed action under the current time window based on the target element position, the target action type and the target action parameter. According to the double-model decoupling architecture, a complex decision process is decomposed into two cooperative sub-tasks of high-level semantic planning and bottom-layer visual grounding, so that the model focuses on the field where the model is skilled, and the specialized capability and generalization performance of the model are improved.