43 results about "Relevance prediction" patented technology

Computer-implemented techniques for multimodal content relevance prediction using neural networks involves processing multimodal content comprising a digital image and text. Initially, dense embeddings are obtained: an image embedding from a pretrained convolutional neural network, and a text embedding from a pretrained transformer network. These embeddings encapsulate the features of the image and text respectively. Two pretrained dense neural sub-networks then reduce the dimensionality of these embeddings. A third dense neural sub-network determines a numerical score for the multimodal content using the reduced embeddings and an additional feature embedding. This score reflects various aspects of the multimodal content, leading to an action taken based on this numerical evaluation, providing a comprehensive and nuanced understanding and management of multimodal digital content.

The invention discloses a time-varying channel estimation method and device based on correlation prediction in an RIS auxiliary system, and the method comprises the steps: assuming that the working mode of the RIS is an ON or OFF mode, and obtaining time domain cascade channel estimation at a historical moment at a base station side; constructing a training sample set by using time domain cascade channel estimation at historical moments; randomly initializing network parameters, and training the long and short-term memory neural network by using the training sample set to obtain a network model with an optimal weight and threshold; predicting the time correlation of the channel based on the trained network model; generating an AR random model by using the time correlation of the channel, and constructing a state space model of the Kalman filter based on the AR random model and the time domain receiving signal; time-varying channel estimation is carried out based on the state space model of the Kalman filter, and cascade channel estimation passing through each RIS unit is obtained. The method is suitable for high-precision acquisition of the time-varying channel information in the RIS-assisted mobile communication system.

Embodiments of the present disclosure provide systems and methods for emotion-based call summarization. One method may include receiving an emotion prediction vector for an utterance text segment from a transcript data object, the emotion prediction vector comprising a plurality of emotion prediction scores respectively corresponding to a plurality of emotion identifiers; generating a domain-specific relevancy prediction for the utterance text segment based on a category-relevant subset of the plurality of emotion prediction scores that correspond to one or more category-specific emotion identifiers of the plurality of emotion identifiers associated with a domain-specific summarization category; identifying the utterance text segment as a relevant utterance from the transcript data object based on a comparison between the domain-specific relevancy prediction and a relevancy threshold; and initiating a performance of a machine learning summarization operation based on the utterance text segment.

The invention discloses a subway short-time OD passenger flow prediction method based on a multi-knowledge graph neural network model. Firstly, based on multi-source data such as subway station position data, AFC data, POI data and mobile phone signaling data, a subway network structure, passenger flow characteristics and subway station surrounding land and population characteristics are quantified, key OD pairs are extracted, and the influence of low-value and random data on a model is prevented; secondly, constructing spatial connectivity, travel mode similarity and function similarity among a plurality of knowledge graph capture nodes; meanwhile, constructing a spatial feature modeling module, and capturing a potential spatial relationship between nodes by using a graph neural network; and finally, constructing a time feature modeling module, and capturing a passenger flow dynamic change process by using a time sequence modeling method. According to the prediction method, complex spatial-temporal correlation existing in subway passenger flow is effectively and quantitatively captured, and the prediction capability and the interpretability of the model are remarkably enhanced.

The application discloses a multi-agent collaborative perception feature enhancement method based on context aggregation, relates to the technical field of automatic driving and intelligent networked vehicles, and comprises the following steps: acquiring current frame features and at least one frame of historical features; adaptively aligning the historical features based on a motion prediction network and a deformable convolution offset; generating semantic consistency weights and identifying time sequence discontinuous regions based on a correlation prediction network; fusing the current and historical features according to the weights and carrying out state space selective scanning, pooling, denoising and aggregation; generating multi-scale features, combining scene complexity, quality evaluation and time sequence consistency constraints to determine scale weights and then fusing; and inputting an LSTM to perform time sequence modeling and output enhanced context features. Invalid alignment is avoided through motion compensation and semantic consistency constraints, noise and scale jitter are suppressed through position-level denoising and multi-scale adaptive smoothing, and the collaborative perception robustness and detection accuracy are improved.

The disclosure provides a relevance prediction model training method and device, and an abstract generation method and device. The relevance prediction model training method comprises: extracting a first semantic feature vector of a first sentence sample and a second semantic feature vector of a second sentence sample; generating a training sample, wherein the training sample comprises the first semantic feature vector and the second semantic feature vector, and a preset semantic relevance label of the first sentence sample and the second sentence sample; inputting the training sample into a machine learning model to obtain a semantic relevance prediction result of the first sentence sample and the second sentence sample; determining a loss function according to the semantic relevance label and the semantic relevance prediction result; and training the machine learning model by using the loss function to obtain a relevance prediction model.

The present application discloses a laparoscopic image segmentation method, system and computer storage medium. The method includes: acquiring a laparoscopic image; inputting the laparoscopic image into a segmentation model to obtain a corresponding segmentation map. The segmentation method of the segmentation model for the laparoscopic image includes: extracting hierarchical semantic features and spatial detail features of the laparoscopic image; fusing the hierarchical semantic features and the spatial detail features to obtain a correlation prediction matrix; using the correlation prediction matrix to optimize the fused features after the fusion of the hierarchical semantic features and the spatial detail features to obtain an optimized feature map; performing semantic segmentation on the optimized feature map to obtain a segmentation map. The present application uses a dual-path feature extraction technology to solve the problem of information loss in a single-path network, and uses a context information guidance technology to model category correlations, and uses this information to optimize the features extracted by the dual-path, improving the segmentation accuracy of laparoscopic images and being particularly suitable for the segmentation of laparoscopic images.

A device receives comment information that is associated with users and includes comments provided by the users, about an entity, via social media sources, and receives transaction information that is associated with the users and includes financial transactions of the users with the entity. The device determines correlations between the comment information and the transaction information, where the correlations between the comment information and the transaction information provide weights to the comment information to generate weighted comment information. The device generates a prediction about a future stock price of the entity based on the weighted comment information, the transaction information, and the correlations between the comment information and the transaction information, and provides the prediction about the future stock price of the entity for display.

The invention discloses a model training method, a knowledge retrieval method, electronic equipment and a storage medium. The model training method comprises the following steps: inputting a first user question of a user, first context information of the first user question and a first knowledge entry into a first model to obtain predicted correlation between the first user question and the first knowledge entry; wherein the first context information is obtained from historical dialogues with the user, and after the first model encodes the first user question, the first context information and the first knowledge item, a coding vector of the first user question and a coding vector of the first context information are fused to obtain a first fusion vector; performing correlation prediction based on the first fusion vector and the coding vector of the first knowledge entry to obtain predicted correlation; parameters of the first model are adjusted based on the predicted correlation and a reference correlation between the first user question and the first knowledge entry.

This invention belongs to the field of image segmentation technology and discloses a method, system, medium, device, and terminal for extracting identification documents. The system utilizes an embedded device to acquire multispectral images of the identification document; it models the document image edge segments based on their straight-line geometric properties and infers from the contextual information of the global image; it extracts global image features using a ResNet network and then encodes them using a Deformable DETR encoder; in the first-stage decoding process, it predicts the document edge segments using an attention mechanism and learnable line segment and position queries; in the second-stage decoding process, it compares the correlation between line segment features and image features to predict the relative order between line segments, and obtains the complete edges, vertices, and bounding boxes of the document image through perspective transformation. This invention uses a two-part matching method to predict line segments, avoiding pre- and post-processing, simplifying the detection channel, and achieving true end-to-end processing.

The application provides a point cloud encoding method based on a multi-level spherical octree and a graph-driven attention entropy model, comprising: encoding point cloud data using an octree entropy model based on multi-level spherical coordinates; constructing an adjacency matrix of a parent graph and a distance graph; using a graph convolution network module to embed context information with the aid of the adjacency matrix; constructing a grouping graph attention module and a cross attention module to learn the relevance of parent node context and sibling node context; predicting the probability of each octree node placeholder symbol; compressing the placeholder symbol sequence into a binary floating point number sequence and converting it into a bit stream; converting the bit stream into an octree placeholder symbol sequence, reconstructing the octree and restoring the point cloud. The application combines multi-level spherical coordinate octree structure, graph convolution, grouping graph attention module and cross attention module, reduces quantization error, improves compression efficiency, balances calculation complexity and compression effect, and significantly enhances the adaptability to high-resolution point cloud data.

The present application provides a kind of shale oil horizontal well volume fracturing ultimate recoverable reserves prediction method, first based on the reservoir geological parameters of the horizontal well to be measured and adjacent horizontal well in the same block, establish productivity prediction model, calculate effective fracture network sweep efficiency using reservoir numerical simulation method, and establish effective fracture network sweep efficiency and maximum recoverable reserves correlation prediction chart;Secondly, the correlation coefficient between the geology and volume fracturing reconstruction parameters of the horizontal well to be measured and the fracture network sweep volume is calculated using grey relational analysis method, and the key control parameters affecting fracture network sweep volume are determined;Finally, the fracture network sweep volume prediction model coupled with key control parameters is established, the fracture network sweep volume of the horizontal well to be measured is obtained, the effective fracture network sweep efficiency is obtained, and the maximum recoverable reserves of the horizontal well to be measured is obtained using the correlation prediction chart. The method can quickly predict the maximum recoverable reserves of any horizontal well by establishing the correlation chart of effective fracture network sweep efficiency and maximum recoverable reserves.

The present disclosure relates to field of wireless communication network that discloses method and system for configuring initial congestion window value in User Equipment (UE (201)). UE (201) determines for each of one or more applications running in UE (201), user experience parameters and UE (201) network parameters. Further, UE (201) predicts using trained Artificial Intelligence (AI) model (203), an optimal initial congestion window value for each of one or more applications based on correlation of user experience parameters and UE (201) network parameters corresponding to each of one or more applications. Finally, UE (201) configures optimal initial congestion window value for each of one or more applications based on prediction. The present disclosure helps in reducing Flow Completion Time (FCT) of each application by predicting optimal initial congestion window value dynamically.

Dynamic conversation alerts are provided within a communication session. In one embodiment, the system presents, to a client device associated with a user of a communication platform, a user interface (“UI”) including a prompt for the user to submit one or more alert phrases, each alert phrase being associated with a category; receives, from the client device, a list of submitted alert phrases; and receives a transcript of a communication session between participants. For each utterance in the transcript, the system determines whether one or more predictions of relatedness are present between the utterance and one or more alert phrases from the list of submitted alert phrases. The system then transmits, to the client device, a list of related categories, each related category including one or more timestamps of utterances for which a prediction of relatedness is present for an alert phrase associated with that category.

The invention discloses a method for predicting the minimum miscible pressure of CO2 and crude oil by using machine learning, and the method comprises the steps: carrying out the high-precision modeling based on a linear SVM algorithm, building a correlation prediction formula, and constructing a correlation prediction model; and training and optimizing the correlation prediction model by using the collected data, and establishing a stacked integrated model in combination with a random forest regression model RF and the correlation prediction model to realize accurate prediction of the CO2-crude oil minimum miscible pressure MMP. According to the method, the time and the cost for determining the CO2-crude oil MMP value can be reduced to the maximum extent. Therefore, a reliable CO2-crude oil MMP value can be provided through prediction, and the miscible phase type of CO2 oil displacement is determined, so that the design of a CO2 oil displacement scheme is improved, the optimal design of an injection-production scheme is guided, and the crude oil recovery rate is increased.

The invention discloses a semantic correlation prediction model, method and device, a storage medium and computer equipment, and relates to the technical field of Internet. The models comprise a student model and an online prompt model, and the student model comprises a parameter sharing twin-tower network used for performing feature extraction on a first keyword and a second keyword to be predicted to obtain a first keyword feature and a second keyword feature; the thinking chain tower network is used for carrying out feature extraction on thinking chain prompt information generated by the online prompt model to obtain thinking chain features, and parameters of the thinking chain tower network and parameters of the double-tower network are not shared; the expert hybrid network is used for carrying out feature fusion on the first keyword feature, the second keyword feature and the thinking chain feature to obtain a keyword fusion feature; and the deep neural network is used for performing semantic correlation prediction based on the keyword fusion features to obtain a semantic correlation prediction result of the first keyword and the second keyword. The above model can improve the accuracy of correlation prediction.

The application belongs to the technical field of audio processing, and a plurality of pre-trained audio text expert models are introduced to construct a multi-expert audio text semantic representation. Meanwhile, the consistency relationship between audio and text at the overall semantic level and the semantic inconsistency between audio and text that is perceptually significant to humans are modeled through a semantic alignment branch and a semantic mismatch branch respectively, and a multi-branch correlation score fusion mechanism is used to output an audio text correlation prediction result that is highly consistent with human subjective evaluation. This method can simultaneously consider semantic consistency and perceptual differences, effectively achieving more accurate and more human subjective evaluation standard-compliant audio text correlation evaluation. Meanwhile, based on the above method, the application constructs an audio retrieval system that evaluates and sorts the correlation between a text query and audio samples in an audio library to realize an audio retrieval function oriented to natural language description.

Methods and systems provide for dynamic conversation alerts within a communication session. In one embodiment, the system presents, to a client device associated with a user of a communication platform, a user interface (“UI”) including a prompt for the user to submit one or more alert phrases, each alert phrase being associated with a category; receives, from the client device, a list of submitted alert phrases; and receives a transcript of a communication session between participants. For each utterance in the transcript, the system determines whether one or more predictions of relatedness are present between the utterance and one or more alert phrases from the list of submitted alert phrases. The system then transmits, to the client device, a list of related categories, each related category including one or more timestamps of utterances for which a prediction of relatedness is present for an alert phrase associated with that category.

The present application relates to a method, apparatus, computer device, and storage medium for determining the access power of sensors. The method includes: obtaining sensor data within a preset time period, inputting the sensor data within the preset time period into a trained correlation prediction neural network, and outputting a prediction matrix, where each element in the prediction matrix includes the occurrence probability of a corresponding emergency event type and the correlation coefficient between the corresponding emergency event type and the corresponding sensor type; determining a target emergency event and a target sensor according to the occurrence probability of the corresponding emergency event type and each correlation coefficient; and determining the scheduling decision and access power of the target sensor through an optimal decision formula. Using this method can improve the accuracy of sensor access and reduce the energy consumption of sensor access.

The application realizes a stream music recommendation method for removing attention deviation based on reward correction through the method in the network security field. The core of the method contains three modules: a reward correction model, an attention prediction module and a correlation prediction module. The reward correction module obtains the unbiased reward after correction by combining the predicted user attention and correlation through the importance sampling method; the attention prediction module models the probability of user attention to each song; the correlation prediction module is used to predict the preference of the user to each candidate song, and the parameter is updated based on the corrected reward. The application eliminates the attention deviation in the user feedback to obtain unbiased reward and improve the prediction accuracy of the model.

This application discloses a method, apparatus, medium, device, and program product for processing pre-training data, relating to the field of artificial intelligence technology. The method includes: acquiring an initial dataset for a target knowledge domain, the initial dataset comprising multiple pre-training data retrieved for the target knowledge domain; performing data quality prediction on the pre-training data in the initial dataset based on a first evaluation model to obtain a quality evaluation result; performing domain relevance prediction on the pre-training data in the initial dataset based on a second evaluation model to obtain a domain relevance prediction result, the relevance prediction result indicating the relevance between the pre-training data and the target knowledge domain, the first and second evaluation models being constructed based on a large-scale language model; filtering the initial dataset based on the quality evaluation result and the relevance prediction result to obtain an intermediate dataset; and performing fine-tuning and filtering on the intermediate dataset to obtain a target pre-training set. This application can efficiently acquire a large amount of high-quality pre-training data.

This disclosure relates to a training method for a video recognition model, a video recognition method, and related equipment. The method includes: acquiring a video sample set, wherein the video samples in the video sample set include display data, audio data, and correlation annotation data between the display data and the audio data; extracting display text data from the display data; performing speech recognition on the audio data to obtain playback text data; inputting the display text data and the playback text data into a first network structure of the video recognition model to obtain a text relation vector; inputting the text relation vector into a second network structure of the video recognition model to obtain correlation prediction data; and training the network parameters corresponding to the second network structure based on the target loss determined by the correlation prediction data and the correlation annotation data to obtain a video recognition model that meets preset conditions. The obtained trained video recognition model can quickly and accurately recognize videos.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting a return to a site. In some implementations, a system obtains data indicative of a time evolving movement of a user interacting with a website shown on the client device. The system determines, using a first trained machine learning model and based on the data indicative of the time evolving movement, a metric associated with an emotion of the user corresponding to the user's interaction with the website. The system obtains, from a metric database, metrics associated with an identifier of the user. The system provides, to a second trained machine learning model, (i) the metric associated with the emotion and (ii) data representing the obtained metrics associated with the identifier. The system generates, using the second trained machine learning model, a prediction indicating whether the user is likely to return to the website.

This invention discloses a method for predicting low-speed turbulent boundary layer noise considering the correlation of flow space, belonging to the field of near-wall noise environment for aircraft. This invention selects key flow field parameters for predicting low-speed turbulent boundary layer noise by analyzing the correlation characteristics of flow space; and obtains ωδ* / U for predicting the power spectrum of turbulent boundary layer noise in different segments. ∞ The values ​​were calculated, and a prediction model for the turbulent boundary layer noise power spectrum was constructed to predict the turbulent boundary layer noise power spectrum; the convection velocity U was analyzed and obtained. c This invention is a key parameter determining the spatial correlation of turbulent boundary layer noise, and it constructs a spatial correlation prediction model for turbulent boundary layer noise. By incorporating the convection velocity into the prediction model, the convection velocity of the turbulent boundary layer noise is predicted, thus obtaining the spatial phase distribution of the noise. This invention can simulate the time-frequency characteristics of noise and accurately reproduce the spatial correlation characteristics of low-speed turbulent boundary layer noise, improving the prediction accuracy of turbulent boundary layer noise.

This invention discloses a commonsense question answering method and system based on dynamic global semantic fusion. The method involves obtaining an initial representation of the question context for the commonsense question to be answered; predicting a soft mask based on the correlation between the initial representation of the question context and the entity representation of the initial knowledge subgraph; generating a dynamic knowledge subgraph based on the soft mask; extracting features from the dynamic knowledge subgraph to obtain its initial representation; fusing word-level local information between the question context and the initial representation of the dynamic knowledge subgraph to obtain a word-level local information fused representation of the question context and the dynamic knowledge subgraph; performing multi-head attention learning on the word-level local information fused representation of the question context and the dynamic knowledge subgraph to obtain an updated representation of the question context and the dynamic knowledge subgraph; pooling the updated representation of the dynamic knowledge subgraph to obtain a pooled representation of the dynamic knowledge subgraph; and obtaining the answer to the commonsense question based on the pooled representation of the dynamic knowledge subgraph.

A prediction method includes a calculation process and a prediction process. The calculation process calculates a correlation between a spatial distribution value of a magnetic field in a chamber when a plasma etching process is performed on a substrate disposed in the chamber, and a process result of the plasma etching process on the substrate. The prediction process predicts the process result of the plasma etching process on the substrate from the spatial distribution value of the magnetic field in the chamber based on the calculated correlation.

The application discloses a navigation route processing method and device, a storage medium and an electronic device, and can be applied to the field of maps. The method comprises the following steps: determining a group of navigation routes and a group of parking lots between a navigation starting point and a navigation ending point, obtaining a group of parking paths, each parking path comprising a combination of a navigation route and a parking lot; determining a vacant parking space feature for describing a historical vacant parking space of each parking lot in a target period, the target period being a period from the navigation starting point to the navigation ending point; determining a traffic flow state feature for describing a historical traffic flow state of each navigation route corresponding to the target period; inputting the vacant parking space feature of each parking lot and the traffic flow state feature of each navigation route into a target prediction model to obtain a target parking path, the target prediction model being used for predicting a selection probability of each parking path based on the correlation between the vacant parking space feature of each parking lot and the traffic flow state feature of each navigation route.

The application discloses a trajectory planning method and device, equipment, vehicle, medium and product. The method comprises the following steps: acquiring first environment information of an environment where a target vehicle is located, the first environment information being used for describing a plurality of first traffic participants in the environment where the target vehicle is located; predicting, by a pre-acquired task correlation predictor, based on the first environment information, to obtain a task correlation of each first traffic participant, the task correlation of the first traffic participant being used for representing an influence degree of the first traffic participant on a current driving behavior of the target vehicle; screening the plurality of first traffic participants according to the task correlation of each first traffic participant to obtain a target traffic participant; and performing trajectory planning on the target vehicle according to the target traffic participant to obtain a target trajectory. The first traffic participant with a larger task correlation can be screened out to participate in subsequent trajectory planning on the target vehicle, so that attention to redundant information is reduced, and the accuracy of trajectory planning is improved.

The present disclosure relates to a text processing method, a model training method, a device, an apparatus and a storage medium. The text processing method comprises: comparing text content in a to-be-labeled document with text content in each initial label respectively, and determining a candidate labeling label from each initial label according to a comparison result; inputting the to-be-labeled document and the candidate labeling label into a relevance prediction model to perform relevance prediction, to obtain an initial relevance parameter of the to-be-labeled document and the candidate labeling label; obtaining, from a preset database, a supplementary document feature corresponding to the to-be-labeled document and a supplementary label feature corresponding to the candidate labeling label; adjusting the initial relevance parameter by using a relevance adjustment model through the supplementary document feature and the supplementary label feature to obtain a target relevance parameter; and determining a target labeling label from the candidate labeling label based on the target relevance parameter. In this way, the accuracy of determining the target labeling label can be improved.

This invention discloses a method for locating generator units based on low-frequency oscillation mode correlation in a power system. It constructs a generator unit location model guided by modal information and combines it with a phased training strategy to achieve coordinated optimization of low-frequency oscillation mode identification and related generator unit location. Specifically, in the first phase, the first branch of the model is trained using generator operation measurement data and power system topology information. In the second phase, the second branch is trained based on the low-frequency oscillation mode parameters output from the first phase and the generator power angle time series. During generator unit location, the two types of features are fused through feature mapping and weight allocation mechanisms, allowing oscillation mode information to participate in the weight calculation of generator time series features, thereby obtaining the correlation prediction results of each generator under different oscillation modes.