70results about How to "Improve personalization" patented technology

This invention discloses a weight management system, method, device, and storage medium for overweight and obese individuals based on standardized operating procedures, belonging to the field of health management technology. It includes an information collection and standardized input module; a health risk analysis and assessment module; a personalized intervention plan configuration module; a dynamic intervention execution and tracking management module; and an effect evaluation and optimization feedback module. The system of this invention achieves standardized collection of multi-dimensional basic data of users, health risk assessment and classification, standardized health risk assessment and classification, and effect evaluation and optimization feedback by constructing a collaborative operating system of the information collection and standardized input module, the health risk analysis and assessment module, the personalized intervention plan configuration module, the dynamic intervention execution and tracking management module, and the effect evaluation and optimization feedback module. Simultaneously, it identifies abnormal behaviors and generates early warning information through early warning rules, and uses the effect evaluation and optimization feedback module to comprehensively evaluate and dynamically optimize the intervention plan execution process.

The application discloses a kind of federal positioning method based on individualized patch and dynamic fusion, belong to indoor positioning and federal learning technical field.When applying federal learning to indoor Wi-Fi fingerprint positioning of Internet of Things, global model is often difficult to capture local data feature preference, local training cost is high and poor adaptability alone, the present application is by setting individualized patch and dynamic parameter fusion mechanism, reduce the influence of above-mentioned problem on positioning performance.The algorithm configures private position decoder as individualized patch for each local client, extracts local data features in training and uploads to server, server calculates the similarity of data distribution between clients to generate combination weight, for local clients with similar data distribution, individualized patch can selectively linearly combine with the decoder of the local model with higher relevance, avoid local model overfitting, improve the positioning performance of global model in local scene.

The invention provides a personalized step speed self-adaptive control method applied to an ankle-assisted exoskeleton, and the method comprises the steps: installing a feedback device at the bottom of a heel and the bottom of a toe, and detecting the acting force of the sole in real time so as to divide four action stages of single-foot walking; and constructing a conversion relation between a first function characterization action duration and the step speed, and constructing a second function characterization step speed and the driving torque. And during walking, the real-time action duration in the middle stage of the supporting action is obtained, the real-time step speed is calculated through the first function, the real-time driving torque is obtained through the second function, and the ankle exoskeleton assistance is controlled according to the real-time driving torque. The driving torque is dynamically adjusted by sensing the gait and the step speed in real time, so that the assisting power fits the walking state of an individual, and the self-adaptive capacity and the assisting power accuracy under the change of the step speed are remarkably improved; and meanwhile, the method has good individuation and generalization capabilities.

The application provides a product retail advertisement pushing method and system, electronic equipment and storage medium, and relates to the technical field of intelligent advertisement pushing and consumer behavior analysis. The method comprises the following steps: collecting user biological data of a target product retail display area; performing multi-dimensional processing on the user biological data to generate a feature change mode set, a pupil response sequence and a physiological response curve; performing multi-modal data fusion on the feature change mode set, the pupil response sequence and the physiological response curve, combining an emotional computing algorithm to determine the emotional feedback intensity value of the user to the current advertisement content; generating a personalized advertisement scheme matched with the emotional state of the user according to the emotional feedback intensity value and combining a generative adversarial network; matching the personalized advertisement scheme with product characteristic data browsed by the user to generate specific advertisement content, and pushing the specific advertisement content to the user. The application improves the real-time accuracy and user emotional matching degree of advertisement pushing.

This invention provides a Pinyin error correction method that simplifies the noise channel model used in most current Pinyin error correction algorithms by employing a real-time frequency counting method, effectively improving the efficiency and lightweighting of the algorithm. Furthermore, this invention uses a direct character substitution method instead of the traditional edit distance calculation method, avoiding frequent edit distance calculations in the Pinyin error correction algorithm. The dictionary database itself is localized to the individual user, resulting in a highly personalized, targeted, and small-scale dictionary that provides accurate candidate words. Compared to existing algorithms, this invention offers significant improvements in error checking rate, candidate word accuracy, and execution efficiency.

The invention provides a commodity recommendation method and system based on consumer behaviors, and the method comprises the steps: firstly extracting explicit behavior features and implicit preference features of consumers in different context scenes, and determining node preference indexes of consumption behaviors in different context scenes; generating a short-term intention chain and a long-term preference chain of the consumer according to the node preference index, and further constructing an element feature map reflecting consumer consumption behavior preference according to a mapping association relationship between the short-term intention chain and the long-term preference chain; extracting the preference weight of each preference node in the element characteristic spectrum under different context scenes; and detecting the matching degree between the element characteristic spectrum and the current consumption behavior characteristics, and when the matching degree is lower than a preset threshold value, performing preference updating on the element characteristic spectrum based on the preference weight corresponding to each preference node and the current consumption behavior characteristics, thereby generating a commodity recommendation list. By adopting the scheme of the invention, personalized commodity recommendation based on different context scenes can be realized in an internet e-commerce platform.

The application discloses a kind of based on the image recognition model training method of multi-agency combination of grouping weighting, which can be applied to medical image recognition and other collaborative learning tasks with multiple participants.In order to alleviate the influence brought by medical image data distribution inconsistency, first, the application groups clients according to the similarity of client models and intra-group global models;For intra-group clients, the application calculates a personalized model for each client according to the similarity between the local models of intra-group clients.Compared with the prior art, the application can reduce the client computing overhead while providing better image recognition models for participants.

This application provides a road rendering method, apparatus, device, and storage medium. The method includes: acquiring road auxiliary points corresponding to each landmark point in a navigation route and the texture coordinates of the road auxiliary points; determining the texture rendering value of each road surface point based on the texture coordinates of the road auxiliary points and a predefined road surface transparency transformation relationship; and rendering the corresponding virtual navigation road based on the texture rendering value of each road surface point. The embodiments of this application can achieve single rendering of multiple effects within a virtual navigation road, eliminating the need for multiple rendering operations to overlay various road effects, simplifying the tedious steps of road rendering, and improving the efficiency and personalization of road rendering.

The invention relates to a physical environment regulation and control method and system for a shared therapy and healing space, belongs to the technical field of intelligent environment regulation and control and healthy therapy and healing, and solves the problems that therapy and healing requirements cannot be accurately responded and equipment is lack of cooperation. Comprising the steps that physiological data, emotion data and behavior data of a user are collected through a multi-mode sensor, feature extraction and data fusion are carried out, and a user state vector is generated; based on the user state vector and a healing demand selected by the user, generating a multi-sensory healing scheme comprising a sound therapy equipment control parameter, a light therapy equipment control parameter, a thermal therapy equipment control parameter and a humidity regulation and control equipment control parameter through an intelligent matching algorithm; a control instruction is sent to a corresponding execution device according to the multi-sense-organ healing scheme; and continuously collecting physiological data, emotion data and behavior data of the user, and dynamically optimizing the control parameters of the execution equipment based on the continuously collected physiological data, emotion data and behavior data of the user.

The present application relates to the technical field of big data, and particularly to a health insurance automatic pricing system based on big data, in which, through a multi-source heterogeneous data acquisition module, sleep depth, health behavior and living environment data are integrated in real time to solve the problem of single data dimension of traditional health insurance; a dynamic health portrait construction module adopts dynamic weight distribution and space-time alignment technology to unify multi-granularity data into a weighted daily dimension vector, eliminating the difference in data value evaluation and the fragmentation of time granularity; based on this, real-time dynamic health risk scores are generated and health behavior entropy values are extracted to quantify behavior stability, breaking through the limitations of static risk assessment; a risk transmission prediction module identifies and maps the disease development rate through the entropy value abnormal duration, constructs a personalized risk evolution path diagram, and finally converts the risk escalation node into an actuarial factor to dynamically generate a ladder-type premium gradient, realizing the accurate matching of premium level and user real-time risk trajectory.

The application discloses a kind of humanoid robot perception interaction method and related equipment, belong to humanoid robot technical field, the method includes: the image data of interactive object is acquired in real time;Wherein, image data includes face near-infrared image and behavior visible light image;Based on image data, extract target data features and identify, determine the physiological monitoring result and behavior intention result of interactive object;Dual-mode feature fusion is carried out to physiological monitoring result and behavior intention result, and fusion perception result is obtained;Combining fusion perception result and preset humanoid robot joint constraint parameter, obtain humanoid robot motion control instruction and execute.This application can realize dual-mode perception by combining physiological monitoring based on near-infrared and behavior recognition based on visible light vision, and introduce joint constraint parameter to improve the scene adaptation of robot motion control, so as to effectively improve the safety and individualization of human-computer interaction of humanoid robot in industrial household scene.

This invention discloses an intelligent management system for SMS marketing, relating to the field of communication technology. It includes: a vector construction module, used to construct the current state vector of each target user based on user reply information of each target user to the target marketing SMS in response to a triggered update mechanism for historical vector clusters; the current state vector represents the real-time interest changes of target users in the target marketing SMS; a cluster update module, used to update the historical vector clusters based on the current state vectors of each target user to obtain the current vector cluster; a value determination module, used to determine the current state action value of the current vector cluster based on the vector samples included in the current vector cluster; and a strategy determination module, used to determine the current recommendation strategy for each target user in the current vector cluster based on the current state action value of the current vector cluster using a Q-learning algorithm. This invention can improve the accuracy and personalization of SMS marketing.

The application provides an athlete muscle strength training and rehabilitation system based on electric stimulation, and relates to the technical field of medical rehabilitation.The application synchronously collects electromyography and kinematic signals, extracts muscle coordination index and joint net output torque double-dimension features in real time, identifies three movement performance states of coordination dominant type, strength dominant type and mixed type, and dynamically generates differentiated electric stimulation parameter sets according to the three movement performance states, so that the precise typing and targeted intervention of the fatigue state of athletes are realized, the muscle coordination efficiency and strength output are synchronously optimized in training, compensatory injury is effectively prevented and functional recovery is accelerated in rehabilitation, and the individualization, safety and overall efficiency of training and rehabilitation are improved.

This invention relates to the technical field of digital advertising and discloses a digital advertising method and system based on virtual communities. The invention uses merchant virtual units to display digital models of target products, while customer virtual units can observe products from multiple dimensions. Based on the transaction records of the target products, offline observation information is obtained from the customer virtual units, and the product digital model is reconstructed accordingly, generating a set of measured reference models. By parallel matching of the product digital models of different merchants within the virtual community and the measured reference model set, a personalized temporary community scene is generated for the customer virtual units. Based on the customer's registration information, logistics delivery prediction is performed on the product information in the temporary community, generating a personalized product prediction model. This helps customers make more accurate shopping decisions. This method enhances the personalization and accuracy of advertising, improves the customer shopping experience and advertising effectiveness, and solves the problem of low advertising effectiveness in existing technologies.

This invention relates to the field of medical data analysis technology, specifically to a patient response assessment method and system based on chemotherapy data analysis. The method includes the following steps: accumulating drug administration time and dose to form a cumulative exposure sequence; matching biomarker time and concentration to form response pairs; extracting adjacent differences and dividing assessment intervals according to correlation; calculating the interval change ratio to obtain an assessment coefficient sequence; and combining the drug administration time window and the response time window to calculate the chemotherapy response result. In this invention, by time matching and difference analysis of cumulative exposure and tumor marker concentration data, the relationship between exposure changes and indicator changes is characterized, reducing fixed threshold assessment bias, improving the utilization of time series data, enhancing the accuracy and individual adaptability of assessment results, improving the reliability of chemotherapy response assessment, and stabilizing assessment consistency.

The present application relates to the technical field of medical data mining, in particular to a method and system for predicting recurrence of atrial fibrillation based on multi-modal data. The method divides the level period based on the historical symptom level duration; determines the atrial fibrillation risk index by combining the symptom level change and the keyword addition similarity in the level period; determines the normal period and filters out the problem period by the periodic fluctuation of the dynamic electrocardiogram; analyzes the problem performance degree according to the difference between the problem period and the normal period, the problem distribution of the level period and the risk index, and determines the prediction attention degree by referring to the frequent confusion of the previous problem distribution, and adjusts the prediction weight based on the prediction attention degree to make prediction and warning. The present application integrates multi-source data and dynamically quantifies the risk, adjusts the prediction weight of the high-risk period, effectively improves the individualization and accuracy of the prediction of recurrence of atrial fibrillation, and provides more reliable support for recurrence warning.

The present invention discloses an intelligent pillow system with multi-zone airbag adjustment. The present invention precisely senses the head position and sleeping posture through a high-density fiber pressure sensor array, and based on this information, drives an innovative "single air pump + multiple electronic control valves" and "integral airbag + partitioned air conduction" structure, achieving remarkable beneficial effects: First, by immediately triggering the overall descent protection when the head is tilted up / down, it can actively prevent hyperextension and hyperflexion of the cervical vertebra, effectively reducing the risk of neck pain and discomfort. Second, the sensor without protrusions embedded and the PWM smooth adjustment of the airbag ensure a comfortable experience; the system adopts a single power unit and logical partition design, greatly simplifying the air circuit, significantly reducing the hardware cost, power consumption, and the failure rate of easy leakage in multi-chambers, and improving the reliability. In addition, the interaction mode combining user-set instructions and automatic recognition takes into account both convenience and personalization, and the self-learning module enables the pillow to continuously adapt to the user's habits, achieving long-term precise personalized support.

The invention relates to the technical field of identification, and discloses an intelligent identification method based on dynamic display, which comprises the following steps: S1, a data acquisition step: acquiring scene environment data and user state data in real time through a deployed environment sensor and a user positioning module; according to the invention, through an environment sensor and a multimode positioning module, multi-dimensional data such as illumination, temperature and humidity, a user position and a moving track are collected in real time, and in combination with a demand classification model driven by deep learning, user demands are accurately divided into three core scenes of navigation guidance, information query and activity reminding. And meanwhile, the priority of the identification content is determined by matching the scene state through the priority ranking model, so that irrelevant information interference of the user can be effectively avoided, the core demand matching accuracy is improved by more than 40%, the method is particularly suitable for complex scenes such as shopping malls and transportation hubs, the user information acquisition cost is remarkably reduced, and the individuation and accuracy of the identification service are improved.

The invention discloses a method for automatically exporting an AM model, and relates to the technical field of ships, and the method comprises the following steps: S1, file configuration: configuring plug-in files for AM software, including configuring an objects file, configuring a Function file and configuring a Forms file; s2, plug-in loading: loading a plug-in in AM software to realize a function of checking a ship model by using Navisworks software; and S3, a use process: exporting the RVM format file through AM software, and realizing conversion and viewing of a model file by using an additional model function of Navisworks software. According to the method for automatically exporting the AM model, secondary development is carried out on the AM model through the PML, so that the ship model can be smoothly checked by using Navisworks software under the condition that an AM environment is not installed, the model checking efficiency is remarkably improved, the operation is simple and easy to learn, the model can be quickly searched, the distance can be quickly measured, interference check can be quickly carried out, and meanwhile, viewpoints can be conveniently marked during model review.

This invention relates to a method and system for optimizing meal replacement intake time based on circadian rhythms, belonging to the field of nutrition management technology. It includes: collecting basic user information, sleep time, and wake-up time; determining a sleep-wake cycle based on sleep and wake-up times; retrieving user age and gender based on the basic information and dividing the sleep-wake cycle into physiological time periods; retrieving health goals and physical fitness parameters based on the basic information to determine the types of meal replacement foods; determining the initial time for meal replacement intake based on physiological time periods and selecting recommended food types; determining a type-based intake adjustment value based on the recommended food types; adjusting the initial time for meal replacement intake based on the type-based intake adjustment value to obtain a recommended meal replacement intake time and outputting it. This invention improves the efficiency of users in achieving their health goals during nutrition management.

The invention provides an interactive broadcasting method and system, and relates to the technical field of intelligent vehicle man-machine interaction. The method comprises the following steps: in response to an operation instruction of a user on a graphical interface, obtaining target material data and a broadcast permission of a target user, and uploading the data and the permission to a cloud platform; generating a callable target data packet of different asset combinations of images, sound rays and permissions in the voice broadcast application based on a driving signal, a pixel-level feature and a broadcast permission determined by the sound ray feature and the audio feature in the target material data; according to the real-time application scene of the current vehicle and the broadcasting authority corresponding to the real-time target user, the corresponding voice broadcasting image and voice broadcasting data are called from the target data packet for synchronous broadcasting, the image and sound ray selection of voice broadcasting are enriched, the computing power and memory occupation of the vehicle are reduced, and the emotional and personalized experience of human-computer interaction is improved.

This invention discloses a closed-loop control method and system for prevention, intervention, and risk feedback in cerebral hemorrhage, relating to the field of cerebral hemorrhage risk intervention technology. It includes the following steps: S1, real-time acquisition of multi-source cerebrovascular monitoring data, data preprocessing, and identification of high-pressure peak events; S2, quantification of key dynamic characteristics of each peak based on the multi-source cerebrovascular monitoring data corresponding to the high-pressure peak events, and assessment of high-pressure peak abnormalities; S3, fusion of multi-source cerebrovascular monitoring data and high-pressure peak abnormality assessment results, quantification of the risk probability of each peak event, identification of physiological fluctuations and risk events, and risk warning and response; S4, real-time monitoring of multi-source cerebrovascular monitoring data after risk warning, assessment of warning accuracy, identification of missed and false alarms, and threshold optimization. This solves the technical problems of existing technologies, such as difficulty in distinguishing between physiological and pathological peaks, susceptibility to misjudgments, and impact on the timeliness and individualized management of risk intervention.

The invention relates to a self-adaptive nutrition planning and recommendation method, system and device based on user characteristics and a medium. The method comprises the following steps: acquiring basic data, dynamic state data and a user preference vector of a user; based on the basic data and the dynamic state data, analyzing a nutrition target of the user in each meal, and generating a meal nutrition target matrix; on the basis of the meal nutrition target matrix, a preset food specification database and the user preference vector, generating a standard meal combination for each meal in the meal nutrition target matrix, and obtaining a standard meal recommendation report; the standard meal recommendation report is used for indicating the standard meal of the user in each meal; the food specification database comprises the calorie supply value of each standard meal and the numerical value of each nutrition type included in each standard meal. By adopting the method, the scientificity and individuation of nutrition planning and recommendation can be improved.

The invention discloses an engineering innovation practical ability cultivation method and system based on a domestic EDA platform and AI-PBL, and belongs to the field of engineering education and intelligent teaching. A domestic EDA tool is taken as an engineering practice carrier, real engineering problems and subject competition tasks are taken as driving, and a stepped progressive project system driven by an ability matrix is constructed, so that the engineering innovation practical ability is cultivated. And systematic fusion of professional knowledge learning, engineering practice training and innovative design capability cultivation is realized. In the implementation process, an artificial intelligence technology is introduced to carry out intelligent support, design assistance, debugging analysis and engineering practice behavior data processing on the learning process, and the engineering practice ability, team cooperation ability and innovative design ability of students are comprehensively evaluated based on a multi-dimensional ability evaluation index system. According to the method, the problems of course content splitting, discontinuous ability cultivation chain and single evaluation mode in existing EDA practical teaching are solved, and the engineering practice task implementation efficiency and ability evaluation consistency are improved.

The invention discloses a multi-mode perception type voice accompanying module system supporting task and learning synchronization, and the system comprises the steps: converting at least one non-voice input and voice input into a structured semantic event in a unified manner; operation is carried out under the combined action of security abstraction, context construction, priority decision, learning synchronization and situation guidance, multi-model arrangement and personality / memory cooperation, so that the multi-modal understanding ability and learning target achievement efficiency are remarkably improved on the premise of privacy compliance and low time delay, and individuation and long-term growth ability are enhanced; and consistent experience in a cross-device scene is ensured.

This application provides a method and system for joint query analysis of multimodal data supporting natural language interaction, relating to the technical field of data query analysis. This application collects historical operation data, audio data, and text data from an intelligent search platform supporting natural language interaction; then models the historical operation data to obtain a user preference vector; processes the audio data through speech recognition and text standardization to obtain standard query data, and integrates it with the text data to form contextual information; then uses an attention-based algorithm for semantic understanding and intent recognition to obtain an intent feature vector, and fuses it with the user preference vector to obtain a classification result; finally, based on this result, a collaborative filtering algorithm is used to query a pre-set multimodal database to obtain joint query results. This enables personalized and accurate querying of multimodal data under natural language interaction, ensuring that the results match user intent and long-term preferences.

The invention discloses a visual perception large model distributed training method based on a general calculation fusion network, and relates to the technical field of information and communication, and the method comprises the steps: an edge server initializes and broadcasts a pre-trained general large model; receiving general large model parameters to form a local general model; the edge device continuously receives visual perception data samples, and online learning is carried out through incremental training and a sliding window mechanism; performing clustering analysis on local sample data, and classifying samples into personalized model processing samples and global model processing samples; respectively inputting the classification samples into the personalized model and the global model, and weighting and aggregating output features through a gating network; the edge device uploads the updated global model parameters to a server for aggregation, and a new global model is generated; and the edge equipment periodically receives the latest global model parameters broadcasted by the server, and iterative training is carried out until the model converges. According to the invention, high-efficiency, high-precision and high-reliability distributed model training and visual perception performance are realized.

This invention relates to the interdisciplinary field of artificial intelligence and intelligent education, and provides a method for adaptively constructing virtual simulation training scenarios based on generative AI. It aims to address the problems of insufficient personalization and poor dynamic adaptability in existing training technologies due to their reliance on static pre-built scenario libraries. The method dynamically constructs and updates the cognitive state graph of trainees to achieve accurate state perception, and generates optimal scenario strategies through multi-agent collaborative inference and multi-objective game evaluation based on this. The optimal scenario strategies are then integrated with a static domain knowledge graph to generate structured instructions that drive a multimodal generative AI model to synthesize specific training scenario content. Finally, the virtual simulation training engine is driven to update the training scenario and collect feedback data, thereby constructing a closed loop of "perception-decision-generation-execution." This allows for the dynamic generation of personalized training content based on the trainees' real-time state, significantly improving the timeliness, adaptability, relevance, and personalization of the training.

The present application relates to the technical field of data processing, and provides a menu generation method and device based on multiple agents, a terminal, a storage medium and a computer program product, the method comprising: obtaining input information of a user, analyzing the input information, and determining demand information and dining scene information of the user; based on the recipe demand information and the dining scene information, determining supplementary information by using RAG technology; generating recipe information according to the recipe demand information, the dining scene information and the supplementary information; determining nutritional information of the recipe information; and optimizing the recipe information according to the nutritional demand information and the nutritional information to obtain target recipe information required as a menu generated based on multiple agents. The present application can combine user demand of multiple modal inputs and multiple scene constraints, generate an executable creative menu scheme that meets individual tastes and takes into account nutritional balance through a feedback mechanism and continuous learning ability, and improve user experience.