44 results about "Speech analytics" patented technology

An AI-driven system for personalized advertising in real time, where: ◯ an analytics unit to monitor and analyze user behavior in real time across multiple digital platforms such as websites, mobile applications, social media, and smart devices; the unit collects data on user engagement, browsing patterns, time spent, and content preferences to enable targeted, personalized advertising; ◯ a prediction module that predicts preferences and interests of a user, operatively connected to the user interaction data collection unit, wherein the prediction module uses machine learning models such as deep learning, recurrent neural networks (RNNs) and transformer-based architectures to predict interests of users based on historical interactions and inferred preferences; ◯ an emotion and sentiment analysis unit that assesses the user's mood in real time through computer vision, natural language processing (NLP) and voice analysis, whereby the analysis of facial expressions, voice pitch and linguistic mood is used to determine emotional states and receptivity to advertising content; ◯ an embodiment of a context awareness component in operational communication with the emotion analysis and mood unit, in some cases further augmented by various environmental and situational data such as the device type and its physical location, date and time, and the content processed in the device, processing methods, etc., in establishing adaptability and automatic ad placement to be as relevant as possible to the user and their status as prescribed; ◯ Use reinforcement learning algorithms and generative AI models to drive advertising with personalization engines. Creative elements, messaging, and presentations are dynamically adjusted in real time based on user responses to ensure advertising is personalized and always optimized for best performance; o a privacy-focused AI system with federated learning, differential privacy methods, and on-device AI processing to reduce targeted advertising while complying with international data protection laws such as the General Data Protection Regulation (GDPR) and California consumer privacy laws; o an operationally adapted ad delivery mechanism to engage with real-time bidding (RTB) networks, programmatic advertising exchanges and demand-side platforms (DSPs) and place advertisements through digital advertising networks, which guarantees the delivery of tailored advertising to the most relevant audience in real time; and o a contextual feedback loop in which the machine learning models used in the preference and interest prediction module and in advertising personalization The engine is continuously updated to reflect the latest user engagement data, improving personalization over time and optimizing advertising performance.

The embodiment of the invention provides a fraud phone real-time recognition method and device based on AI semantic understanding, and the method and device achieve the precise understanding of the dialogue content through the innovative construction of a voice analysis mechanism, the grammatical feature extraction and the semantic role marking. And designing a scene discrimination model based on verbal skill recognition, and establishing a fraud verbal skill recognition strategy for intelligent classification in combination with semantic pattern matching and hierarchical clustering algorithms. A residual fusion assessment mechanism is introduced, and accurate assessment and timely prevention and control of call risks are realized through historical case feature fusion and risk scoring. According to the method, the defects of the traditional technology in the aspects of speech understanding, verbal skill recognition, risk assessment and the like are effectively overcome, and the accuracy and reliability of fraud phone recognition are remarkably improved.

The invention discloses a customer voice analysis system based on a large model, and the system comprises a voice data collection and preprocessing module which is used for collecting and preprocessing customer voice data; the voice recognition module is used for recognizing the preprocessed customer voice data; the feature fusion module is used for extracting semantic features of the text data and fusing customer voice data features to generate voice text representation; the large model fine tuning module is used for performing fine tuning on the pre-trained starry fire cognition large model; the semantic analysis and emotion recognition module is used for processing the voice text representation, extracting semantic information, recognizing an emotion state and generating an analysis result; and the dialogue context management module is used for associating customer historical voice data, establishing dialogue context information and updating a voice analysis result. According to the method, the stellate fire cognitive large model and the improved quantitative low-rank adaptation are utilized, and the method has the advantages of high precision, low calculation cost and high intelligent adaptability.

The embodiment of the invention provides a smart city rule engine system and method based on a large model. The system is applied to the technical field of Internet of Things, comprises a data acquisition receiving port, a character and voice analysis module, an intention instruction recognition module, a rule chain module and a smart city service driving module, and comprises the steps of performing intention recognition processing on character and voice data of a client to obtain an intention recognition result; performing analysis processing on the intention recognition result to obtain an instruction analysis list; an instruction set is combined based on the intention recognition result and the instruction analysis list, and a rule chain conforming to rule engine logic description is generated through the instruction set; and automatically executing the service scene based on the rule chain and the service data. In this way, through the large model technology, the rule chain is automatically generated and optimized, and the service scene is automatically processed through the rule engine, and compared with a traditional rule engine system, the system has higher intellectualization, automation and expansibility.

This application relates to a human-computer voice interaction control method and system based on a smart TV. The method includes acquiring voice data within a preset range, processing the voice data to obtain voice feature data carrying control commands; performing feature analysis on the voice feature data using a preset voice analysis model to extract wake-up keywords and compare them with a preset command library to obtain control command comparison results; sending a secondary confirmation request to the user based on the control command comparison results, and combining the confirmation voice information from the user feedback to perform command recognition evaluation and correct command deviation processing to obtain the corrected control command; performing function switching processing on the TV according to the correct control command, and optimizing the display effect by linking and adjusting related devices based on the program display requirements after the switch to obtain human-computer voice interaction control data. This application has the effect of improving the intelligence of voice interaction control of smart TVs.

The invention relates to the field of work satisfaction evaluation, and discloses an express industry work satisfaction evaluation method based on a productivity model, and the method comprises the following steps: collecting the work data of a courier, including the workload, the work efficiency and the customer evaluation; collecting emotional data of the courier, wherein the emotional data comprises emotional fluctuations and emotional labels obtained through voice analysis and text analysis; constructing a productivity evaluation model based on the collected work data, and evaluating the work productivity of the courier; constructing an emotion calculation model based on the collected emotion data, and evaluating the emotional state of the courier; calculating the work satisfaction degree of the courier based on the results of the productivity evaluation and the emotional state evaluation; and based on a work satisfaction evaluation result, constructing an optimization algorithm to dynamically adjust the work task and the work time. Through a dynamic evaluation method based on a productivity model, in combination with a multi-dimensional working state score and a real-time feedback mechanism, the accuracy of work satisfaction evaluation is effectively improved.

Systems and methods detect and prevent vishing attacks through an integrated framework combining SIP header customization, STIR / SHAKEN frameworks, AI / ML analysis, and real-time speech analysis using the Viterbi algorithm. The system begins with call initiation, embedding authentication information in the SIP header. The SIP data is transmitted and verified using STIR / SHAKEN frameworks, ensuring the authenticity of the caller's identity. Verified data is cross-referenced with third-party databases and analyzed by an AI / ML engine to detect anomalies. If potential fraud is detected, the call is blocked, and the customer is notified. Calls that pass initial checks are further analyzed using the Viterbi algorithm, which converts speech to text and identifies suspicious patterns. An anomaly pattern detector processes the converted text to detect vishing indicators, terminating the call if a match is found. This multi-layered approach ensures robust protection against vishing, enhancing the security and reliability of voice communications while safeguarding users from fraud.

The embodiment of the present application discloses a conversation processing method, device, terminal and storage medium based on emotion recognition. In the process of communication between an agent and a customer based on a business scenario, this solution collects the customer's first voice information and the agent's second voice information, identifies the customer's emotion based on the first voice information, and if the customer's emotion is negative, generates a conversation voice based on the first voice information and the second voice information according to the time point when the customer's negative emotion appears, then performs voice analysis on the conversation voice to obtain a voice analysis result, and performs text analysis on the conversation voice to obtain a text analysis result, and generates customer soothing strategy information based on the voice analysis result and the text analysis result, so that the agent can soothe the customer's emotion according to the customer soothing strategy information, thereby improving the customer's business experience.

The invention discloses a voiceprint recognition and identity authentication method and system based on public safety management and a medium, and relates to the field of voiceprint recognition, and the method comprises the following steps: S1, carrying out the preprocessing of an original voice signal, extracting three types of features in parallel, and splicing the three types of features into a joint feature for representation; s2, coding the joint feature representation, outputting a high-level shared feature vector, and generating a voiceprint identification vector and a living body attribute vector through branches; s3, when high-security-level authentication is judged, dynamic challenge content is sent, response voice is received, and instant behavior characteristics are analyzed; and S4, calculating voiceprint matching similarity and living body attribute probability, fusing the voiceprint matching similarity, the living body attribute probability and the selectable dynamic evidence score, and generating a final authentication decision. By introducing the joint feature coding method based on the correlation constraint, the deep discrimination information can be better reserved, the defense capability on advanced speech synthesis and conversion attacks is enhanced, and the overall security level of the system is improved.

The invention relates to the technical field of intelligent outbound calls, in particular to an intelligent outbound call processing method and device based on artificial intelligence and a server. Real-time voice data input by a user is converted into text data based on a voice analysis model, intention analysis is conducted on the text data based on an intention analysis model, and the intention of the user is obtained. And performing sentiment analysis on the text data based on the sentiment analysis model to obtain a sentiment analysis result of the user, generating a dialogue strategy matched with the user according to a user intention analysis result and the sentiment analysis result, and finally executing dynamic adjustment operation according to the dialogue strategy and dialogue skill data to obtain verbal skill data matched with the user. According to the method, the accurate dialogue strategy is generated for the user not only through the intention of the user but also in combination with the sentiment analysis result of the user, so that the verbal skill data to be provided for the user is adjusted, the determination accuracy and precision of the verbal skill data needed by the user are improved, and the interaction accuracy of the intelligent outbound dialogue process is improved.

There is provided a method for executing a session of a video game, comprising the operations of: recording voice of a player participating in game progress of the session of the video game; analyzing a game state generated by execution of the session of the video game, wherein analyzing the game state identifies a scenario in which the game is performed; analyzing the recorded speech using the recognized scenario of the game and the speech recognition model to recognize text content of the recorded speech; the identified textual content is applied as a game input of a session of the video game.

A computer-implemented method for determining whether to introduce a delay in an audio stream from a particular speaker includes an audio stream from a transmitting device. The method further includes providing the audio stream and the speech analysis score, information about the one or more sound emotion parameters, and one or more sound emotion scores of a first user associated with the transmitting device as inputs to a trained machine learning model, the trained machine learning model is iteratively applied to the audio stream, and each iteration corresponds to a respective portion of the audio stream. The method further includes generating, as an output, a degree of detrimental of the audio stream using the trained machine learning model. The method further includes transmitting the audio stream to a receiving device, where the transmitting is performed to introduce a time delay in the audio stream based on the degree of detrimental.

An AI-powered automated recruitment system for intelligent talent identification and selection, consisting of: a processing unit configured to execute machine learning procedures for adaptive candidate evaluation; a storage unit that is operationally connected to the aforementioned processing unit and is configured to store candidate data, trained model parameters, and records of the recruitment process; a video interaction unit consisting of a camera and an audio recording unit configured for conducting real-time or asynchronous interviews, wherein the processing unit analyzes the captured audiovisual data using natural language processing, facial recognition and sentiment analysis to generate behavioral assessment metrics; a speech analysis unit consisting of a speech recognition processor and an acoustic characterization subsystem configured to receive and analyze telephone candidate responses, with the processing unit deriving linguistic competence, tonal clarity, and confidence indices from the responses; a technical evaluation unit configured to deliver interactive programming tasks to candidate terminals, with the processing unit evaluating code execution output and logic efficiency and detecting plagiarism through code similarity comparison techniques; a data analysis unit configured to combine the outputs of the aforementioned video interaction unit, speech analysis unit, and technical evaluation unit to generate predictive suitability scores for candidates and recruitment performance metrics; and a communication interface unit configured to establish secure data transmission connections between the system and external personnel databases via encrypted network protocols.

An electronic processor that may be configured to perform a first voiceprint analysis of first audio data of a first call from a first caller to determine a first voiceprint of the first caller, and determine that the first voiceprint matches a first previously-stored voiceprint. The first previously-stored voiceprint may identify and be associated with first previously-stored information regarding a previous call from the first caller that is anonymized so as not to include a name of the first caller or a phone number of the first caller. The electronic processor may also receive a permission command based on a user input from the first caller to indicate that a call taker handling the first call is allowed to have access to the first previously-stored information. The electronic processor may output the first previously-stored information on a call taker device in response to receiving the permission command.

The system according to this embodiment aims to support workplace communication and employee mental health in real time. [Solution] The system according to the embodiment comprises an acquisition unit, an analysis unit, a decision unit, and a proposal unit. The acquisition unit acquires images and voices of employees. The analysis unit analyzes the employee's facial expressions, tone of voice, and language use based on the information acquired by the acquisition unit. The decision unit determines an action in accordance with the analysis results obtained by the analysis unit. The proposal unit proposes the action determined by the decision unit to the employee.

A method and non-transitory computer-readable medium for automated cognitive assessment and generation of cognitive assessment scores from patient oral cognitive assessment data using an artificial-intelligence ("AI") agent is provided. The method and computer-readable medium include delivering one or more oral cognitive assessments and storing collected audio data for analysis by a backend analysis server. Collected cognitive assessment data is processed by a backend analysis server using a feature-embedding model to extract derived features from collected audio data. The backend analysis server uses machine-learning classification to generate cognitive assessment scores from the features. Assessment scores may be transmitted to a clinician-accessible device for review. The method enables automated, scalable, and reproducible patient cognitive assessments on patient-accessible devices.

An input audio signal is received from a client device. A connection to a communication session with a plurality of participants is maintained for the client device. An audio feed corresponding to the input audio signal is transmitted to a plurality of other participant devices. While the communication session is in progress and the connection to the communication session is maintained by the client device, the following are periodically performed. An absence of audible speech originating at the client device within the input audio signal is detected. The audio feed transmitted to the plurality of other participant devices is muted. The input audio signal is analyzed for audible speech. A presence of audible speech within the input audio signal is detected. The audio feed transmitted to the plurality of other participant devices is unmuted.

A method and non-transitory computer-readable medium for generating cognitive assessment scores from patient speech using an artificial-intelligence (“AI”) agent is provided. The method and computer-readable medium include initiating a patient audio communication session, conducting one or more oral cognitive assessments including narrative recall, engaging in a context-dependent discussion phase, assessing responses, and generating scores from audiometric, lexical, prosodic, and semantic-coherence features using multi-modal machine learning architectures. Audio data is processed by a backend analysis server using automatic speech recognition to generate a time-aligned transcript, and the backend analysis server uses a feature-embedding model to extract prosodic, acoustic, lexical, and semantic-coherence features from the audio and corresponding transcript. The backend analysis server uses machine-learning classification to generate cognitive assessment scores from extracted features. Assessment scores, feature data, and session records may be transmitted to clinician-accessible locations for review. The method enables automated, scalable, and reproducible speech-based patient cognitive assessments on patient-accessible devices.

To provide a system which accurately reads emotions from utterances of a baby and suggests appropriate support to a parent in order to suggest an appropriate support to the parent who are anxious about communications with the baby.SOLUTION: In a data processing system provided with a data processing unit and a smart device, specification processing by a specification processing part of the data processing unit includes: receiving means receiving a baby's utterance; analysis means analyzing the received utterance by a speech recognition technology; discriminating means for discriminating a baby's emotions on the basis of a result of an analysis; and suggestion means suggesting an appropriate support to a parent in accordance with discriminated emotions.SELECTED DRAWING: Figure 2

The invention discloses a subtitle rendering method, system and device based on the AI technology and a storage medium, and the method comprises the steps: obtaining a to-be-rendered target file, and carrying out the voice analysis processing of the target file through employing the AI technology, and obtaining corresponding text information; obtaining statement core components and keywords in the text information so as to generate structured subtitle metadata; performing paging layout processing on the structured subtitle metadata by using preset layout parameters to obtain corresponding page data; on the basis of the page data, word-by-word rendering is carried out on structured subtitle metadata in the page data in combination with a timestamp; and according to a set display mode, performing paging rendering on the page data, and outputting and displaying the page data. According to the method, the target file is analyzed and processed through the AI voice, word-by-word rendering is carried out in combination with the layout parameters and the timestamps, and finally paging rendering output is completed in the set display mode, so that the subtitle rendering effect can be remarkably improved, and user requirements are met.

The invention discloses an outbound call quality monitoring and optimizing system based on big data, and belongs to the technical field of outbound call quality optimization. Comprising a call-out voice acquisition unit, a voice grouping unit, a call-out quality evaluation unit, an optimized voice library generation unit and a call-out optimization unit, the call-out voice acquisition unit acquires all call-out voices in a time period from the current to a certain previous time node to obtain a voice data array # imgabs0 #, the voice grouping unit is connected to a voice analysis module, and the call-out quality evaluation unit is connected to the call-out optimization unit. Each element in the voice data array # imgabs1 # is grouped, and the outbound quality evaluation unit communicates with the voice grouping unit; the invention discloses an outbound call quality monitoring and optimizing system based on big data, the outbound call quality monitoring and optimizing system comprises an outbound call quality monitoring and optimizing unit, the outbound call quality monitoring and optimizing unit obtains an evaluation intersection # imgabs2, counts the number of # imgabs3 # and then calculates the outbound call quality ratio # imgabs5 #, and the optimization voice library generating unit conducts intersection calculation on # imgabs6 #, # imgabs7 # and # imgabs8 # to obtain an intersection # imgabs9 #. The outbound call quality monitoring and optimizing system based on the big data can monitor the outbound call quality in real time, optimizes outbound call behaviors according to the outbound call quality, and improves the outbound call quality.

The application embodiment provides a fraud call real-time identification method and device based on AI semantic understanding, innovatively constructs a voice analysis mechanism, realizes accurate understanding of conversation content through grammar feature extraction and semantic role labeling, designs a scene discrimination model based on dialogue identification, combines semantic pattern matching and hierarchical clustering algorithm, establishes a fraud dialogue identification strategy for intelligent classification, introduces a residual fusion evaluation mechanism, realizes accurate evaluation and timely prevention and control of call risk through historical case feature fusion and risk scoring. The method effectively solves the shortcomings of traditional technologies in voice understanding, dialogue identification and risk assessment, and significantly improves the accuracy and reliability of fraud call identification.

This enables the use of the VOX function on multiple VOX communication devices. [Solution] The system comprises an extraction unit 22 that extracts voice data from received data, an analysis unit 23 that analyzes the voice data extracted by the extraction unit 22 as a voice analysis result, a calculation unit 26 that calculates priority information regarding the priority of the VOX function of the communication device based on the voice analysis result analyzed by the analysis unit 23, and a control unit 28 that determines whether to enable or disable the VOX function based on the priority information calculated by the calculation unit 26.

To solve the problem of inquiry processing in a call center.SOLUTION: An inquiry processing system in a call center includes voice analysis means, means of utilizing a generative AI, and summary output means.SELECTED DRAWING: Figure 1

System for assigning one of a plurality of operators to an interaction amongst one or more interactions each between a respective user and a user assistance centre comprising an artificial operator implemented on a computer, wherein each interaction comprises a real-time interaction between a user and the user assistance centre, the system comprising a controller configured to: obtain, during an information-collection phase of an interaction between a respective user and the user assistance centre, first information concerning a matter relating to the respective user; and determine an operator suitable to resolve the matter, based on said first information and on second information output by an AI-based operator assignment unit, the AI-based operator assignment unit being trained based on a dataset including at least the following information associated to each other: (i) past transcript information representing a transcript of an earlier interaction between a user and the user assistance centre; (ii) speech analysis information comprising parameters output from a speech analysis for the earlier interaction; and (iii) operator information identifying an operator assigned to said earlier interaction.

We provide the system. [Solution] A means for obtaining call content, electronic messages, and visitor images from a user's mobile device, A method using generative AI to analyze acquired information and detect the risk of fraud, Image recognition and voice analysis means for recording and identifying the audio and video of visitors, A means of alerting and notifying users or designated contacts when suspicious activity or visitors are identified, A means of exchanging information with users and providing instructions, A system that includes this.

The invention introduces the concept of subjecting wirelessly transmitted and received official warning messages to speech analysis in order to identify a scenario and to carry out an action by the vehicle according to the scenario.

The invention provides an intelligent call forwarding method, device and equipment and a readable storage medium, and relates to the technical field of call forwarding, and the method comprises the steps: sending a first response voice to a call object after a call is connected; receiving a first response voice of the calling object for the first response voice; analyzing the first reply voice to obtain a voice analysis result; if the voice analysis result meets the configured switching rule, respectively converting the first response voice and the first reply voice into a first response text and a first reply text; and transferring the call with the calling object to the artificial customer service, and sending the identity information, the historical call data, the to-be-mediated item data, the first response text and the first reply text to the artificial customer service. According to the invention, the intelligent and automatic degree of intelligent call forwarding is improved, and the mediation efficiency is also improved.

A system is provided.SOLUTION: A system comprising: data collection means for collecting facial data of a customer; data collection means for collecting voice data of the customer; facial expression recognition means for recognizing facial expressions of the customer from the collected facial data; voice analysis means for analyzing a tone of the customer from the collected voice data; emotion analysis means for determining an emotional state of the customer based on data obtained by the facial expression recognition means and the voice analysis means; and feedback generation means for suggesting a service improvement based on the determined emotional state.SELECTED DRAWING: Figure 1

The invention provides a pipe network engineering field expert experience guiding method based on voice recognition, and the collection and sharing efficiency of construction experience is remarkably improved through a voice recognition technology and an intelligent guiding mechanism. Considering specific implementation thoughts and related technologies of expert construction experience guidance, an expert construction experience intelligent guidance framework based on voice is provided. The implicit experience of experts is automatically converted into dominant knowledge, and high-precision speech analysis is realized through a lightweight speech recognition model. By means of transfer learning and knowledge distillation technologies, the speech recognition model still maintains excellent performance while reducing computing resource consumption. According to the automatic knowledge matching mechanism, the conformity of expert experience and construction specifications is further ensured, and standardization and systematization of construction experience are helped to be achieved. By intelligently generating the guiding problem, the experience of experts can be guided more deeply and comprehensively, efficient sharing and application of construction knowledge are promoted, and the efficiency and quality of enterprises in engineering management are greatly improved.