Some embodiments optimize a local, pending transaction queue of transactions locally executed by a client application to be synchronized with a server. Multiple transactions, which may be executed by the client application for a same local object, need to be synchronized with a server. The transactions, in queue for synchronization with the server, may include transactions to add, delete, or modify objects. Each set of transactions for a particular object in queue for synchronization with the server corresponds to a particular net effect on the server if synchronized individually to the server. The client application may replace the set of transactions for the particular object with a reduced set of transactions corresponding to the net effect, or entirely removed without replacement if there is no net effect of the set of transactions on the server.

A resource-efficient technique is described for producing and utilizing a set of trained embeddings. With respect to its training phase, the technique receives a group of sparsely-expressed training examples of high dimensionality. The technique processes the training examples using a distributed training framework of computing devices. With respect to its inference stage, the technique draws on the embeddings produced by the training framework. But in one implementation, the inference-stage processing applies a different prediction function than that used by the training framework. One implementation of interference-stage processing involves determining a distance between a query embedding and a candidate item embedding, where each such embedding is obtained or derived from the trained embeddings produced by the training framework. Another manifestation of inference-stage processing involves adjusting click counts based on identified relations among items embeddings.

Digital content search techniques are described that overcome the challenges found in conventional sequence-based techniques through use of a query-aware sequential search. In one example, a search query is received and sequence input data is obtained based on the search query. The sequence input data describes a sequence of digital content and respective search queries. Embedding data is generated based on the sequence input data using an embedding module of a machine-learning model. The embedding module includes a query-aware embedding layer that generates embeddings of the sequence of digital content and respective search queries. A search result is generated referencing at least one item of digital content by processing the embedding data using at least one layer of the machine-learning model.

A system and method of processing a search request is provided. Identification of a desired content item is based on comparing a topic of the search request to previous user input. The method includes providing access to a set of content items with metadata that describes the corresponding content items and providing information about previous searches. The method further includes receiving a present input from the user and determining a relatedness measure between the information about the previous searches and an element of the present input. If the relatedness measure is high, the method also includes selecting a subset of content items based on comparing the present input and information about the previous searches with the metadata that describes the subset of content items. Otherwise, the method includes selecting a subset of content items based on comparing the present input with the metadata that describes the subset of content items.

A computer-implemented technique is described herein for expediting a user's interaction with a digital assistant. In one implementation, the technique involves receiving a system prompt generated by a digital assistant in response to an input command provided by a user via an input device. The technique then generates a predicted response based on linguistic content of the system prompt, together with contextual features pertaining to a circumstance in which the system prompt was issued. The predicted response corresponds to a prediction of how the user will respond to the system prompt. The technique then selects one or more dialogue actions from a plurality of dialogue actions, based on a confidence value associated with the predicted response. The technique expedites the user's interaction with the digital assistant by reducing the number of system prompts that the user is asked to respond to.

A method may include executing, on a processing unit, a received search query for an electronic messaging data store according to an entity type and a directory scope; in response to the executing: presenting, on a computing device, results of the search query from the electronic messaging data store; and presenting a user interface option configured to initiate storing the search query as a pinned search; receiving, by the processing unit, an indication that the user interface option was activated; based on receiving the indication, generating a pinned search data structure identifying the entity type and the directory scope; subsequent to the generating, receiving a request to initiate a search of the electronic message data store; and in response to receiving the request, presenting a search panel that identifies the pinned search data structure as an activatable search query.

Systems, methods, and computer-readable media for fuzzy-searches on encrypted messages include maintaining, in an indexer, a dictionary of words appearing in a message history. Upon receiving a query including at least one search term, a fuzzy search of the dictionary using the at least one search term is performed to determine one or more fuzzy-matching words in the dictionary, and one or more search tokens are generated from the one or more fuzzy-matching words, the one or more search tokens including encrypted versions of the one or more fuzzy-matching words. The one or more search tokens are provided to a search service for searching a database of encrypted messages of the message history, where the at least one search term may not have an exact match with any of the words in the dictionary.

Some embodiments locally cache search results received from a server in an object-agnostic data structure for the local execution of subsequent queries by a client application. The object-agnostic data structure stores objects returned, from a server, in a set of query results corresponding to one or more initial queries. The locally cached objects correspond to a partial set of the objects stored at a server. The object-agnostic data structure allows for the local execution of a subsequent query on the partial set of objects to generate a partial result set corresponding to the subsequent query. The partial result set for the subsequent query may be desired when the server is inaccessible or query executions on the server take longer than a threshold amount of time. The client application may update the partial result set with the complete result set received from the server upon availability of the complete result set.

The present invention is directed towards systems and methods for conducting legal research across multiple jurisdictions, which comprises receiving a first user search query at a graphical user interface comprising one or more search terms, executing the first search query against a first jurisdictional database and identifying one or more search results that are responsive to the first search query. Upon receiving a selected search result, a modified search query is generated comprising the one or more search terms of the first search query and one or more relevant terms from the selected search result and is executed against a second jurisdictional database. One or more secondary search results that are responsive to the modified search query are identified and then presented at the graphical user interface.

Systems and methods are provided for improved interactive graphical user interfaces. The system allows users to visually construct queries of a database. The system can dynamically generate visual representations of supplemental data and cumulative data. Users can share annotations related to user input queries via the graphical user interface.