Abstract: Various embodiments of the present disclosure provide a query handling process that improves the functionality of a computer in various aspects. The techniques comprise receiving a data request that comprises a target data input and a target contextual attribute for the target data input. The techniques comprise receiving a subset of candidate reference entries, receiving a genericism metric for the candidate reference entries, and generating input vectors for the candidate reference entries by concatenating a similarity metric, a genericism metric, a target contextual attribute, and an entry contextual attribute for up to each of the candidate reference entries. The technique comprises generating, using a machine learned classifier, match probability scores for the candidate reference entries based on the input vector, and, in response to a determination that a match probability score meets or exceeds a similarity threshold, providing the candidate reference entry in response to the data request.

Abstract: Various embodiments of the present disclosure provide a feedback integrated retrieval augmented generation (RAG) framework that improves the functionality of a computer in various aspects. The techniques comprise receiving a query comprising a request content for a generative model and generating, using an embedding model, a query embedding for the query by embedding the request content. The techniques comprise receiving, from a RAG datastore, a content chunk based on the query embedding and a content embedding of the content chunk, wherein the content chunk is assigned a preprocessing tag. The techniques comprise generating a generative prompt for the query and that comprises (i) prompt instructions, (ii) the request content, and (iii) the content chunk and generating, using the generative model, a generative output based on the generative prompt. The techniques comprise routing the generative output based on the preprocessing tag.

Abstract: Various embodiments of the present disclosure provide a feedback integrated retrieval augmented generation (RAG) framework that improves the functionality of a computer in various aspects. The techniques comprise receiving a query comprising a request content for a generative model and generating, using an embedding model, a query embedding for the query by embedding the request content. The techniques comprise receiving, from a RAG datastore, a content chunk based on the query embedding and a content embedding of the content chunk, wherein the content chunk is assigned a preprocessing tag. The techniques comprise generating a generative prompt for the query and that comprises (i) prompt instructions, (ii) the request content, and (iii) the content chunk and generating, using the generative model, a generative output based on the generative prompt. The techniques comprise routing the generative output based on the preprocessing tag.

Abstract: Various embodiments of the present disclosure provide a mixed prediction scheme that improves the functionality of a computer in various aspects. The techniques comprise receiving an input sequence comprising a set of input classes and a set of elapsed time indicators respectively corresponding to the set of input classes and generating, using an embedding layer of a mixed model pipeline, a hidden state representation of the input sequence. The techniques comprise inputting the hidden state representation to (i) a connected classification layer of the mixed model pipeline to receive a class prediction based on the hidden state representation and (ii) a connected regression layer of the mixed model pipeline to receive a time prediction based on the hidden state representation and providing an aggregate output for the input sequence that identifies the class prediction and the time prediction.

Abstract: Various embodiments of the present disclosure provide a data filtering technique that improves the functionality of a computer in various aspects. The technique comprises receiving a set of historical quantiles based on a set of historical datasets; receiving an input dataset for a new time period; determining a quality class for a data subset of the input dataset based on (a) a set of input quantiles for the data subset and (b) the set of historical quantiles; iteratively generating a filtered dataset from the input dataset by: filtering a set of data objects from the data subset based on an initial quantile threshold, determining a dataset volume of the initial filtered dataset failing to meet a volume tolerance, and modifying the target quantile based on the quality class and refiltering the data subset; and providing the filtered dataset to a receiving computing entity.

Abstract: Embodiments of the present disclosure provide systems and methods for matching data entities across disparate datasets. One method may include generating a reference numeric vector for a reference data object and generating a plurality of match candidate numeric vectors for a plurality of match candidate data objects. The method may also include identifying a subset of match candidate data objects from the plurality of match candidate data objects and generating a plurality of data features based on a comparison between one or more reference object attributes corresponding to the reference data object and one or more matching object attributes corresponding to the subset of match candidate data objects. The method may also include identifying a match between the reference data object and a match candidate data object of the subset of match candidate data objects based on the plurality of data features.

Abstract: Various embodiments of the present disclosure provide data storage, processing, and prediction techniques for providing predictive insights within large data prediction domains. The techniques may include generating, using a plurality of source tables for a prediction domain, a global graph for the prediction domain. The techniques may include generating, using a graph-based machine learning model, a plurality of node-level weights for the plurality of graph nodes based on a plurality of node attributes corresponding to the plurality of graph nodes. The techniques may include generating, using the graph-based machine learning model, a plurality of semantic-level weights for the plurality of weighted edges based on a designated predictive task for the global graph. The techniques may include generating plurality of graph node embeddings and initiating the performance of the designated predictive task based on the plurality of graph node embeddings.

Abstract: Embodiments of the present disclosure provide systems and methods for matching data entities across disparate datasets. One method may include generating a reference numeric vector for a reference data object and generating a plurality of match candidate numeric vectors for a plurality of match candidate data objects. The method may also include identifying a subset of match candidate data objects from the plurality of match candidate data objects and generating a plurality of data features based on a comparison between one or more reference object attributes corresponding to the reference data object and one or more matching object attributes corresponding to the subset of match candidate data objects. The method may also include identifying a match between the reference data object and a match candidate data object of the subset of match candidate data objects based on the plurality of data features.

Abstract: Various embodiments of the present disclosure provide data storage, processing, and prediction techniques for providing predictive insights within large data prediction domains. The techniques may include generating, using a plurality of source tables for a prediction domain, a global graph for the prediction domain. The techniques may include generating, using a graph-based machine learning model, a plurality of node-level weights for the plurality of graph nodes based on a plurality of node attributes corresponding to the plurality of graph nodes. The techniques may include generating, using the graph-based machine learning model, a plurality of semantic-level weights for the plurality of weighted edges based on a designated predictive task for the global graph. The techniques may include generating plurality of graph node embeddings and initiating the performance of the designated predictive task based on the plurality of graph node embeddings.

Abstract: Various embodiments of the present disclosure provide data storage, processing, and prediction techniques for providing predictive insights within large data prediction domains. The techniques may include generating, using a plurality of source tables for a prediction domain, a plurality of subdomain-specific graphs for the prediction domain. The techniques may include generating a plurality of subdomain-specific embeddings for the plurality of subdomain-specific graphs and a composite graph embedding based on the plurality of graph embeddings and a designated predictive task. The techniques may include initiating the performance of the designated predictive task based on the composite graph embedding.

Abstract: Embodiments of the present disclosure provide systems and methods for matching data entities across disparate datasets. One method may include generating a reference numeric vector for a reference data object and generating a plurality of match candidate numeric vectors for a plurality of match candidate data objects. The method may also include identifying a subset of match candidate data objects from the plurality of match candidate data objects and generating a plurality of data features based on a comparison between one or more reference object attributes corresponding to the reference data object and one or more matching object attributes corresponding to the subset of match candidate data objects. The method may also include identifying a match between the reference data object and a match candidate data object of the subset of match candidate data objects based on the plurality of data features.

Abstract: Various embodiments of the present disclosure provide data storage, processing, and prediction techniques for providing predictive insights within large data prediction domains. The techniques may include generating, using a plurality of source tables for a prediction domain, a global graph for the prediction domain. The techniques may include generating, using a graph-based machine learning model, a plurality of node-level weights for the plurality of graph nodes based on a plurality of node attributes corresponding to the plurality of graph nodes. The techniques may include generating, using the graph-based machine learning model, a plurality of semantic-level weights for the plurality of weighted edges based on a designated predictive task for the global graph. The techniques may include generating plurality of graph node embeddings and initiating the performance of the designated predictive task based on the plurality of graph node embeddings.