Abstract: A method, apparatus, non-transitory computer readable medium, and system for obtaining a query comprising a document request. A language generation model is configured to retrieve a first document based on the query and generate an expanded query based on the first document using a knowledge graph that connects the first document to a second document. A third document is retrieved based on the expanded query in response to the query.

Abstract: A search system facilitates efficient and fast near neighbor search given item vector representations of items, regardless of item type or corpus size. To index an item, the search system expands an item vector for the item to generate an expanded item vector and selects elements of the expanded item vector. The item is index by storing an identifier of the item in posting lists of an index corresponding to the position of each selected element in the expanded item vector. When a query is received, a query vector for the item is expanded to generate an expanded query vector, and elements of the expanded query vector are selected. Candidate items are identified based on posting lists corresponding to the position of each selected element in the expand query vector. The candidate items may be ranked, and a result set is returned as a response to the query.

Abstract: A method, apparatus, non-transitory computer readable medium, and system for data processing include obtaining a query relating to a document and identifying metadata for the document based on the query, where the metadata describes a structure including a plurality of portions of the document. Some embodiments including generating, using a machine learning model, a retrieval command based on the query and the metadata, selectively retrieving at least one of the plurality of portions of the document based on the retrieval command, and generating, using the machine learning model, a response to the query based on the at least one of the plurality of portions of the document.

Abstract: In various examples, direct and indirect feedback is obtained and used to update a machine learning model. For example, feedback indicating interactions with the machine learning model are obtained from various entities. Continuing this example, the feedback is used to determine a set of scores associated with a particular response generated by the machine learning model. In various embodiments, the set of scores includes a response score, a multi-turn score, and a session score. Furthermore, the set of scores, in this examples, are combined to generate a single score associated with the response that is then used to update the machine learning model.

Abstract: Techniques for data question answering with auxiliary recommendations are described to enable efficient querying of data sets for answers to data questions based on a natural language input. In an example, a processing device is operable to receive a natural language input including a query, determine an additional query based on a context of the query, and query a machine-learning model using the query and the additional query. The processing device is further operable to receive, from the machine-learning model, a result including a quantitative answer to the query, an additional answer based on the additional query, and an explanation by the machine-learning model of how the machine-learning model generated the quantitative answer or the additional answer in response to the querying. The processing device is operable to present the result for display in a user interface.

Abstract: A method, apparatus, non-transitory computer readable medium, and system for query disambiguation include obtaining a query including an ambiguous element, where the ambiguous element corresponds to an ambiguity category, and selecting a plurality of candidate elements by retrieving the plurality of candidate elements based on the ambiguity category and computing a distance between the ambiguous element and each of the plurality of candidate elements. Some embodiments include generating a plurality of modified queries based on the query by replacing the ambiguous element from the query with each of the plurality of candidate elements, respectively.

Abstract: Techniques for data question answering with auxiliary recommendations are described to enable efficient querying of data sets for answers to data questions based on a natural language input. In an example, a processing device is operable to receive a natural language input including a query, determine an additional query based on a context of the query, and query a machine-learning model using the query and the additional query. The processing device is further operable to receive, from the machine-learning model, a result including a quantitative answer to the query, an additional answer based on the additional query, and an explanation by the machine-learning model of how the machine-learning model generated the quantitative answer or the additional answer in response to the querying. The processing device is operable to present the result for display in a user interface.

Abstract: Embodiments provide systems, methods, and computer storage media for determining string similarity and pattern matching in strings that arrive in a stream. A stream representing string of characters is received and used to compute mapping values that are compared to a mapping value of a query string to identify a match between strings in the stream of characters and the query string. The stream of characters is searched in a single sequential pass to detect a match or the longest matching substring with a query string. An identified match or absence of a match is provided.

Abstract: Techniques for data question answering with auxiliary recommendations are described to enable efficient querying of data sets for answers to data questions based on a natural language input. In an example, a processing device is operable to receive a natural language input including a query, determine an additional query based on a context of the query, and query a machine-learning model using the query and the additional query. The processing device is further operable to receive, from the machine-learning model, a result including a quantitative answer to the query, an additional answer based on the additional query, and an explanation by the machine-learning model of how the machine-learning model generated the quantitative answer or the additional answer in response to the querying. The processing device is operable to present the result for display in a user interface.

Abstract: Change-of-thought machine-learning model debiasing techniques and systems are described. A query is received and context data is produced based on the query, e.g., from an external source. A prompt is generated that includes the context data, the query, and a chain-of-though prompt, which is processed by a machine-learning model. A candidate result based on processing of the prompt using the machine-learning model. The candidate result includes a candidate answer and a chain-of-thought result describing reasoning indicated by the machine-learning model as used in generating the candidate answer.

Abstract: A method, apparatus, non-transitory computer readable medium, and system for data processing include obtaining a query relating to a document and identifying metadata for the document based on the query, where the metadata describes a structure including a plurality of portions of the document. Some embodiments including generating, using a machine learning model, a retrieval command based on the query and the metadata, selectively retrieving at least one of the plurality of portions of the document based on the retrieval command, and generating, using the machine learning model, a response to the query based on the at least one of the plurality of portions of the document.

Abstract: Systems and methods for training a neural network are described. One or more embodiments of the present disclosure include training a neural network based on a first combined gradient of a loss function at a plurality of sampled elements of a dataset; receiving an insertion request that indicates an insertion element to be added to the dataset, or a deletion request that indicates a deletion element to be removed from the dataset, wherein the deletion element is one of the plurality of sampled elements; computing a second combined gradient of the loss function by adding the insertion element to the dataset or by replacing the deletion element with a replacement element from the dataset; determining whether the first combined gradient and the second combined gradient satisfy a stochastic condition; and retraining the neural network to obtain a modified neural network based on the determination.

Abstract: In implementations of systems for generating node embeddings for multiple roles, a computing device implements an embeddings system to cluster nodes of a graph into clusters. An initial role membership vector is computed for each of the nodes based on the clusters. The embeddings system generates a first set of role embeddings for a particular node of the nodes based on the initial role membership vector for the particular node and nodes connected to the particular node in the graph. The embeddings system determines an indication of at least one of a node classification or a link prediction for the graph based on the first set of role embeddings and a second set of role embeddings for an additional node of the nodes.

Abstract: Embodiments described herein provide methods and systems for facilitating actively-learned context modeling. In one embodiment, a subset of data is selected from a training dataset corresponding with an image to be compressed, the subset of data corresponding with a subset of data of pixels of the image. A context model is generated using the selected subset of data. The context model is generally in the form of a decision tree having a set of leaf nodes. Entropy values corresponding with each leaf node of the set of leaf nodes are determined. Each entropy value indicates an extent of diversity of context associated with the corresponding leaf node. Additional data from the training dataset is selected based on the entropy values corresponding with the leaf nodes. The updated subset of data is used to generate an updated context model for use in performing compression of the image.

Abstract: Embodiments provide systems, methods, and computer storage media for management, assessment, navigation, and/or discovery of data based on data quality, consumption, and/or utility metrics. Data may be assessed using attribute-level and/or record-level metrics that quantify data: “quality”—the condition of data (e.g., presence of incorrect or incomplete values), its “consumption”—the tracked usage of data in downstream applications (e.g., utilization of attributes in dashboard widgets or customer segmentation rules), and/or its “utility”?a quantifiable impact resulting from the consumption of data (e.g., revenue or number of visits resulting from marketing campaigns that use particular datasets, storage costs of data). This data assessment may be performed at different stages of a data intake, preparation, and/or modeling lifecycle.

Abstract: Embodiments of the present invention provide systems, methods, and computer storage media for generating and recommending responsive visualizations. In an example embodiment, a design specification of a source visualization and an author's preferences are used to identify and rank compatible sets of candidate responsive transformations (e.g., using answer set programming). Each set is evaluated and ranked according to one or more cost metrics that quantify changes in information density, messaging, or popularity. Some embodiments generate a transformation specification in a declarative grammar that represent the sets of candidate responsive transformations independent of the structure of the source visualization specifications, compile each declarative transformation specification into a rendering grammar specification, and generate a responsive visualization by compiling the rendering grammar specification using a rendering grammar compiler.

Abstract: Methods and systems are provided for generating and using a unified fine-tuning dataset to fine-tune a language model. In embodiments described herein, a first labeled dataset having a first format of human feedback associated with performance of a pre-trained language model is accessed. Additionally, a second labeled dataset having a second format of human feedback associated with performance of the pre-trained language model is accessed. Thereafter, a unified fine-tuning dataset is generated by converting the second labeled dataset to a refined labeled dataset having the first format of human feedback and aggregating the first labeled dataset having the first format of human feedback with the refined labeled dataset having the first format of human feedback. The pre-trained language model is fine-tuned using the unified fine-tuning dataset and output for subsequent utilization.

Abstract: Methods and systems are provided for using a fine-tuned language model to reduce representations of structured data. In embodiments described herein, training data is accessed that includes structured data, a set of queries for the structured data, and each portion of the structured data that is relevant to each query of the set of queries. A language model is fine-tuned to maximize a cumulative reward based on the training data. The cumulative reward includes a reward for determining a correct portion of the structured data, a first penalty for failing to determine the correct portion of the structured data, and a second penalty for determining an incorrect portion of the structured data. The fine-tuned language model is then output.

Abstract: Embodiments provide systems, methods, and computer storage media for management, assessment, navigation, and/or discovery of data based on data quality, consumption, and/or utility metrics. Data may be assessed using attribute-level and/or record-level metrics that quantify data: “quality”—the condition of data (e.g., presence of incorrect or incomplete values), its “consumption”—the tracked usage of data in downstream applications (e.g., utilization of attributes in dashboard widgets or customer segmentation rules), and/or its “utility”—a quantifiable impact resulting from the consumption of data (e.g., revenue or number of visits resulting from marketing campaigns that use particular datasets, storage costs of data). This data assessment may be performed at different stages of a data intake, preparation, and/or modeling lifecycle.

Abstract: Systems and methods for dynamic user profile management are provided. One aspect of the systems and methods includes receiving, by a lookup component, a request for a user profile; computing, by a profile component, a time-to-live (TTL) refresh value for the user profile based on a lookup history of the user profile; updating, by the profile component, a TTL value of the user profile based on the request and the TTL refresh value; storing, by the profile component, the user profile and the updated TTL value in the edge database; and removing, by the edge database, the user profile from the edge database based on the updated TTL value.