20 results about "Time sensitivity" patented technology

The invention belongs to the field of large-model retrieval, and discloses a large-model knowledge retrieval enhancement method and a related system for the field of power dispatching. A dynamic weighting mechanism of time sensitivity and credibility factors is introduced on the basis of a traditional knowledge retrieval enhancement RAG retrieval framework, so that a retrieval result not only depends on semantic similarity, but also depends on credibility factors; and the timeliness and the source reliability of the information can be comprehensively considered. In traditional knowledge retrieval enhancement (RAG), document sorting often mainly depends on semantic similarity, and differentiation processing of content release time and credibility is lacked, so that the problem that information is outdated although the information is similar or the content source is unreliable easily occurs in a scheduling scene. According to the method, after keyword retrieval and vector retrieval are combined to obtain initial semantic related data, the timeliness weight is calculated through the interval between the timestamp of the query content and the current time, and it is ensured that high-timeliness knowledge such as real-time trend data and an operation mode is preferentially presented in sorting.

The invention discloses a risk assessment stabilization method for senile chronic disease detection, and relates to the technical field of medical data analysis, and the method comprises the steps: carrying out federal dynamic time warping processing on encrypted patient data streams of a plurality of medical institutions, generating a joint feature space mapping matrix, and obtaining a standard feature tensor through homomorphic encryption; inputting the standard feature tensor into a dynamic medical knowledge graph construction module, and generating a knowledge graph embedding matrix through a space-time sensitivity enhanced cross-modal attention mechanism; and fusing the standard feature tensor and the knowledge graph embedding matrix, constructing a dynamic hypergraph structure, executing dual-channel hypergraph convolution calculation, and outputting a patient-knowledge joint embedding matrix. According to the method, in the construction of the dynamic knowledge graph, the time sensitivity weight of the entity relationship is quantified through the exponential decay function, and the time-space enhanced embedded matrix is generated in combination with the cross-modal attention mechanism, so that the time sequence discrimination of the concurrent disease association strength is improved.

The invention discloses a power grid dispatcher cognitive load dynamic adjustment system and method based on multi-modal biofeedback, and the method comprises the steps: carrying out the real-time quantification of the cognitive load of a power grid dispatcher through fusing the multi-modal biofeedback signals, such as electroencephalogram, eye movement and galvanic skin; and a human-computer interface information presentation mode and a task allocation strategy are dynamically adjusted according to the load level. The system specifically comprises a biological signal acquisition module, a cognitive load calculation engine module, a dynamic adjustment execution module and a closed-loop feedback adjustment module. According to the method, the problems of information overload, poor individual adaptability, high misoperation risk and the like in a traditional power grid dispatching system can be effectively solved, the efficiency and safety of dispatchers in complex fault handling are improved, and the time sensitivity requirement of power control is met.

A memory-enhanced time-sensitive model control method for super-long video understanding is applied to the field of super-long video processing, and comprises the following steps: dynamically sampling an input original video frame sequence to obtain a compressed first frame sequence, the first frame sequence comprising a selected frame obtained by dynamic sampling and a corresponding timestamp; performing time dynamic storage on the first frame sequence according to the sliding window and a first-in-first-out architecture to obtain a frame representation set which is located in a memory buffer area and has time perception and an output exceeding frame representation; based on the sliding window, performing space-time weighted integration on the frame representation set and the exceeding frame representation to generate a comprehensive frame representation; and performing iterative coding according to the comprehensive frame representation and the user query to obtain a response to the user query. According to the method, the calculation burden is reduced, the time sensitivity is improved, the understanding and reasoning of a long-time video are enhanced, the overall performance and accuracy of the model are improved, and the excellent zero sample learning ability is shown.

The application provides a dynamic knowledge retrieval and iterative generation method based on time sensitivity quantification. The system extracts the semantic features, time entities and historical answer update frequency of the user's question, calculates the time-sensitive sensitivity score of 0~1, and calls the large model API to assist calibration for complex or uncertain questions. The dynamic retrieval module assigns weights to the local knowledge base and Internet documents according to the score, constructs a time sequence candidate document set, and prioritizes the latest Internet documents for high sensitivity questions, mainly uses the local knowledge base supplemented by the Internet for medium sensitivity questions, and only uses the local knowledge base for low sensitivity questions. In the review stage, the time sequence and version conflict are detected to form a preliminary review document set, and in the refine stage, the answer is optimized through dynamic time weight fusion, obsolete information filtering and time constraint, and a closed loop is formed through secondary recall. Finally, the answer is generated with the source, timestamp and time-sensitive confidence. The method effectively solves the problems of traditional static question and answer system in time-sensitive adaptation, fixed document weight and time sequence conflict processing difficulty, and improves the accuracy, timeliness and system stability of the answer.

The invention relates to an early warning method for infectious shock and a storage medium. Firstly, the limitation of a single mode is broken through by obtaining multi-mode physiological feature parameters of a user; then, constructing a graph neural network according to the multi-modal physiological feature parameters to obtain spatial correlation features, and obtaining time evolution features based on short-term time sequence features and long-term time sequence features; the method comprises the following steps of: obtaining a time-space feature vector by fusing a space correlation feature and a time evolution feature, outputting a risk probability and a time sensitivity index of the infectious shock of a user through a two-channel mechanism, and predicting the risk of the infectious shock from two dimensions of time and space, so that the risk level of the user can be more accurately predicted, and the risk prediction accuracy is improved. And the timeliness of risk early warning is improved, and timely intervention on the infectious shock of the user is realized, so that the treatment effect is improved.

The present invention provides a method for comparative analysis of Parkinson's disease gait based on time-series characteristic curves. The method comprises acquiring video data containing gait information, wherein the gait information includes gait information of a user before and after taking medication; performing posture recognition on a human body in the video data to extract position information of key body parts; calculating gait features based on the extracted key point information, and constructing a time-series curve based on the changes of these features over time; performing data processing on the extracted gait features using a digital filter; integrating local frequency and amplitude into a distance metric using a frequency-amplitude enhanced dynamic time warping method to quantify differences in gait motion; obtaining a curve difference distance; and calculating a metric based on the cumulative frequency difference and amplitude difference in the optimal warping path. The present invention exhibits significant advantages in terms of accuracy, time sensitivity, robustness, adaptability, comprehensive evaluation capability, quantitative analysis, and clinical application potential.

The application relates to the technical field of personnel position determination, in particular to a key personnel position determination method, device and equipment, wherein the accuracy of historical personnel position determination, a historical average time attenuation factor and an average spatial correlation degree are used to dynamically and adaptively determine the weight corresponding to the perception data type. Based on the determined weight corresponding to the perception data type, the time attenuation factor can improve the time sensitivity of the personnel position determination, and the address confidence degree is fused to finally accurately determine the comprehensive score of the address, and the key personnel position is determined based on the comprehensive score of the address, so that the working efficiency can be improved, the personnel cost can be reduced, and the response speed can be improved.

The invention discloses an air conditioner load prediction method and system based on artificial intelligence, and the method comprises the steps: carrying out the normalization processing of an original load, weather and personnel data flow, and intelligently carrying out the time sensing coding of a normalized feature set based on a current prediction time point and the size of a playback window, a feature snapshot vector and a time delay vector are explicitly extracted, further, a feature value and time delay joint embedding mechanism is adopted, feature value information and time sensitivity are deeply fused, the limitation of simple information superposition is avoided, and on the basis, through context association coding based on cross-feature attention, the feature value information and the time delay information are integrated. And dynamically capturing complex relevance between different characteristics and time lags thereof, and comprehensively constructing context information of the air conditioner load. Thus, the inherent problems of information loss and noise amplification of a traditional data processing method can be effectively solved, and complex air conditioner load influence factors and time sequence dynamics thereof are described more comprehensively, so that high-precision prediction of the air conditioner load is realized.

The present invention belongs to the field of knowledge tracking technology, and discloses a knowledge tracking method and system that supports dynamic context. The method includes: constructing a knowledge tracking model that supports dynamic context length based on the student's question set; and obtaining the final prediction result of the student's answer based on the knowledge tracking model that supports dynamic context length. The present invention maps the student's historical answer sequence into a high-dimensional feature embedding through a question encoder and an answer encoder, and generates a question-level perception vector and a joint question-answer interaction representation. The attention score matrix is ​​dynamically adjusted based on a directional penalty function matrix to enhance the model's adaptability to dynamic changes in context length. The multi-head self-attention mechanism and scaled dot product attention are used to extract time-sensitive and context-dependent knowledge states from student interaction data.

The invention discloses a computing resource allocation method for mobile detection task scheduling in an extreme environment, which comprises the following steps of: calculating a priority coefficient of a task to be scheduled based on the time sensitivity, importance weight and resource occupancy rate of the task; outputting the actual energy efficiency ratio of the calculation core based on the basic energy efficiency of the calculation core and the temperature attenuation coefficient and the radiation attenuation coefficient in the current environment; dynamic load balancing is executed, a two-dimensional scheduling model is constructed based on the priority coefficient and the actual energy efficiency ratio, high-priority tasks are directionally distributed to a calculation core with the high actual energy efficiency ratio, and low-priority tasks are distributed to a calculation core with the low actual energy efficiency ratio; and an energy consumption prediction module is introduced to predict the calculation load in a future set time period, and the working state of the calculation core is adjusted in advance according to a prediction result. The total power consumption of the calculation module can be reduced, the cruising ability of the detector can be prolonged, and collaborative optimization of the task real-time requirement and the system energy consumption can be realized.

The invention relates to the technical field of electric digital data processing, in particular to a question and answer processing method and device based on time sensitivity and a medium. The method comprises the steps of receiving a question input by a user; judging whether the question input by the user comprises a word representing time or not, and if yes, determining the question input by the user as a target question; otherwise, performing intention analysis on the question input by the user, judging whether the question input by the user needs to consider a time factor or not, and if not, determining the question input by the user as a target question; otherwise, acquiring a target time range matched with the question input by the user, and constructing a target question according to the target time range and the question input by the user. The question answering accuracy of the question answering system can be improved.

The invention discloses a power grid emergency material dispatching double-layer optimization method and system under an extreme weather. The method comprises the following steps: firstly, acquiring real-time meteorological monitoring data and vulnerability characteristics of power grid equipment, calculating a dynamic time sensitivity coefficient, and generating a material demand priority parameter in combination with a disaster point load level; secondly, constructing a double-layer material scheduling model, and introducing constraints to realize degraded supply under resource limitation by taking the maximization of a key load recovery index as a target at the upper layer; the lower layer achieves dynamic risk avoiding and time-space synchronization of people and goods through constraints with the purpose of minimizing the comprehensive scheduling cost; and finally, a double-layer iteration solving strategy based on interlayer interaction is adopted, an upper-layer topology construction mask is utilized to execute dimension reduction locking and dynamic truncation on lower-layer flow, and real-time weather and road condition data are accessed in the solving process to execute online rolling correction. According to the method, the problem that traditional scheduling lacks environment perception and supply and demand flexibility is solved, and the real-time adaptability, decision robustness and cooperation efficiency of power grid emergency repair under the extreme weather are improved.

The invention relates to the technical field of quality detection, in particular to a quick-frozen strawberry after-unfreezing quality detection method based on visual identification, which comprises the following steps: collecting sequence images of quick-frozen strawberries in an unfreezing process to obtain an image sequence including time dimension information; condensate water interference elimination processing is carried out on the image sequence, a corrected image sequence is obtained, juice exudation dynamic characteristics are extracted from the corrected image sequence, and the juice exudation dynamic characteristics comprise exudation area change rate and exudation area texture evolution characteristics; and inputting the juice exudation dynamic characteristics into a quality evaluation model, and outputting a quality prediction result when the thawing is completed. According to the method, the high-consistency image data flow covering unfreezing all the time is constructed, and compared with a traditional quality evaluation method based on a static image, the whole-process dynamic behavior of juice exudation and surface change can be captured, and the time sensitivity and the reflection authenticity of unfreezing quality detection are improved.

The present invention belongs to the technical field of distributed system management and maintenance, and discloses a system log detection method, device and storage medium based on a pulse neural network. The present invention uses a pre-trained T5 model to implement semantic embedding of each log in a log sequence, thereby obtaining a word embedding sequence with contextual semantic association; then, pulse feature encoding is performed on the pre-trained T5 model to capture the semantic information and time dynamics in the embedded words, thereby achieving effective encapsulation of log semantics and time information, and obtaining a spatiotemporal pulse sequence; finally, the pulse sequence is input into the trained pulse neural network model to obtain a log anomaly detection result; thus, the present invention can effectively encapsulate the time and semantic information of the log, and simultaneously use the T5 model and the pulse neural network model to capture the association relationship between logs, thereby achieving anomaly detection of logs based on log association and time sensitivity; thus, the effectiveness and accuracy of log anomaly detection are improved.

The invention discloses a data recommendation method and apparatus, and an electronic device. The method comprises the steps of obtaining a target recommendation model; the target recommendation model is obtained by dynamically adjusting popularity adjustment parameters in a user preference module in the recommendation model and time sensitivity parameters in a popularity module in the recommendation model by utilizing reinforcement learning; the user preference module is used for calculating preference characteristics of a user, and the popularity module is used for capturing a change trend of article popularity along with time; and according to the target recommendation model, generating a prediction score between the to-be-predicted user and the to-be-predicted article, and based on the prediction score, displaying a recommendation result for the prediction user. Through the technical scheme provided by the embodiment of the invention, the displayed recommendation result better conforms to user preferences and can fully reflect the popularity trend of the articles, so that the satisfaction and experience of the user can be improved.

The invention discloses a session recommendation method based on a time interval perception graph neural network and collaborative enhancement, and aims to improve the modeling capability of a recommendation system for time sensitivity, structure discrimination and noise robustness. The method comprises the following four key steps: firstly, constructing a session graph of multiple types of edges containing time interval information based on a session sequence; secondly, fusing time and semantic context features by utilizing a gating attention mechanism, and generating high-quality project embedding; thirdly, by splicing relative time embedding and project embedding, combining GRU and an attention mechanism, extracting dynamic and general preferences of the user, and generating session representation; and finally, a collaborative enhancement mechanism is introduced, information is extracted from similar sessions, and current session representation is enhanced. According to the method, the accuracy and stability of recommendation are improved while the structural modeling precision is ensured.

The invention relates to a knowledge pushing method and device in a multi-round dialogue scene, equipment and a medium, and belongs to the technical field of knowledge pushing. The method and the device are used for solving the problem of how to improve pushing opportunity and pushing content accuracy. The method comprises the steps of obtaining multi-round dialogue content of a user and a system; inputting the multi-round dialogue content into a large language model to obtain dialogue state information; performing knowledge keyword trigger identification on the user query of the target round and the system reply of the target round, evaluating the pushing emergency degree based on the context confusion, the time sensitivity factor and the emotion emergency degree factor, and determining a pushing strategy based on the pushing emergency degree; determining active and passive push modes based on the trigger identifier, and determining a push content type based on the knowledge keyword; searching candidate knowledge in a knowledge base corresponding to the push content type according to active and passive push modes in combination with the dialogue state information; and determining final push content based on the multi-round dialogue content and the candidate knowledge through a large language model.

The invention discloses a user behavior analysis system and method for digital commemorative inspection industry big data, and relates to the technical field of data processing, and the method comprises the steps: collecting user operation log data, business approval process data, policy sensitive period data and relative relationship data in an organization personnel system; calculating a post basic value beta p; generating a dynamic behavior baseline Bt; constructing a relation graph containing the following associated edges; outputting a five-dimensional quantitative report and generating a visual space-time evidence chain; and when P is greater than or equal to theta and theta belongs to [0.6, 0.9], triggering a risk alarm. According to the method, the effect of real-time calibration of post characteristics * time sensitivity * group variance is achieved by establishing a dynamic baseline triple linkage mechanism; through a third-order penetration algorithm, the limitation of traditional single-layer analysis is broken through, and the effect of cross-department hidden association mining is achieved; through a five-dimensional quantitative model, the policy sensitive period is taken as an independent risk dimension to be incorporated into an evaluation system.

The disclosed smart electronic faucet system controls the flow of water being dispensed using voice commands. The disclosed smart electronic faucet system is a hybrid system that uses the computing resources available locally within the faucet itself for time-sensitive commands and the computing resources of a network connected server system or computing device for commands that do not have a time-sensitivity. After receiving a voice command, the disclosed smart electronic faucet system analyzes the command to determine if the command is time-sensitive and if so, uses computing resources located within the faucet to process the command to determine the control action to be taken by the faucet. Otherwise, the faucet sends the command to a server computer or computing device that is remotely located but communicatively connected to the faucet, which then processes the command and returns the control action to be taken by the faucet.