Abstract: Systems and methods for establishing user persona from audio interactions are disclosed, including a voice-based conversational AI platform having an acoustic analytical record engine and audio signal codification optimizer. The engine receives an audio sample indicative of voice conversation between an end user and a bot and transforms it into quantifiable and machine-ingestible power spectrum and acoustic indicators that uniquely represent the audio sample in the form of a feature vector. The optimizer ingests and processes the indicators to estimate likelihood of an attribute value representing the audio sample by constructing a convolutional neural network model for each attribute category. The optimizer establishes user persona attribute values across different attribute categories for the audio sample based on the estimated likelihood.

Abstract: Examples of a process optimization system are provided. The system may obtain a query from a user and process data associated with the query from a plurality of data sources. The system may identify a plurality of factors and a target variable from the process data. The system may identify a factor range for the plurality of factors. The system may deconstruct the factor range to identify a plurality of data partitions. The system may identify a data pruning activator based on the plurality of data partitions and a preponderant data partition therefrom. The system may identify a plurality of clusters associated with the preponderant data partition. The system may identify a preponderant cluster from the plurality of clusters. The system may identify a confidence score associated with the preponderant cluster. The system may generate a process optimization result based on the preponderant cluster and the confidence score.

Abstract: An utterance recommendation system may obtain a plurality of utterance influencing attributes influencing a response recommended for a user query from a plurality of sources. The system may collate the utterance influencing attributes to provide enriched user data and may identify a set of preconfigured potential utterance options associated with the user query. The system may implement a plurality of utterance recommendation techniques to analyze the enriched user data based on associated predefined rules, and provide a preferential rank ordering of the preconfigured potential utterance options to be recommended based on the analysis. The system may determine a single optimized rank ordering of the preconfigured potential utterance options representing a ranking order of the preconfigured potential utterance options. The system may provide the response to the user query, based on a potential utterance option ranked highest in the single optimized rank ordering.

Abstract: Systems and methods for establishing data augmentation and enrichment within a data repository are disclosed, including a plug-in including an information relation charting engine, an analytical metric study engine, a derived record construction engine, and a configurational layer. The information relation charting engine receives raw data metrics from a production environment, determines a correspondence between field elements associated with each of the source and target schema, and correspondingly generates a mapping. The analytical metric study engine performs analytical analysis on the field elements present in the generated mapping and uses machine learning and previous data log to generate recommendations for the database. The derived record construction engine involves a deep learning model to generate relevant and usable data metrics based on the generated recommendations, which is then mapped to the relational database to generate augmented and enriched database.

Abstract: Systems and methods for establishing user persona from audio interactions are disclosed, including a voice-based conversational AI platform having an acoustic analytical record engine and audio signal codification optimizer. The engine receives an audio sample indicative of voice conversation between an end user and a bot and transforms it into quantifiable and machine-ingestible power spectrum and acoustic indicators that uniquely represent the audio sample in the form of a feature vector. The optimizer ingests and processes the indicators to estimate likelihood of an attribute value representing the audio sample by constructing a convolutional neural network model for each attribute category. The optimizer establishes user persona attribute values across different attribute categories for the audio sample based on the estimated likelihood.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media supporting automated document redaction in compliance with data privacy requirements. To facilitate data redaction, a document template may be subjected to a template expansion process that generates multiple instances of the template, each instance having data fields populated with arbitrary data (e.g., data that is not subject to the data privacy requirements). The populated templates may then be converted to synthetic data that includes information about the location of the data contents within the populated templates, a copy of the data contents, and structural information. A set of content features and structural features may be generated based on the synthetic data and used to train a model to identify data that should be redacted within a document. Once trained, documents may be evaluated using the model to perform data redaction prior to providing access to the documents.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support automated summarization of documents using machine learning and artificial intelligence. The summaries include document category-specific summaries for different categories of documents. To illustrate, a document processing device may receive input data representing a document and provide first feature data based on the input data to one or more machine learning (ML) models to select a document category of a plurality of predefined document categories. The document processing device may provide second feature data based on the input data to a second set of ML models associated with the selected document category to generate annotation data associated with the document. The document processing device may provide the annotation data to a third set of ML models to generate a summary of the document, and the document processing device may generate an output based on the summary.

Abstract: Examples of a process optimization system are provided. The system may obtain a query from a user and process data associated with the query from a plurality of data sources. The system may identify a plurality of factors and a target variable from the process data. The system may identify a factor range for the plurality of factors. The system may deconstruct the factor range to identify a plurality of data partitions. The system may identify a data pruning activator based on the plurality of data partitions and a preponderant data partition therefrom. The system may identify a plurality of clusters associated with the preponderant data partition. The system may identify a preponderant cluster from the plurality of clusters. The system may identify a confidence score associated with the preponderant cluster. The system may generate a process optimization result based on the preponderant cluster and the confidence score.

Abstract: Examples of network bandwidth management are provided. A system may access, time series distribution data including a distribution of data points corresponding to an average count of network packets. Further, the system may determine, from the time series distribution data, trend data component, seasonal data component, cyclical data component, and irregular data component. Furthermore, the system may compute, baseline distribution traffic data based on the trend data component, the cyclical data component, and the seasonal data component. Furthermore, the system may identify auto regression data component and moving average data component based on the irregular data component. The system may further, determine an estimated network traffic for a future time interval based on the baseline distribution traffic data, a lagged covariate factor, and covariate metrics. Furthermore, the system may provide the estimated bandwidth based on the estimated network traffic to configure the server.

Abstract: Examples of a process optimization system are provided. The system may obtain a query from a user and process data associated with the query from a plurality of data sources. The system may identify a plurality of factors and a target variable from the process data. The system may identify a factor range for the plurality of factors. The system may deconstruct the factor range to identify a plurality of data partitions. The system may identify a data pruning activator based on the plurality of data partitions and a preponderant data partition therefrom. The system may identify a plurality of clusters associated with the preponderant data partition. The system may identify a preponderant cluster from the plurality of clusters. The system may identify a confidence score associated with the preponderant cluster. The system may generate a process optimization result based on the preponderant cluster and the confidence score.

Abstract: An utterance recommendation system may obtain a plurality of utterance influencing attributes influencing a response recommended for a user query from a plurality of sources. The system may collate the utterance influencing attributes to provide enriched user data and may identify a set of preconfigured potential utterance options associated with the user query. The system may implement a plurality of utterance recommendation techniques to analyze the enriched user data based on associated predefined rules, and provide a preferential rank ordering of the preconfigured potential utterance options to be recommended based on the analysis. The system may determine a single optimized rank ordering of the preconfigured potential utterance options representing a ranking order of the preconfigured potential utterance options. The system may provide the response to the user query, based on a potential utterance option ranked highest in the single optimized rank ordering.

Abstract: Examples of network bandwidth management are provided. A system may access, time series distribution data including a distribution of data points corresponding to an average count of network packets. Further, the system may determine, from the time series distribution data, trend data component, seasonal data component, cyclical data component, and irregular data component. Furthermore, the system may compute, baseline distribution traffic data based on the trend data component, the cyclical data component, and the seasonal data component. Furthermore, the system may identify auto regression data component and moving average data component based on the irregular data component. The system may further, determine an estimated network traffic for a future time interval based on the baseline distribution traffic data, a lagged covariate factor, and covariate metrics. Furthermore, the system may provide the estimated bandwidth based on the estimated network traffic to configure the server.

Abstract: Examples of a process optimization system are provided. The system may obtain a query from a user and process data associated with the query from a plurality of data sources. The system may identify a plurality of factors and a target variable from the process data. The system may identify a factor range for the plurality of factors. The system may deconstruct the factor range to identify a plurality of data partitions. The system may identify a data pruning activator based on the plurality of data partitions and a preponderant data partition therefrom. The system may identify a plurality of clusters associated with the preponderant data partition. The system may identify a preponderant cluster from the plurality of clusters. The system may identify a confidence score associated with the preponderant cluster. The system may generate a process optimization result based on the preponderant cluster and the confidence score.

Abstract: In some implementations, a device may receive unstructured interaction data identifying an interaction of a user with a user device. The device may receive historical unstructured interaction data identifying historical interactions of users and historical unstructured resolution data identifying historical resolutions to the historical interactions. The device may process the historical unstructured interaction data and the historical unstructured resolution data to determine historical structured interaction data and historical structured resolution data. The device may process the unstructured interaction data and the historical structured interaction data to determine pretext identifiers for the interaction of the user. The device may process the pretext identifiers and the historical structured resolution data to generate a resolution network identifying possible resolutions to the interaction of the user.

Abstract: A wireless signal strength monitoring system employs a wireless signal strength driver model to monitor and maintain wireless signal strengths for mobile devices on a wireless communication network. The monitoring system parses activity logs of network entities including the mobile devices on the wireless communication network to generate a plurality of multilevel token sets. A selected one of the plurality of multilevel token sets is used to generate a wireless signal strength driver model along with topics from clusters of homogenous logs of the activity logs. The drivers of the wireless signal strength are identified from the wireless signal strength driver model. The drivers are used to monitor and improve the wireless signal strengths for the mobile devices on the wireless communication network.

Abstract: A wireless signal strength monitoring system employs a wireless signal strength driver model to monitor and maintain wireless signal strengths for mobile devices on a wireless communication network. The monitoring system parses activity logs of network entities including the mobile devices on the wireless communication network to generate a plurality of multilevel token sets. A selected one of the plurality of multilevel token sets is used to generate a wireless signal strength driver model along with topics from clusters of homogenous logs of the activity logs. The drivers of the wireless signal strength are identified from the wireless signal strength driver model. The drivers are used to monitor and improve the wireless signal strengths for the mobile devices on the wireless communication network.