16 results about "Content filtering" patented technology

The application relates to the technical field of artificial intelligence, in particular to a multi-modal human-computer interaction content filtering method, device and medium. The method comprises the following steps: collecting multi-modal interaction data; obtaining a current application scene, calculating single invalidity scores of various data, obtaining joint invalidity scores in combination with multi-modal weights corresponding to the scene, and judging whether the joint invalidity scores are lower than a preset threshold. If yes, scene-specific violation judgment weights are called, multi-modal violation features are extracted from the data, and current fusion features are obtained by combining the weights; the latest N (N>1) historical fusion features are obtained, the current fusion features are combined to generate modified fusion features, the modified fusion features are input into a pre-trained deep learning model, and a risk confidence is obtained. When the risk confidence exceeds a preset threshold, an interaction instruction transmission link is immediately cut off, and multi-modal data is encrypted and stored as evidence, so that accurate filtering and high-risk prevention and control of multi-modal interaction content are realized. The application can effectively improve the judgment accuracy.

Methods and a system for personalizing self-checkout terminals based on customer loyalty information includes receiving customer preferences during registration, generating themed personas with specific color schemes, graphics, and animations relevant to the customer's registered interests, and dynamically applying personalized interface themes when customers are identified at self-checkout terminals. The system maintains cached personas tagged with attributes like interests, religious preferences, and dietary restrictions, enabling sub-second personalization by matching customer preferences to pre-generated themes. The system avoids showing inappropriate content by filtering themes based on offensive content, collects user feedback on displayed themes through a rating mechanism, and employs time-delay factors to prevent theme repetition. New personas and corresponding themes are periodically generated to maintain customer engagement while keeping processing costs manageable.

The application discloses a content filtering method and system based on role portraits and a storage medium. The method comprises the following steps: a client establishes a first role portrait library for a user role, and sends the first role portrait library to a server; the server learns each received first role portrait library, and generates a second role portrait library for each role; when the client receives conversation information sent by the user role, it is judged whether the user role matches the first role portrait library; if yes, the first role portrait library is used to filter the content of the conversation information; otherwise, the client sends the conversation information to the server, the server matches a corresponding second role portrait library according to the user role, and uses the second role portrait library to filter the content of the conversation information. The application solves the problems of the prior art, such as rigid filtering rules, incapability of combining with context, and uneven load of the client and the server, and realizes intelligent, accurate and adaptive content filtering.

This invention relates to a method, apparatus, device, and storage medium for generating pet companion voice based on voiceprint modeling. The method includes: acquiring the voiceprint features of a target user and constructing a personalized timbre model based on a quality score; acquiring pet behavior data collected by a pet companion terminal, identifying the pet's behavior categories and the intensity of the behavior; determining a basic tone template based on the behavior categories, and dynamically adjusting the speech parameters in the basic tone template according to a preset mapping relationship based on the behavior intensity to obtain target tone parameters; and synthesizing target companion voice based on the target tone parameters, the personalized timbre model, and target text matching the behavior category, and sending it to the pet companion terminal for playback. This invention achieves emotional and adaptive automated pet companionship, and ensures modeling accuracy and usage security through a recording quality scoring mechanism and a sensitive content filtering mechanism.

This invention discloses a multi-application message monitoring method and system based on audio features, belonging to the field of computer application technology. The method includes: acquiring process information of target applications; creating independent audio capture streams for each target application based on content filtering functionality using screen capture technology to achieve application-level audio isolation; acquiring audio data in real time through the audio capture streams and calculating volume; using a state machine to identify short-duration prompt tone features in the audio data according to preset volume thresholds and duration ranges; filtering repetitive sound events for each target application within a preset time window; and generating and outputting message events when a sound event matching the features is identified. This invention combines application-level audio isolation with state machine recognition to achieve non-intrusive, highly accurate unified monitoring of multi-application messages, effectively solving the problems of strong dependence and high intrusion in existing technologies, and is suitable for multi-platform message management scenarios such as e-commerce customer service.

This invention provides an intelligent customer acquisition and automated marketing method based on multimodal behavioral sequences and causal gain attribution. The method includes: performing identity parsing, data cleaning and standardization, and data quality monitoring on the original event table to obtain a basic data table; processing the basic data table to obtain multimodal time series representations and user vector representations; constructing a causal gain attribution framework to obtain the true incremental contribution of a single user to a marketing action; performing cross-channel attribution and interpretability analysis to obtain attribution results; performing content representation and embedding, and content filtering to obtain personalized content recommendations; performing automated marketing orchestration and execution, as well as privacy protection and ethical risk detection; and finally, monitoring and evaluating the model. This invention enables unified representation of multimodal data, accurate attribution based on causal gain, automated cross-channel marketing orchestration, and privacy and ethical monitoring, significantly improving customer acquisition effectiveness, resource allocation efficiency, and decision-making transparency.

This invention discloses a dynamic insurance product recommendation system based on the user lifecycle, comprising: a data collection and user profiling system, a big data processing system, and a front-end display and interactive interface. The big data processing system is connected to the data collection and user profiling system and the front-end display and interactive interface. The big data processing system includes a dynamic recommendation engine, a user lifecycle model, and an insurance product knowledge base. The dynamic recommendation engine combines the user's current lifecycle stage, user profile, and a preset protection gap analysis model to assess the user's protection needs; it then uses collaborative filtering, content filtering, or a hybrid recommendation algorithm to select the most suitable products from the insurance product database. This invention can provide personalized and timely insurance product recommendations based on the user's specific needs and risk status at different life stages. The system utilizes data analysis, machine learning, and the user lifecycle model to ensure the accuracy and relevance of the recommendations.

The application provides a multi-party system data pushing method and system based on SaaS, and a storage medium. The method comprises the following steps: obtaining a differentiated pushing rule pre-configured by a target tenant, wherein the differentiated pushing rule comprises a field selection rule, a data filtering rule and a data format conversion rule for a target interface; obtaining original data to be sent to the target tenant, performing field screening on the original data based on the field selection rule to obtain first intermediate data; performing content filtering on the first intermediate data based on the data filtering rule to obtain second intermediate data; performing format conversion on the second intermediate data based on the data format conversion rule to obtain to-be-transmitted data meeting the customized format requirement of the target tenant; and sending the to-be-transmitted data to a receiving system corresponding to the target tenant. The application can safely, efficiently and at low cost push multi-party system data.

Disclosed herein are system, apparatus, article of manufacture, method and / or computer program product embodiments, and / or combinations and sub-combinations thereof, for user-based content filtering. A type of content indicated by a user profile associated with a request for a content item may be identified. An occurrence of the type of content within the content item may be identified. Filtering data that comprises one or more instructions for a media device to modify (e.g., blur, alter, obfuscate, etc.) a visual representation of the identified occurrence of the type of content during playback may be generated. The content item and the filtering data may be sent to a media device associated with the user profile.

This invention relates to the intersection of railway information security and multimedia processing, specifically a method for graded masking and content filtering of privacy data in railway communications. By configuring multi-level user permission rules, it applies varying degrees of privacy processing to keyframe images output by sensitive target detection targets. This allows for accurate identification of sensitive targets and masking only them to preserve background information. The masking intensity is dynamically adjusted based on user permissions, achieving effective voice desensitization and accurate keyword detection. Simultaneously, it establishes a comprehensive operation log and reverse audit closed loop. The advantages of this invention are: achieving a balance between information preservation and accurate masking; a role-based dynamic masking mechanism that meets the information detail requirements of different users, significantly improving flexibility compared to static masking techniques; high comprehensiveness of voice processing; a reliable audit closed loop ensuring traceability of operations and controllability of decryption; and wide applicability, suitable for railway communication scenarios in various fields such as government affairs and emergency response.

The application provides a multi-modal content aggregation and expression normalization method and device for social big data. It relates to the technical field of natural language processing. The method comprises: performing content filtering on an original social text set to obtain an effective text content set; performing text coding on each effective text content and splicing the effective text content with a time feature vector corresponding to the effective text content to obtain a corresponding multi-modal fusion vector; performing similarity calculation on any two multi-modal fusion vectors and grouping the effective text content set into multiple content groups according to the similarity; selecting key information by using a multi-path query driven attention Top-k mechanism and obtaining an aggregation result vector through fusion gating; determining the center content of each content group; and selecting a preset number of center contents as the content output of multi-modal content aggregation and expression normalization. The application realizes accurate aggregation and primary and secondary discrimination of multi-modal social content under a label-free condition.

A recommendation system includes a content similarity analyzer configured to determine a first set of content item identifiers similar to a set of viewed content items based on respective similarity scores and add them to a first list. A similar user content extraction module identifies a set of similar user identifiers from a user similarity database; obtains, based on respective viewing histories of the set, a second set of content item identifiers; and adds them to the first list. The recommendation system includes a content filter configured to select a subset of content item identifiers from the first list based on the corresponding similarity scores between content item identifiers of the first list and a viewing history. The content filter is configured to transmit the subset of content item identifiers for display of the corresponding content items via a web portal on a user interface of a first user device.

Various approaches relate to user defined content filtering in media playing devices of undesirable content represented in stored and real-time content from content providers. For example, video, image, and / or audio data can be analyzed to identify and classify content included in the data using various classification models and object and text recognition approaches. Thereafter, the identification and classification can be used to control presentation and / or access to the content and / or portions of the content. For example, based on the classification, portions of the content can be modified (e.g., replaced, removed, degraded, etc.) using one or more techniques (e.g., media replacement, media removal, media degradation, etc.) and then presented.

Techniques for filtering out undesirable generative machine learned model (or LLM) output are discussed herein. A system may receive a subset of an LLM output. That is, the system may stream the LLM output to a user device by receiving one or more tokens from the LLM and outputting such token(s) to a user device. However, prior to outputting the token(s) to the user device, the system may determine whether the token(s) include undesirable content that is to be blocked. The system may use synchronous blocking components (e.g., blocks the undesirable token(s) before such token(s) get output to the user device) and / or asynchronous blocking components (e.g., blocks the undesirable token(s) after the token(s) have been output to the user device) to filter out undesirable content. The synchronous and / or asynchronous blocking components may be designed to block one or more undesirable topics such as hateful speech, profanity, bias, toxicity, factualness, etc.