Abstract: Techniques are provided for dynamic context-based evaluation of language models. One method comprises accessing a data structure comprising information characterizing a user input and a corresponding context of the user input; dynamically identifying, in response to the user input, evaluation criteria, from a pool of evaluation criteria, based on the context of the user input; dynamically generating evaluation metrics based on the identified evaluation criteria; applying the user input to at least one language model, wherein the at least one language model generates a response based on the user input; evaluating the response from the at least one language model using the dynamically generated evaluation metrics to assign a score to the response; and initiating an automated action based on the score assigned to the response. A weight may be assigned to the evaluation metrics based on the context of the user input.

Abstract: Systems and methods are described for generation and execution of modified queries. An input can be received via a visualization of a user interface. The input may identify a first field value and a first field for execution of a query. A set of data for execution of the query can be identified based on the input. Alias data may identify a second field that is associated with the first field. Using the alias data, a modified query can be generated based on the query and the second field. The modified query can be executed to generate query results. The query results can be displayed via a visualization of the user interface based on the first field.

Abstract: An automated data cross-checking method includes executing following steps through a processor: receiving a question string and verifying a key field content of the question string; reading and executing agent programs to parse the key field content; determining whether to execute a data cross-check based on the key field content; in response to determining to execute the data cross-check, searching and reading datasets corresponding to the key field content, executing the agent programs to parse fields of each dataset, generating a cross-checking instruction, and outputting a cross-checking prompt, where each dataset includes a plurality of fields; receiving a verification string related to the cross-checking prompt; executing the agent programs to generate a program code based on the cross-checking instruction and the verification string; executing the program code to generate a cross-checking result file; outputting a cross-checking result string corresponding to the question string based on the cross-chec

Abstract: Methods and systems are described configured to generate and distribute content over a network and cause the content to be rendered on one or more devices. A computer system may comprise a processing device comprising one or more arrays of tensor cores, matrix multipliers, and/or SIMD (Single Instruction, Multiple Data) units. The processing device may comprise a dataflow or systolic array architecture. The computer system receives and segments text. Claims may be extracted and may be determined to be conclusion or premise claims. A graph structure may be generated reflecting the relationship between the claims. A characteristic of respective claims may be determined. Using the characteristic, a determination may be made as to whether additional content is to be rendered in association with a given claim. In response to such a determination, additional content may be caused to be rendered on a device adjacent to the given claim.

Abstract: Systems and methods are provided herein for automatically arranging media assets that are contextually related to a requested media asset for access prior to a start time of the requested media asset. In some embodiments, user input circuitry may receive a user request to access a media asset. Control circuitry may identify, by consulting a database, a first set of media assets that share a context with the requested media asset and were released prior to the release date of the requested media asset. Control circuitry may determine a subset of media assets of the first set of media assets by performing a comparison of each media asset of the first set of the media assets against a viewing history of the user. Control circuitry may access a user profile to determine a user constraint, and generate a recommendation for display based on a first aspect of the user constraint.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for iteratively generating different sections of a textual electronic document (“TED”) using one or more large language models (LLMs). In one aspect, a method comprises receiving a request to generate a TED, generating a generation schema of the TED specifying two or more sections of the TED for separate generation, inputting a first set of commands instructing one or more LLMs to generate a first set of text for a first section of the generation schema, inputting a second set of commands including at least a portion of the first set of text and instructing the LLMs to generate a second set of text for a second section of the generation schema, and creating the TED based on an aggregation of the first set of text and the second text according to the generation schema.

Abstract: The implementations of the present disclosure provide a method and a system for generating a semantic response to a query. The method comprises receiving unstructured continuous data of a user, classifying the unstructured continuous data into a first type of data, by a first classification model, a second type of data, by a second classification model, and a third type of data, by a third classification model, wherein the second and third type of data includes name related data and event related data respectively. The method further comprises storing the first, second and third type of data in a first, a second and a third database respectively, and receiving a query to search for one or more details corresponding to the user, wherein the one or more details are based on the unstructured continuous data of the user.

Abstract: Change detection and updates using probabilistic data structures are described. In one or more examples, a change is detected to a dataset record that is used as a basis to generate a first sketch as a probabilistic data structure. A second sketch is generated as a probabilistic data structure based on the change to the dataset record. The first sketch is replaced with the second sketch as stored in a database, the database supporting a probabilistic result to a query operation.

Abstract: A method can include: storing, on an electronic platform, a vectorized item database corresponding to a plurality of items; tracking electronic interactions between the electronic platform and a user computer during a user session; generating, using an embedding model, a session context embedding based, at least in part, on the electronic interactions tracked during the user session; in response to receiving the search query from the user computer, querying the vectorized item database to identify search results corresponding to one or more of the plurality of items, at least in part, using the session context embedding; and transmitting, by the electronic platform, the search results to the user computer. Other embodiments are disclosed.

Abstract: An overlay system is provided that includes a plurality of storage elements and processing circuitry coupled thereto. The plurality of storage elements store a plurality of executable graph-based models such that first and second storage elements store first and second executable graph-based models, respectively. Each executable graph-based model includes a plurality of nodes. The processing circuitry receives a stimulus to share a first node of the first executable graph-based model with a second node of the second executable graph-based model. The processing circuitry instantiates a tenant overlay node that is associated with the second node and includes a set of constraints to be adhered to by the second node while sharing the first node. The processing circuitry creates a sharing channel as a medium between the first and second storage elements. The sharing channel and the tenant overlay node enable sharing of the first node with the second node.

Abstract: Storage service version configuration data is obtained at a time when a data storage node cluster starts a data storage service based on a debugging mode. Based on the storage service version configuration data and data storage node cluster topology information, graph metadata of the distributed graph database is created at a time when no data storage service is provided, where the graph metadata includes a storage service version configuration field populated with the storage service version configuration data and a node metadata field of each data storage node. Node metadata information of the data storage node is obtained at a time when the data storage node cluster starts the data storage service based on a normal mode. The obtained node metadata information of the data storage node is filed into a corresponding node metadata field of the graph metadata to obtain managed graph metadata.

Abstract: This disclosure relates to systems and methods related to smart sharing options for a shared digital asset (DA) library. An example method performed by a system includes: initiating a camera session; identifying a sharable DA trigger; capturing a DA during the camera session after identifying the sharable DA trigger; and selecting the captured DA for the shared DA library based on the identified sharable DA trigger.

Abstract: Systems, methods, and articles of manufacture, including computer program products, provide a system including at least one data processor and at least one memory storing instructions which, when executed by the at least one data processor, cause operations comprising: generating, by a database execution engine, a query plan including a plurality of operators; inserting, by the database execution engine, an enforce compilation operator into the query plan that includes the plurality of operators, the plurality of operators comprising a first operator, the enforce compilation operator, and a second operator; executing at least the first operator of the query plan; in response to executing the first operator, evaluating, by the database execution engine, an output of the first operator to determine whether a condition is satisfied; and in response to the condition being satisfied, triggering, by the database execution engine, a just-in-time compilation of the second operator.

Abstract: In a cloud database system employing multiple types of storage, such as external object store, managed object store. block storage, and compute node memory, each type of storage having different kinds of file organization, different types of data organization, different forms of storage access, and different latency and throughput costs, a system and method for caching different data transformations created during query executions involving different data stores. Transformed versions of data read from external object storage are saved to a multi-layered warehouse cache for use in subsequent query executions.

Abstract: Systems and methods for multi-image search can include obtaining two or more images and determining one or more search results that are based on the two or more images. The one or more search results can be determined based on determined shared attributes of the two or more images. The one or more search results may be based on feature embeddings associated with the two or more images. The two or more images may be obtained based on one or more user interactions with one or more databases.

Abstract: Disclosed herein is a method for predictive and real-time adaptive music recommendation. The method comprises the steps of: acquiring data, the data comprising biometric data and user input data, the user input data indicating a context in which a playlist of music items is to be used; analyzing the data to determine a user state profile including a physiological state and a psychological state; and generating the playlist based on the user state profile. A corresponding system for predictive and real-time adaptive music recommendation is also disclosed.

Abstract: A data processing method applied to the data warehouse system includes that a coordinator node allocates a data write request of a data table to a first computing node, and the first computing node checks whether target data corresponding to the data write request meets a constraint condition. If the check succeeds, a metadata cluster checks the target data again. If the check performed by the metadata cluster also succeeds, a cloud storage cluster writes the target data into the data table. It can be learned that the data warehouse system uses a two-layer indexing mechanism and the metadata cluster of a cloud native architecture to check whether data to be written into the data table meets the constraint condition. Therefore, when a data constraint capability is provided, data reliability is ensured, and that system running performance is not affected is also ensured.

Abstract: A method for determining a complexity of a natural language query includes converting a first natural language query into executable program code, the first natural language query being a query for a first video to be answered by one or more video question answering (VideoQA) models. The method also includes generating, via a complexity model, an abstract syntax tree (AST) based on the executable program code. The method further includes determining, via the complexity model, a complexity of the first natural language query based on quantity of subtrees, from a group of subtrees, that are present in the AST.

Abstract: A client application sends an application programming interface call to an interface of a content management system. The call specifies one or more content item search criteria. Sending the call causes the content management system to perform a lookup in a content item index to identify at least one content item that satisfies the one or more content item search criteria. Based on sending the call, the client application receives from the content management system a suggestion to attach the at least one content item to a text being displayed by at the computing system.

Abstract: A generalized natural language—operating system (“GNL-OS”) for constraining a large language model (“LLM”) to provide deterministic output may be provided. The system may include delivery frameworks. Each delivery framework may deliver generalized natural language-program architecture natural language code (“GNL-PA NLC”) to the LLM. The LLM may operate on one or more hardware processors. The delivery frameworks may include a system framework. The system framework may deliver a system prompt structured in GNL-PA NLC code to the LLM. The system prompt may constrain the LLM to operate as an operative system. Upon receiving the system prompt, the LLM may operate as an operative system under one or more rules specified in the system prompt. The delivery frameworks may include a command framework. The command framework may deliver GNL-PA code to the LLM. The GNL-PA code may correspond to one or more commands executable within the operative system.