Abstract: Mechanisms are provided for training a computer implemented model. The mechanisms perform multiple instances of training of the computer implemented model, where each instance of training of the computer implemented model comprises training the computer implemented model using a different training data set to generate a different instance of a trained computer implemented model. The mechanisms generate computer implemented model results after each instance of training by executing the corresponding instance of the trained computer implemented model. The mechanisms record differences in the instances of training of the computer implemented model in association with corresponding identifiers of the instances of trained computer implemented model and corresponding computer implemented model results.

Abstract: Embodiments disclosed herein generally relate to a method for searching a database. The application receives an input identifying one or more users. The application matches the input to one or more user accounts, each user account corresponding to a user that was identified. The application presents to a requestor public accounts of each identified user account. The application receives a keyword input that filters one or more files in the public accounts. The application receives an input that saves a search comprising the user accounts and the keyword input. The application queries the database using the search. The application presents to the requestor a list of all files found using the search. The application dynamically updates the list to include new files matching a criteria of the search as new files satisfying the search are uploaded to the database.

Abstract: A method includes receiving a search query, identifying a document based on the search query, and providing a search result based on the document. The search result includes, for example, an image associated with the document, an excerpt from the document that is associated with the search query, and links to other excerpts in the document that are associated with the search query. The method may also include providing other information associated with the document.

Abstract: Systems and methods are provided for generating a rank vector comprising an array of values for each term position of a document and setting a value at a position of each term identifier in the document to a distance score for each term identifier, calculating snippet frame quality scores for each frame of a plurality of frames in the rank vector, determining whether a length of the character string associated with the frame with the highest frame quality score is similar to a requested length for a snippet, if necessary, adjusting the size of the character string to be similar to the requested length for the snippet and, once the length of a character string for one or more frames is similar to the requested length for the snippet, generating the character string to be provided as a snippet.

Abstract: Examples are provided for a column browser for navigating hierarchical data structures. The column browser can include multiple columns for displaying different levels of the hierarchical data structure and the information, options, and other data stored in nodes of the different levels. The column browser can navigate to different levels of the hierarchical data structure by traversing through parent node and child node links in the hierarchical data structure. Based on the links, the column browser can navigate up or down levels of the hierarchical data structure.

Abstract: A system and method is provided to selectively process and store tables of a relational database by calculating an overall data sensitivity score for each table based on predefined attribute rules; performing column-wise splitting of at least one of the tables into a first table and a second table based on the overall data sensitivity score of each table, thereby generating a total number of relational database tables; storing a first subset of the total number of relational database tables in a private cloud storage database in a distributed storage environment based on the overall data sensitivity scores of each of the total number of relational database tables; and storing a second subset of the total number of relational database tables in a public cloud storage database of the distributed storage environment based on the overall data sensitivity scores of each of the total number of relational database tables.

Abstract: A computer-implemented method for a crash recovery for linked databases may be provided. The linked databases comprise a source and related target database. Selected queries of the source database are transferred to the target database. The method comprises synchronizing selected portions of the source database with tables of an in-memory portion of target database and, storing persistently applied changes to the in-memory target database portion asynchronously and persistently. Upon a database crash of the target database system, the method also comprises restoring, the in-memory target database portion with the latest snapshot available, and applying, changes from the source database recovery log file that have a later timestamp than the latest snapshot available in the persistent target database storage of the in-memory target database portion.

Abstract: A system executes a set of database operations and determines counts of instances that each key is specified for a corresponding column by any database operations on a database table. The system identifies each key which is associated with any determined count that satisfies a threshold as a corresponding frequently accessed key. The system creates an optimized index for each column which stores any frequently accessed key. The system inserts each frequently accessed key into a corresponding optimized index. The system receives a database operation that specifies a specific key for a specific column in the database table. If any optimized index matches the specific column and stores any frequently accessed key that matches the specific key, then the system references a matching frequently accessed key in a matching optimized index to access a record, which is associated with the specific column and the specific key, via the database table.

Abstract: Technologies described reduce the amount of time between requesting content and receiving the requested content. In some embodiments, a client provides a search query. A search result page is generated and, prior to returning the search result page to the client, some or all of the search results are asynchronously pre-fetched and cached. The search query can include custom parameters that are also included in the pre-fetch requests. Web addresses in the search results page can be modified to indicate whether a prefetch operation was initiated. Once a user activates one of the search results, the responding web server uses the cached data while generating a response.

Abstract: Venn diagrams are computed for a given plurality of input sets. The process of computing the Venn diagrams is executed on columnar database systems for efficient execution. The computation of various subsets of the Venn diagrams is performed by determining subsets of various combinations of the input sets and computing set differences of the intersection sets. The process orders the execution of various steps of computing the subsets for the Venn diagram in an order that reduces the number of times an input set is loaded. Information describing various subsets of a Venn diagram is used to render the Venn diagram for display, for example, on a client device.

Abstract: A computer-implemented method for automatically adjusting a Uniform Resource Locator (URL) seed list. The method includes crawling for documents based on a seed URL list. The method generates relations data from the documents using a Natural Language Processing (NLP) model. The method analyzes the relations data using an auto-seed model. The method modifies the seed URL list.

Abstract: Techniques for improved searching and querying in computer-based reasoning systems are discussed and include receiving multiple new multidimensional data element to store in a computer-based reasoning data model; determining a feature bucket for each feature of each data element and storing a reference identifier in the feature bucket(s). A query on the computer-based reasoning system includes input data element (e.g., an actual data element, or a set of restrictions on features). For each feature in the input data element, feature buckets are determined, candidate results are determined based on whether cases have related feature buckets, and the results are determined based at least in part on the candidate results. In some embodiments, control of controllable systems may be caused based on the results.

Abstract: Provided is a system and method for automated feature engineering of multidimensional data. The system and method may generate a flat data set for use in training a machine learning model from a multidimensional data set. In one example, the method may include extracting metadata of multidimensional data, generating declarative statements of features that are to be derived from the multidimensional data based on hierarchical attributes of the multidimensional data which are identified from the extracted metadata, querying, via a query engine, the multidimensional data based on the generated declarative statements to generate the derived features, and storing the derived features in a training data set.

Abstract: An information retrieval method is provided, applicable to a database. The database includes a plurality of pieces of information and a plurality of index tags, and each of the pieces of information is associated with at least one of the index tags. The information retrieval method includes: receiving a piece of retrieval information; extracting pieces of first information associated with the retrieval information from the pieces of information of the database; extracting a plurality of first index tags corresponding to the pieces of first information from the index tags; requesting a display field of a first index tag to be adjacent to or located in a display field or display page of first information corresponding to the first index tag; receiving a filter tag selected from the first index tags; and extracting pieces of second information associated with the filter tag from the pieces of first information.

Abstract: Methods and system for recommending components to a user. Implementations are directed to receiving, from a user, a user query for a design recommendation, the user query comprising a plurality of terms; determining, from the plurality of terms, a user intent including at least one query-derived design function and not including a named component for the design recommendation; receiving additional contextual information for the design recommendation, wherein the additional contextual information comprises a plurality of action attributes for the at least one query-derived design function; determining, from a database, a plurality of design recommendations, wherein each design recommendation comprises a respective recommended component and a plurality of recommended component actions; selecting, from the plurality of design recommendations, a subset of design recommendations; and providing, to the user, the subset of design recommendations.

Abstract: Systems and methods are provided herein for enabling a computing system to search and interact with service records containing natural language text to aid in the analysis of those records by: displaying in a user interface an image of a complex system; receiving, from the user interface, a high-level selection of criteria of the complex system; querying issue maps associated with individual natural language service records of a corpus of natural language service records based on the high-level selection, wherein the issue maps specify at least one term related to the complex system and a location on the complex system associated with the at least one term; and returning at least one issue map, wherein the at least one issue map returned specifies a term or location correlated to the criteria of the complex system indicated by the high-level selection.

Abstract: A system and method for generating an electronic catalog booklet for online computer users is disclosed. The system and method in an example embodiment includes providing access to a database of item listings; enabling selection of a plurality of item listings from at least some of the item listings provided in the database; enabling arrangement of the plurality of selected item listings on a plurality of pages in a computer-generated electronic catalog booklet; recording an identifier of the computer-generated electronic catalog booklet; and recording information associating each of the plurality of selected item listings with the identifier of the computer-generated electronic catalog booklet.

Abstract: Embodiments of the present invention relate to a method, system, and computer program product for multiplexing data operation. In some embodiments, a method is disclosed. A query for at least one table comprising a plurality of data records is received. The query indicating a plurality of data operations to be performed on the plurality of data records. The plurality of data operations are combined into a target data operation. An intermediate result of the query is generated by performing the target data operation on the plurality of data records. A final result of the query is determined based on the intermediate result. In other embodiments, a system and a computer program product are disclosed.

Abstract: Data-deduplicating includes comparing a first record of a data-store with a second record of the data-store but instead of using a static weight for a field, the present data-deduplicating dynamically assigns a first weight for the first score to generate a first weighted score, wherein the first weight is based on one or both of the first value or the second value; and assigns a second weight for the second score to generate a second weighted score. A composite score is calculated based on the first weighted score and the second weighted score; and it is determined whether or not the first record and the second record are duplicate records, based on the composite score.

Abstract: Systems and methods for garbage collection are disclosed. A garbage collection is triggered in a memory allocation based on a free list and selected amount of free space in an inner closed feedback loop. The free list is determined from a selected memory load and a current memory load as generated from the garbage collection in an outer feedback loop.