176 results about "Multiple category" patented technology

A large language model (LLM) may be used to classify an input into one of a plurality of categories. However, given the machine-learning operation of the LLM, the output of the LLM does not represent a definitive statement, but is based on probability computations of the machine learning model. Therefore, the classification performed by the LLM might not be correct. Classification into the wrong category by the LLM results in downstream technical problems. In some implementations, when an LLM generates a response that classifies an input, one or more probability values associated with a token that forms the basis of the response may be used to determine a confidence value. The confidence value is indicative of confidence in the classification performed by the LLM. An action may be taken based on the confidence value.

To efficiently and effectively classify a large amount of document and text data by an LLM.SOLUTION: The present invention relates to a text classification system 1 which classifies documents accumulated in a document DB 14 by categories, and the text classification system has a classification processing part 12 which lets an LLM 2 proposes one or more categories based upon a classification policy specified by a user, a search processing part 13 which searches the document DB 14 for documents belonging to the respective proposed categories, and a UI processing part 11 which presents the respective categories and the numbers of documents belonging to the respective categories to the user.SELECTED DRAWING: Figure 1

A large language model (LLM) may be used to classify an input into one of a plurality of categories. However, given the machine-learning operation of the LLM, the output of the LLM does not represent a definitive statement, but is based on probability computations of the machine learning model. Therefore, the classification performed by the LLM might not be correct. Classification into the wrong category by the LLM results in downstream technical problems. In some implementations, when an LLM generates a response that classifies an input, one or more probability values associated with a token that forms the basis of the response may be used to determine a confidence value. The confidence value is indicative of confidence in the classification performed by the LLM. An action may be taken based on the confidence value.

A fraud call recognition method and system based on causal graph and agent collaboration comprises the steps of obtaining a fraud ASR text, mapping extracted multi-category to-be-mapped objects into semantic concept variable vectors, then learning a causal structure based on a data matrix formed by all semantic concept variable vectors, constructing a fraud causal graph, and recognizing a fraud call in the fraud ASR text. Calculating a causal effect of the reason variable on the result variable, and storing the causal effect as a relation weight in a causal graph; the method comprises the steps of obtaining a to-be-recognized ASR text, mapping multiple categories of to-be-mapped objects extracted from the to-be-recognized ASR text into semantic concept variable vectors, then retrieving matched causal chains in a causal graph to obtain a plurality of candidate causal chains, and then generating fraud judgment and risk degree evaluation values based on a large language model inference engine. And determining whether the to-be-identified call is a fraud call. The invention relates to the field of telecommunication anti-fraud, can deeply fuse a causal knowledge base and a large language model, and effectively improves the effectiveness, adaptability and reliability of communication anti-fraud identification.

Methods, systems, and devices for classification. In one aspect, a method includes receiving an input and a request to classify the input into one of a plurality of categories, processing the input using a multi-modal model to generate (i) a description of the input and (ii) a category prediction, a description of the input and the category prediction are processed using a text encoder embedding neural network to generate (i) a text description feature embedding and (ii) a predictive feature embedding, a query feature embedding representing the input being generated from at least the description feature embedding and the predictive feature embedding, and classifying the input into one of a plurality of categories using the query embedding.

A database query processing method includes receiving a natural language request for information contained within a database, categorizing the natural language request in a particular category from a plurality of categories, and selecting a large language model and one or more database structures for which a SQL request is to be directed based in part on the particular category. The selected large language model is prompted to generate a SQL request directed to the selected set of one ore more database structures and the received SQL request is validated for the particular category. The validated particular SQL request is used to access data within the database and one or more visualizations is caused to be displayed based at least in part on a result of the validated particular SQL request.

The disclosure presents systems and methods for hierarchical classification of input data across a plurality of categories. A machine learning model processes various data formats, starting with dimensional reduction using tokenization techniques, such as Bert-tiny tokenization, to create model-readable representations. The system predicts super-categories, sub-categories, and granular categories through selective activation of sub-layers tied to identified super-categories, optimizing computational efficiency. Label smoothing during training mitigates overconfidence in predictions, while softmax normalization refines inference outputs. Synthetic data generation using Large Language Models (LLMs) supplements training datasets, and an automated data labeling pipeline efficiently generates hierarchical labels. Modifications to the model, such as stop word removal and file size limitations, further reduce latency. Inference analyzes logits to predict hierarchical paths, providing detailed classifications with clear outputs. The method is adaptable for multimodal formats, ensuring scalable and accurate predictions across diverse data types while minimizing computational costs and improving reliability.

Aspects of the present disclosure relate to hierarchical content organization techniques, which may thus enable improved content indexing and / or retrieval, among other benefits. In examples, a content platform maintains a content hierarchy, where, for example, each category within the content hierarchy has an associated set of categories / subcategories. It will be appreciated that any number of hierarchical levels may be used. Accordingly, a post or other content item may be associated with one or more categories / subcategories (e.g., one or more nodes) of the content hierarchy, thereby facilitating subsequent retrieval of the content (e.g., responsive to a user query and / or to provide to a search platform). Thus, in contrast to discussion boards and other content management systems, the disclosed aspects ensure content is organized when created, thereby improving later retrieval and / or any of a variety of subsequent processing, among other benefits.

The invention discloses an overstocked material automatic inventory method based on multi-element collaborative cause tracing, relates to the technical field of material management, and solves the problems that a standardized multi-factor quantitative model is not established for evaluation of market demands during external docking, the matching logic of single-category and multi-category materials is not clear, and the market demands are not evaluated. According to the method, multi-dimensional identification is carried out by combining material attributes and dynamic states, misjudgment of a single warehouse age standard is avoided, the accuracy of accumulated material identification is improved, and the accuracy of the accumulated material identification is improved by quantitatively comparing the accumulated remaining amount with a preset value and combining accurate parameter matching of required materials and defining an internal reuse boundary, so that the accuracy of the accumulated material identification is improved. The internal consumption rate of reusable materials is improved, and a weighted quantization model of purchase quantity matching degree + date of delivery collaboration degree + cooperation stability is established for single-category materials; for multiple categories of materials, the combination matching logic of the same subject demand combination score is adopted, the external demand matching precision is improved, and the disposal period is shortened.

This application discloses a display method, apparatus, electronic device, and computer-readable medium. The method includes: when a client is displaying a live video page and a first live video is displayed on the live video page, after the client receives a first operation triggered on the live video page, displaying an object aggregation interface on the live video page, and displaying a first candidate object corresponding to first category description information on the object aggregation interface, the first candidate object being used to describe the first candidate live video; then, after the client receives a trigger operation on a first control on the object aggregation interface, displaying at least one second category description information on the object aggregation interface, so that the user can view live videos under multiple categories through the object aggregation interface, thereby better meeting the user's live video viewing needs and effectively improving the user's live video viewing experience.

Systems and methods for providing information responsive to a user search are provided. Such a method includes receiving a search query from a client device and generating a plurality of categories associated with the search query using a first artificial intelligence (AI) model. The method also includes selecting a plurality of sponsored content options that are responsive to the search query and, for each category, generating respective freeform descriptive text for the category using a second AI model. The method also includes causing the client device to present a user interface that includes (i) the respective freeform descriptive text for each category and (ii) links to third party information resources associated with the selected plurality of sponsored content options. An arrangement of the links within the user interface indicates which of the plurality of sponsored content options correspond to which of the plurality of categories.

The invention relates to the technical field of model lightweight, in particular to a construction full-process simulation method for construction of a green assembly type comprehensive pipe gallery, and the method comprises the steps: dividing all components in a pipe gallery model into a plurality of category clusters; acquiring a central component of each category cluster; acquiring a vertex important index of each vertex according to the density and size of the triangular surface at each vertex; according to the average level of the vertex important indexes of the two vertexes of each edge, the triangular face included angle of each edge and the edge degree, the folding delay degree of each edge is obtained, then the folding edges in the center components of all the category clusters are obtained, and then the folding edges of all the center components are folded; and obtaining the simplification rate of the center member of each category cluster according to the ratio of the number of the triangular facets of the center member of each category cluster after folding to the number of the triangular facets before folding, and then carrying out lightweight processing on all the members in each category cluster. According to the method, the model simplification effect is improved by adaptively calculating the simplification rate of the component in each category cluster.

Fine-tuning AI models is described. According to some aspects, one of a number of pre-trained AI models is selected based on the explicit input and the implicit input. In addition, one of a number of fine-tuning methods is selected. Also, a set of one or more of a plurality of categories is selected, where a categorized data set associated with an organization was classified into the categories using a classifier, and where the selected set of categories identify a selected subset of the categorized data set. A version of the selected subset is used to fine-tune the selected AI model using the selected fine-tuning method.

For at least a selected class attribute of the multiple class attributes, one or more bias metrics are determined that estimate a degree to which a particular workflow (having a set of processing stages) is biased in association with the class attribute. Each user of a set of users is associated with a set of user data to be processed by the particular workflow. At least one of the set of processing stages includes executing a machine-learning model. It can be detected that a bias-mitigation option corresponding to a specific class attribute has been selected. For each of at least two of the set of processing stages: a de-biasing technique is selected; and the processing stage is modified by applying the de-biasing technique. A modified version of the particular workflow (which includes the modified processing stages) is applied to each of a set of input data sets.

Systems or methods for the selection of large language models (LLMs). A system receives a request from a service that hosts an application. The request is configured to be processed by an LLM to generate a response. The system applies a classification model to the request to determine the class of the request. The classification model is a language model trained to receive text and classify the text into a plurality of classes. The system selects an LLM from a plurality of candidate LLMs based in part on the determined class of the request and recommends the selected LLM to the application.

To provide an information processing program and an information processing device capable of improving efficiency of processing related to an inquiry.SOLUTION: An information processing program causes a computer to execute the steps of: receiving, from a user, an inquiry about at least one category among a plurality of categories associated with an ID of the user; specifying an answering entity in response to the inquiry based on content of the inquiry and information associated with the plurality of categories; and responding in accordance with providing means of an answer in response to the inquiry.SELECTED DRAWING: Figure 2

A student performance prediction method, apparatus, electronic device, and computer-readable storage medium, wherein the student performance prediction method includes: obtaining learning behavior data corresponding to different target behaviors among multiple behavior categories of a student within a preset time period; and aggregating the learning behavior data corresponding to the different target behaviors in a preset time unit, and performing feature fusion on the aggregated data separately for each behavior category to obtain a category feature set, determining the multiple category feature sets organized in chronological order as a category feature time series set, inputting the category feature time series set into a pre-trained feature reconstruction network to obtain a reconstructed time series set, and inputting the reconstructed time series set into a student performance prediction model to obtain a performance prediction result for the student. The above student performance prediction method can objectively and efficiently predict the student performance.

The present disclosure relates to systems, software, and computer-implemented methods for automatically identifying handwritten tips. An example method includes obtaining an image of a receipt, where the image includes a reference region in a pre-set format, a plurality of category identifiers, and at least one set of handwritten characters. The reference region and the plurality of category identifiers can be printed. The plurality of category identifiers can be located at pre-set positions relative to the reference region. The method further includes obtaining optical character recognition (OCR) information of the image and identifying the reference region from the OCR information based on the pre-set format of the reference region. The method further includes determining a plurality of amounts based on the OCR information and the reference region, where at least one of the plurality of amounts is associated with a tip category.

The invention discloses a transformer substation multi-source heterogeneous data processing method and system based on space layer superposition, and relates to the technical field of electric power data processing, and the method comprises the steps: classifying the multi-source heterogeneous data of a transformer substation based on a data structure type, and obtaining a plurality of category data sets; pre-processing the plurality of category data sets to obtain a plurality of pre-processed category data sets, and storing the pre-processed category data sets in different domains; establishing a unified model based on equipment common attributes of the transformer substation, and embedding the plurality of preprocessed category data sets into a bottom geographic distribution map based on the unified model; the bottom geographic distribution map comprises a plurality of space map layers, and each space map layer corresponds to one data type; and performing dynamic coupling on the plurality of space layers, and constructing a three-dimensional fusion model with a space-time correlation characteristic. According to the method, the technical problems of long time consumption and weak spatial dimension analysis capability of spatial correlation analysis of multi-source heterogeneous data in the prior art are solved.

A machine learning model is applied to extract keywords from an event associated with network operations. A keyword vector is constructed where each keyword extracted is represented as a unique dimension in the keyword vector. The keyword vector is compared against event category vectors associated with event categories to identify potential matches. A relevance score is calculated for each category based on the comparing. The event is classified into one or more categories based on the relevance scores exceeding a predetermined threshold. A user interface configured to visually display and obtain feedback regarding the classifying is generated. At least one of the event categories is updated based on a user feedback received via the user interface and machine learning retraining.

Certain aspects involve evaluating modeling algorithms whose outputs can impact machine-implemented operating environments. For instance, a computing system generates, from a comparison of a set of estimated attribute values of an attribute to a set of validation attribute values of the attribute, a discretized evaluation dataset with data values in multiple categories. The computing system computes, for a modeling algorithm used to generate the estimated attribute values, an evaluation metric. The computing system provides a host computing system with access to the evaluation metric, one or more modeling outputs generated with the modeling algorithm, or both. Providing one or more of these outputs to the host computing system can facilitate modifying one or more machine-implemented operations.

An artificial intelligence (AI) training method is disclosed. Training data associated with a training task is received. The training data is categorized into a plurality of categories. A set of category groups are generated, wherein each category group of the set of category groups includes one or more of the plurality of categories. A first AI system is trained for the training task using a first subset of the training data. The first subset of the training data corresponds to the one or more of the plurality of categories included in a first group of the set of category groups. A second AI system is trained for the training task using a second subset of the training data. The first AI system is used to generate a first output for the task and the second AI system to generate a second output for the task.

The application provides an attack detection method and device for an electric power private network and a public network interaction node, pre-processes each piece of traffic data obtained from a node to obtain target traffic data containing feature information, classifies all the target traffic data to obtain multiple category data sets, samples and optimizes the target traffic data in each category data set to obtain an optimized category data set, selects features for each feature information in the optimized category data set by using a mutual information feature selection algorithm to obtain a feature selection result, updates the optimized category data set according to the feature selection result, reduces the redundancy of the data in the optimized category data set, obtains a more refined to-be-detected traffic data set, shortens the time required for an attack detection classification model to detect the to-be-detected traffic data set, and improves the classification accuracy of the attack detection classification model.

The embodiment of the invention provides a data transmission method and a data transmission device, and relates to the technical field of communication. According to the method, the data of the first type can be received by sending the first information used for indicating to report the data of the first type. Wherein the data of the first category is one of a plurality of categories, and the data of each category in the plurality of categories is used for AI model training. For example, data of each of the plurality of categories is included in a training set. The training set is used for AI model training. Therefore, under the condition that the first category is the category with insufficient data volume in the training set, the data of the category with insufficient data volume can be collected in a targeted manner, so that the data volume of each category in the training set is sufficient, that is, the balance of the data features of the training set is relatively good, and the training efficiency is improved. And therefore, the generalization of the AI model trained by adopting the training set is relatively good.

Certain aspects of the disclosure provide techniques for categorization by a device. An example method includes receiving input information regarding a plurality of classification targets and a plurality of categories for classification of the plurality of classification targets; generating a plurality of embeddings for the input information using a first model, the plurality of embeddings including: a set of first embeddings associated with the plurality of classification targets, and a set of second embeddings associated with the plurality of categories; determining that a similarity score for the set of first embeddings and the set of second embeddings fails to satisfy a threshold; generating, based on the similarity score failing to satisfy the threshold and using a graph neural network (GNN), a classification of the plurality of classification targets in accordance with the plurality of categories; and outputting information regarding the classification.

The present disclosure relates to a data processing method, an electronic device, a storage medium and a program product, and relates to the field of data processing. The data processing method includes: processing each sample in a first sample set by using a first machine learning model to obtain a prediction probability that each sample is classified into each category of one or more categories; and determining, for the each category, a probability threshold corresponding to the category to maximize a number of positives of the category, wherein a confidence corresponding to the category is not lower than a confidence threshold, the probability threshold is for determining a category to which the each sample pertains based on a prediction probability of the each sample, and the confidence corresponding to the category is a confidence at which an actual precision of classification based on the probability threshold meets a precision condition.

The application discloses an Excel-based data processing and visualization method and system, which comprises the following steps: loading a visualization user interface in an Excel workbook, wherein the user interface comprises interactive controls for selecting independent variable fields, dependent variable fields and grouping fields; receiving non-continuous selection of multiple categories in the grouping field by the user through the interactive controls, and automatically classifying the unselected categories into other groups to form an extended grouping mapping; matching and reorganizing the original data based on the extended grouping mapping by using a hybrid indexing mechanism; automatically generating a dynamic chart according to the reorganized data, wherein the coordinate axis range, legend position, series color and auxiliary line style of the dynamic chart are automatically configured according to the number of groups and the dependent variable value range; performing statistical analysis and multivariate linear regression on the data of each group, and converting the regression results into an executable formula string containing cell references and outputting the executable formula string to a specified cell.

Similarity sensitive diversity is utilized to measure variation in a distribution of item listings along one or more categories. A similarity between category vectors of each category pair in a set of categories is determined and utilized to generate a pairwise similarity matrix. The pairwise similarity matrix may be pruned to remove category pairs below a threshold. Utilizing the pairwise similarity matrix, similarity sensitive diversity between one or more items of a plurality of items may be determined. In various aspects, the similarity sensitive diversity may be utilized to: generate a list of relevant items in an appropriate distribution, suggest refinements of a search query; generate navigation modules; categorize or recategorize the plurality of items; or generate autosuggestions.

A method includes constructing a schema segment from a schema segment definition. The schema segment includes a multitude of schema elements. A multitude of embeddings corresponding to the multitude of schema elements is generated by an embedding model. The multitude of schema elements and the multitude of embeddings is grouped into a multitude of categories. A semantic collision between a first embedding of a first schema element and a second embedding of a second schema element of a category of the multitude of categories may be detected. A set of semantic collisions may be obtained. A large language model (LLM) may resolve the set of semantic collisions to obtain a corresponding multitude of updated schema elements. The schema segment is updated with the corresponding multitude of updated schema elements.

Embodiments of the disclosure provide a method, an apparatus, a device and a storage medium for object processing. The method for object processing includes: in response to receiving a predetermined operation by a user on a first selection control for pre-trained at least one generic model presented in a user interface, selecting the at least one generic model; at least acquiring at least one generic feature that is generated by the at least one generic model and associated with a sample of an object of a target category among a plurality of categories; and training an individual model for processing the object of the target category at least based at least on the at least one generic feature and annotation information of the sample of the object of the target category. Therefore, the efficiency of model training can be improved.