Abstract: The present disclosure relates to processing operations configured to uniquely utilize indexing of content to improve content retrieval processing, particularly when working with large data sets. The techniques described herein enables efficient content retrieval when working with large data sets such as those that may be associated with a plurality of tenants of a data storage application/service. Among other technical advantages, the present disclosure is applicable to train a classifier using relevant samples based on text search in tenant-specific scenarios, where accurate searching can be executed for content associated with one or more tenant accounts of an application/service concurrently in milliseconds even in instances where there may be millions of documents to be searched. As an example, exemplary data shards may be generated and managed for efficient and scalable content retrieval processing including training of a classifier (e.g.

Abstract: The present disclosure relates to processing operations configured to uniquely utilize indexing of content to improve content retrieval processing, particularly when working with large data sets. The techniques described herein enables efficient content retrieval when working with large data sets such as those that may be associated with a plurality of tenants of a data storage application/service. Among other technical advantages, the present disclosure is applicable to train a classifier using relevant samples based on text search in tenant-specific scenarios, where accurate searching can be executed for content associated with one or more tenant accounts of an application/service concurrently in milliseconds even in instances where there may be millions of documents to be searched. As an example, exemplary data shards may be generated and managed for efficient and scalable content retrieval processing including training of a classifier (e.g.

Abstract: The present disclosure relates to processing operations configured to uniquely utilize indexing of content to improve content retrieval processing, particularly when working with large data sets. The techniques described herein enables efficient content retrieval when working with large data sets such as those that may be associated with a plurality of tenants of a data storage application/service. Among other technical advantages, the present disclosure is applicable to train a classifier using relevant samples based on text search in tenant-specific scenarios, where accurate searching can be executed for content associated with one or more tenant accounts of an application/service concurrently in milliseconds even in instances where there may be millions of documents to be searched. As an example, exemplary data shards may be generated and managed for efficient and scalable content retrieval processing including training of a classifier (e.g.

Abstract: Described herein is a mechanism for improving the accuracy of a language model interpreting short input utterances. A language model operates in a stateless manner, only ascertaining the intents and/or entities associated with a presented input utterance. To increase the accuracy, two language understanding models are trained. One is trained using only input utterances. The second is trained using input utterance-prior dialog context pairs. The prior dialog context is previous intents and/or entities already determined from the utterances in prior turns of the dialog. When input is received, the language understanding model decides whether the input comprises only an utterance or an utterance and prior dialog context. The appropriate trained machine learning model is selected and the intents and/or entities associated with the input determined by the selected machine learning model.

Abstract: Described herein is a mechanism for improving the accuracy of a language model interpreting short input utterances. A language model operates in a stateless manner, only ascertaining the intents and/or entities associated with a presented input utterance. To increase the accuracy, two language understanding models are trained. One is trained using only input utterances. The second is trained using input utterance-prior dialog context pairs. The prior dialog context is previous intents and/or entities already determined from the utterances in prior turns of the dialog. When input is received, the language understanding model decides whether the input comprises only an utterance or an utterance and prior dialog context. The appropriate trained machine learning model is selected and the intents and/or entities associated with the input determined by the selected machine learning model.

Abstract: A method is provided of enriching an entry for an entity in a local index of a search engine with tags. The method comprises obtaining location-related social media messages from within a neighborhood of an entity; determining from the obtained messages one or more terms that are unique to the entity; individually determining one or more co-occurring terms for the one or more unique terms; and using the one or more co-occurring term as tags to label the entity in the local index. Furthermore, a method is provided of retrieving social media messages associated with search results.

Abstract: A system for retrieving/identifying a document comprising text stored in a document repository is described. A memory stores a graphical structure comprising a first plurality of nodes each representing a person, and a second plurality of nodes each representing a document in the document repository, the nodes being connected by edges according to automatically observed interactions between the represented people and documents. At least some of the nodes have one or more annotations each denoting a topic. A node relatedness calculator computes distances between nodes of the graphical structure using the topic annotations. An input receives an identifier of a user who is represented by one of the first plurality of nodes. An identifier/retriever identifies one or more documents from the document repository by using the identifier and using the computed distances between nodes.

Abstract: Retrieving and/or storing images associated with events is described. For example, streams of event data comprising text are analyzed to detect an event and a language component builds an event language model for the event, comprising a plurality of words. In various examples, images extracted from web or other sources have associated text. In examples, images with associated text that is similar to the event language model are identified as images of the event. In various examples, associations between images and events are used to update an image retrieval system and/or an image storage system. In various examples, query terms about an event are received at an image retrieval system which returns images related to the event on the basis of associations between image text and event language models.

Abstract: A system for retrieving/identifying a document comprising text stored in a document repository is described. A memory stores a graphical structure comprising a first plurality of nodes each representing a person, and a second plurality of nodes each representing a document in the document repository, the nodes being connected by edges according to automatically observed interactions between the represented people and documents. At least some of the nodes have one or more annotations each denoting a topic. A node relatedness calculator computes distances between nodes of the graphical structure using the topic annotations. An input receives an identifier of a user who is represented by one of the first plurality of nodes. An identifier/retriever identifies one or more documents from the document repository by using the identifier and using the computed distances between nodes.

Abstract: A system for retrieving/identifying a document comprising text stored in a document repository is described. A memory stores a graphical structure comprising a first plurality of nodes each representing a person, and a second plurality of nodes each representing a document in the document repository, the nodes being connected by edges according to automatically observed interactions between the represented people and documents. At least some of the nodes have one or more annotations each denoting a topic. A node relatedness calculator computes distances between nodes of the graphical structure using the topic annotations. An input receives an identifier of a user who is represented by one of the first plurality of nodes. An identifier/retriever identifies one or more documents from the document repository by using the identifier and using the computed distances between nodes.

Abstract: A system for retrieving/identifying a document comprising text stored in a document repository is described. A memory stores a graphical structure comprising a first plurality of nodes each representing a person, and a second plurality of nodes each representing a document in the document repository, the nodes being connected by edges according to automatically observed interactions between the represented people and documents. At least some of the nodes have one or more annotations each denoting a topic. A node relatedness calculator computes distances between nodes of the graphical structure using the topic annotations. An input receives an identifier of a user who is represented by one of the first plurality of nodes. An identifier/retriever identifies one or more documents from the document repository by using the identifier and using the computed distances between nodes.

Abstract: Retrieving and/or storing images associated with events is described. For example, streams of event data comprising text are analyzed to detect an event and a language component builds an event language model for the event, comprising a plurality of words. In various examples, images extracted from web or other sources have associated text. In examples, images with associated text that is similar to the event language model are identified as images of the event. In various examples, associations between images and events are used to update an image retrieval system and/or an image storage system. In various examples, query terms about an event are received at an image retrieval system which returns images related to the event on the basis of associations between image text and event language models.

Abstract: A method is provided of enriching an entry for an entity in a local index of a search engine with tags. The method comprises obtaining location-related social media messages from within a neighborhood of an entity; determining from the obtained messages one or more terms that are unique to the entity; individually determining one or more co-occurring terms for the one or more unique terms; and using the one or more co-occurring term as tags to label the entity in the local index. Furthermore, a method is provided of retrieving social media messages associated with search results.