Abstract: An example method includes (a) generating a request to obtain log data descriptive of operation of an autonomous vehicle control system in a subject scenario, the log data indexed by an indexing parameter; (b) submitting the request to a log repository service, the log repository service configured to receive the request and serve, responsive to the request, log metadata; (c) loading, based on the log metadata and an index value of the indexing parameter, a log data sketch associated with the index value; and (d) loading, based on the log metadata and the index value, and responsive to an inspection indicator, detailed log data associated with the index value.

Abstract: Systems and methods for integration of anchor information for retrieval augmented generation are provided. A communication platform accesses communication data associated with an enterprise on the communication platform. The communication platform divides the communication data into multiple data chunks. The communication platform generates one or more extended queries for each data chunk of the multiple data chunks using a generative AI model. The communication platform receives a user query, and identifies, from the multiple data chunks, one or more relevant data chunks for the user query at least by comparing the user query with the one or more extended queries associated with each data chunk of the multiple data chunks. The communication platform provides an answer to the user query based on the one or more relevant data chunks.

Abstract: The present application generally relates to a method for training a machine learning, ML, model using class incremental learning, and to a computer-implemented method and apparatus for using the trained machine learning, ML, model. The method may learn how to update semantic representations of old concepts (classes) by modelling drift of semantic representations.

Abstract: In an example embodiment, prompt compression is used in two different places in the LLM query pipeline to reduce the size of LLM queries, which reduces the time it takes for an LLM to respond to a query and improves scalability. Specifically, the conversation history itself may be run through a token compressor, which acts to eliminate unnecessary tokens from the conversation history and thus also from a rewritten version of the query. Additionally, the context chunks retrieved during RAG are themselves run through a token compressor to eliminate unnecessary tokens from the context chunks. The compressed context chunks and the rewritten query are then passed to the LLM for processing, and since both the context chunks and the rewritten query contain compressed elements, the result is that the LLM processes fewer tokens to respond to the query.

Abstract: Embodiments described herein are generally related to systems and methods for generating data visualization scores, for use with data analytics environments. In accordance with an embodiment, the system can operate in the manner of an expert system, or according to a series of processes or rules, to examine a data visualization of interest, compare a list of found elements with element types specified by an analytics data visualization score matrix, and generate, based on matching found elements with the analytics data visualization matrix, a data visualization score associated with the data visualization. In accordance with an embodiment, the system can operate as a data visualization advisor, during preparation of a data visualization, to provide a user with recommendations or score values indicative of a quality or complexity of their data visualization, which may be helpful in improving their data visualization, for example from a beginner-level to a more advanced-level.

Abstract: Techniques for utilizing machine learning to generate a question and answer pair for a conversational agent are described.

Abstract: Systems and methods for information retrieval and knowledge discovery are disclosed. The method includes, receiving a user input from a user via a user interface, wherein the user input comprises a user requirement and a user preference, identifying a user intent and a context of the received user input by parsing a plurality of keywords within the received user input, and extracting multimedia content from a structured data source based on the identified user intent, the context, the user requirement, and the user preference, wherein the structured data source stores the multimedia content retrieved from a plurality of data repositories.

Abstract: A method for automatically categorizing groups of users in a social network, and using the categories to propose matches between groups or users and groups includes accessing content shared within a group in a social network application, processing the shared content with at least one machine learning model to determine labels for the content, determining categories for the group based on the labels for the content, accessing matching criteria for a user of the social network application, generating a match between the user and the group based on the matching criteria for the user and the categories for the group, supplying data indicative of the match to a client device of the user, receiving a match acceptance from the client device, and in response to receiving the match acceptance, providing the user access to the group in the social network application.

Abstract: An example computing system receives a media content search query from a controller configured to control media playback by one or more playback devices in a media playback system associated with a plurality of media content services. Based on the search query, the system generates a first set of search queries for the first media content service and a second set of search queries for the second media content service. The computing system receives search results from the first and second media content services and aggregates the search results such that the search results are organized based on (i) categories of media content items corresponding to the search results and (ii) the media content services from which the search results are received. The computing system sends the aggregated search results to the controller for display in their aggregated order.

Abstract: Methods and systems for providing customized services to users of data processing systems are disclosed. To provide the customized services, a query may be obtained from a user. Contextual information for the query may be attempted to be obtained from a first data source associated with the user. If sufficient contextual information is unable to be obtained from the first data source, at least one other user may be identified based on the user. Corresponding portions of supplementary contextual information may be obtained from a plurality of other data sources associated with the at least one other user. A consensus response may be obtained based on the corresponding portions of the supplementary contextual information to indicate a most common response for a topic indicated by the query. The query may be serviced using at least the consensus response and any other information obtained from the first data source.

Abstract: According to some embodiments, a digital core for a cloud computing environment is associated with an enterprise (e.g., the digital core may comprise enterprise resource planning software optimized for an in-memory database). Moreover, an integration inspector framework may establish a checkpoint between an application system (e.g., a source application system to provide information via the digital core and/or a target application system to receive information via the digital core) and the digital core. The integration inspector framework then automatically validates a communication arrangement associated with the application system and performs a data replication test, in connection with the application system and the digital core, using test data. A result of the data replication test is then evaluated and indicated (e.g., to a user via a user interface).

Abstract: Artificial intelligence techniques for query management are described. A method comprises generating, by a context detection module, context information for a first query comprising natural language information to request a result from one of a plurality of machine learning models, modifying, by a query modification module, the first query based the context information to form a first modified query, determining, by an intent module, an intent type for the first modified query, selecting, by a routing module, a machine learning model from the plurality of machine learning models based on the intent type, and routing, by the routing module, the first modified query to the selected machine learning model. Other embodiments are described and claimed.

Abstract: Current generative AI systems using large language models (LLMs) face challenges including non-deterministic outputs, hallucinations, outdated information, and resource-intensive training. This disclosure introduces topic maps for constrained retrieval augmented generation to address these issues. The technique leverages existing LLMs while constraining outputs to specific, user-defined content domains. Topic maps, composed of topic names, descriptions, and relevant resource references, create a curated knowledge base that guides agent responses. This approach reduces hallucinations, improves consistency, and allows for dynamic updates without model retraining. The method involves receiving a query, identifying relevant topic maps, transmitting the query and references to an AI agent, and generating constrained responses. By providing a structured, updatable knowledge framework, this method enhances the accuracy, reliability, and adaptability of generative AI systems.

Abstract: A method and a system for determining similarities between entities are provided. The method includes: receiving a request to determine an estimate for a first instrument; generating a query to identify corresponding first data sources that contain data about an entity; identifying, using the generated query, the corresponding first data sources from a system of computer networks; filtering the corresponding first data sources in order to determine at least one primary data source; extracting corresponding first information that relates to the respective entity from the at least one primary data source; identifying, using a large language model, corresponding entity-specific information based on the extracted corresponding first information; comparing the entity-specific information of a first entity with entity-specific information for a plurality of entities; and generating a similarity ranking that list each respective entity from among the plurality of entities in relation to the first entity.

Abstract: Some embodiments provide a method for generating a graphical user interface (GUI) for a research system. The method receives a request from a user of the research system for information about a particular category. The method generates a chart that displays a set of events associated with the particular category over a particular period of time. The method incorporates the chart into a GUI for the particular category for transmission to the user. Some embodiments generate a list of events associated with the particular category and generate a GUI that displays the list of the events. Each event is represented in the list by a title of a document identified by the research system as representative of the event.

Abstract: A data processing method, a data processing apparatus, and a health management apparatus. The data processing method includes: performing data processing on device-related data from a device associated with a health management apparatus. The data processing apparatus and the health management apparatus can improve the uniformity of memories provided to be associated with the health management apparatus.

Abstract: Provided are a method of processing encryption of database (DB) storage data and a database management system (DBMS). The method includes, when an encrypted column inquiry request is received through an application, executing a hooking module, inquiring about, by the executed hooking module, encryption data in a DB of a DBMS, and decrypting retrieved encryption data and transmitting the decrypted data to the application. According to the method, a DB client intervenes in structured query language (SQL) transmission to a DB server and a data receiving process to automatically perform encryption and decryption. Accordingly, it is possible to automatically decrypt encrypted data without modifying an application source, or conversely, input data to a DB in an encrypted state.

Abstract: Systems and methods for retrieving information from a corpus of documents that is relevant to a query. The method comprising: subdividing each document in the corpus of documents into one or more chunks; using a synthetic generation LLM to generate at least one piece of synthetic information related to each chunk; generating, using an embedding model a plurality of vectors for each chunk, the plurality of vectors for a chunk comprising a vector generated from the chunk, and a vector generated from each of the at least one piece of synthetic information related to that chunk; and using an information retrieval system to identify, from the plurality of vectors for each chunk, a set of chunks that are relevant to a query.

Abstract: Techniques for ingesting data using an attribute definition file are disclosed. The system receives an attribute definition file that defines an attribute name, an entity type, and one or more validation criteria for an attribute. The system receives a dataset comprising a plurality of attribute values with a respective plurality of attribute names. The system determines whether an attribute name in the dataset maps to an attribute name in the definition file and maps the attribute value in the dataset to the attribute definition of the attribute. The system then determines whether the attribute value meets the validation criteria in the definition and stores the attribute value in a data repository when the value meets the validation criteria.

Abstract: A method, an apparatus, and a non-transitory computer-readable storage medium for video compression using a generative adversarial network (GAN) are provided. The method includes obtaining, by a generator of the GAN, a reconstructed target frame based on a reference frame and a raw target frame to be reconstructed; concatenating, by a transformer-based discriminator of the GAN, the reference frame, the raw target frame and the reconstructed target frame to obtain a paired data; determining, by the transformer-based discriminator of the GAN, whether the paired data is real or fake to guide reconstruction of the raw target frame; and determining a generator loss and a transformer-based discriminator loss, and performing gradient back propagation and updating network parameters of the GAN based on the generator loss and the transformer-based discriminator loss.