Abstract: A method may include training a machine learning model to perform a first task before training the machine learning model to perform the second task. The machine learning model includes a generator network and a discriminator network. The training includes training, based on a first training sample associated with the first task, the discriminator network to perform the first task. The generator network may be trained to generate a first synthetic training sample emulating the first training sample. The discriminator network trained to perform the first task may be reinitialized in order for the discriminator network to be trained, based on a second training sample, to perform the second task. The reinitialized discriminator network may be further retrained, based on the first synthetic training sample, to perform the first task. Related systems and articles of manufacture, including computer program products, are also provided.

Abstract: A method may include generating, based on a representation of a hierarchy stored in a database, a visiting sequence data structure. The hierarchy may be stored in a table in the database. Each of a plurality of rows comprising the table may correspond to one of a plurality of nodes comprising the hierarchy. The visiting sequence data structure may include a row vector specifying an order for traversing the plurality of nodes in the hierarchy. A hierarchical window function may be executed by iterating through the plurality of rows in the table in accordance with the order specified by the row vector. The execution of the hierarchical window function may further include determining, for a first node in the hierarchy, a summary value corresponding to a first value of the first node and a second value of a second node descendent from the first node.

Abstract: A specialized in-memory database health check process is utilized to resolve dependencies in a resource indicating requirements for an instance of an in-memory database. Specifically, when an instance of an in-memory database is created in response to a request, a list of one or more component handlers are obtained. These component handlers are modular functions, separate from each other but potentially dependent on one or more other component handlers, and act to validate various requirements listed in a resource for the request. Each of the component handlers are executed individually during execution of a Reconcile function. To the extent that the execution of any component handlers in the list is unsuccessful, the Reconcile function is rerun for another iteration. These iterations continue until all component handlers report back as successful. Instance creation is then considered successful and the instance of the in-memory database can be utilized by users.

Abstract: Calls received by a data center that are associated with a request are monitored. Features are subsequently extracted from the monitored calls so that a machine learning model may use such features to determine that the request will cause the data center to malfunction. The machine learning model can be trained using data derived from a transaction log for the data center. At least one correction action to prevent the data center from malfunctioning can then be initiated in response to such determination. Related apparatus, systems, techniques and articles are also described.

Abstract: A request for an entry of a dictionary having multiple dictionary blocks may be received, where the dictionary stores string values associated with corresponding value identifiers (IDs) for dictionary compression. One of the dictionary blocks may be selectively loaded into memory, and the dictionary block that has been loaded into memory may be searched. A value ID directory may be constructed in memory, where the value ID directory includes last value IDs for the dictionary blocks, and each of the last value IDs is mapped to an index of one of the dictionary blocks that includes a string value for that last value ID. A separator directory may also be constructed in memory, where the separator directory includes separators for the dictionary blocks, and each of the separators is mapped to an index of one of the dictionary blocks that includes a string value for that separator.

Abstract: Methods and systems are used for achieving interface design consistency across micro services. As an example, a user interface (UI) training request including at least a set of reference objects is received, the set of reference objects including at least a set of reference UIs. A user interface behavior reference model (UIBRM) is trained to generate a trained UIBRM by analyzing reference UI displays rendered on a browser in response to interactions with the set of reference UIs. A UI displays assessment request including at least a set of development objects is received, the set of development objects including at least a set of development UIs. A UI displays assessment is performed to generate an assessment of development UI displays by comparing the trained UIBRM to the development UI displays rendered on the browser in response to interactions with at least a subset of the set of development UIs.

Abstract: Methods, systems, and computer-readable storage media for receiving, by a content transfer service of a content management system and from a source system, a first content file comprising first content and first content metadata, the first content metadata being stored in a first format, processing the first content file using a set of metadata retrieval definitions to extract file-type-specific metadata from the first content metadata and map at least a portion of the file-type-specific metadata to a first uniform content metadata file having a second format that is different from the first format, each metadata retrieval definition comprising a computer-executable, declarative procedure, and transferring, by the content transfer service, the first content file and the first uniform content metadata file to a target system, the target system consuming the content at least partially based on the first uniform content metadata file.

Abstract: Embodiments associate a relevant semantic data type (e.g., date) with incoming raw data (e.g., a column of digits) which lacks metadata. Assignment of semantic data type is inferred from a plurality of features. A first step determines a first feature comprising success rate in converting the raw data into various semantic data types. Then, alignment between observed/reference distributions of other features (e.g., data first digit, data length) is determined per-semantic data type. Total scores for each semantic data type are calculated from the combined features, and used as a basis for ranking the semantic data types. The total scores may reflect a weighting of the various features. In a second step, top-ranked semantic data types may be further differentiated from one another by applying additional features. User feedback regarding accuracy of semantic data type assignment, may be incorporated into training data used to modify the feature reference distributions.

Abstract: According to some embodiments, a system to manage database management system backups may include a plurality of database services. For each database service, a database backup resource (e.g., a human-readable structured document) may specify desired backup state information (e.g., a buffer size, a schedule, an amount of memory, etc.). A stateless backup operator component may then compare a present backup state of each database service with the associated desired backup state information. When a mismatch is identified between the present backup state of a database service and the associated desired backup state information, that database service may be automatically reconfigured (e.g., in connection with a recover job, backup job, backup configuration job, keystore backup job, etc.) in accordance with the desired backup state information.

Abstract: Systems and methods extract flat data units from a non-flat input, such as a spreadsheet comprising tables organized according to a hierarchy. First, the non-flat input is read (e.g., using pandas in combination with openpyxl) to create a flat dataframe comprising the content of the non-flat input. Next, individual flat data units (e.g., spreadsheet tables) are recognized and split from the dataframe based upon the appearance of blank rows and/or columns. Headers present in the flat data units are determined (e.g., based upon alphabetic cell text, bolded cell text, and/or early position of the cell in a column), and then connections between the flat data units are identified. Based upon the connections, individual flat data units are merged together. The resulting merged flat data units are subsequently available for consumption, for example user reports of content, and/or conversion to a new non-flat format (e.g., relational database schema).

Abstract: An analytics platform or another system is provided to store multiple sets of credentials and other connection information in each user profile and to map the stored credentials to different resources within the platform. Each set of credentials can be associated with a credential mapping key, and each resource in the platform can be mapped to the credentials using a connection associating the resource with the credential mapping key. By mapping multiple sets of credentials and other connection information to the resources within the platform, the user profile can be transparently authenticated when accessing resources in the platform, different credentials can be used to authenticate the user profile with accessing different resources, and administrators can benefit by having greater control over permissions with the platform.

Abstract: Methods, systems, and computer-readable storage media for a machine learning (ML) model and framework for training of the ML model to enable the ML model to correctly match entities even in instances where new entities are added after the ML model has been trained. More particularly, implementations of the present disclosure are directed to a ML model provided as a neural network that is trained to provide a scalar confidence score that indicates whether two entities in a pair of entities are considered a match, even if an entity in the set of entities was not accounted for in training of the ML model.

Abstract: Systems, methods, and computer media for securing software applications are provided herein. The multi-factor fingerprints allow attackers to be distinguished from authorized users and allow different types of attacks to be distinguished. The multi-factor fingerprint can include, for example, a session identifier component, a software information component, and a hardware information component. The different components can be separately compared to components of stored fingerprints to determine whether an application session request is malicious, and if so, what type of attack, such as session cookie theft or a spoofing attack, is occurring.

Abstract: Various examples are directed to systems and methods for executing a web application with client-side encryption. A web application may execute in a web browser at a client computing device. The web browser may generate a document comprising a secure display element. The web browser may request to render the document at the client computing device. A cryptographic tool of the web browser may decrypt the first encrypted value to generate a first clear value. The web browser may render the document at an output device of the client computing device using the clear value. The web browser may also be programmed to prevent the web application from accessing the first clear value.

Abstract: The presentation of data models (a story) on a mobile device having limited memory and processing resources includes instantiating widgets one at a time for interaction with a user. When the user completes their interaction, the instance of that widget is removed from memory (destroyed); widgets are therefore stateless. A context framework that manages the application of filters to widgets that comprise the story. When a filter is applied, widget snapshot images used to display corresponding widgets in a collection view can be updated to reflect state changes of affected corresponding widgets so that users can see the state changes despite the widget being stateless.

Abstract: A data manager in an enterprise provides data management of users' personal data, which is used by enterprise applications to support operations in the enterprise. The data manager obtains personal data from the enterprise applications via the use of corresponding connector modules. The data manager communicates with each connector module, which then communicates with its corresponding enterprise application to obtain personal data used by that application. The data manager can also communicate with a central database that can store personal data used by the enterprise applications. The central database can serve as a central store of personal data for the enterprise applications.

Abstract: Technologies are provided for assigning developers to source code issues using machine learning. A machine learning model can be generated based on multiple versions of source code objects (such as source code files, classes, modules, packages, etc.), such as those that are managed by a version control system. The versions of the source code objects can reflect changes that are made to the source code objects over time. Associations between developers and source code object versions can be analyzed and used to train the machine learning model. Patterns of similar changes to various source code objects can be detected and can also be used to train the machine learning model. When an issue is detected in a version of a source code object, the model can be used to identify a developer to assign to the issue. Feedback data regarding the developer assignment can be used to re-train the model.

Abstract: Disclosed herein are system, method, and computer program product embodiments for providing a natural, pen-and-paper user experience that harnesses cutting edge artificial intelligence and machine learning to enhance user interactions with data in a suite of software applications. The pen-and-paper user experience may mimic the time-honored human tradition of writing with pen and paper, allowing users to work in an intuitive, productive, familiar fashion, while incorporating artificial intelligence and machine learning techniques to interpret a user's natural language and other pen-strokes in order to better interpret a user's natural language and resolve ambiguity in user inputs. A pen-and-paper-based user experience may be extended and enhanced to allow a user to interact with a suite of applications, a launchpad, a key performance indicator tool, a digital boardroom, or other application using the same pen-and-paper-based design paradigm.

Abstract: Methods, systems, and articles of manufacture, including computer program products, are provided for synthesizing images for machine learning. The method may include selecting one or more image preprocessing transformations to apply on the foreground object image; applying the selected one or more image preprocessing transformations to the foreground object image; selecting a background image from a set of background images depicting a variety of different backgrounds which may be associated with the foreground object image; merging the selected background image with the foreground object image to form a synthesized image; selecting one or more image transformations to apply on the synthesized image; applying the selected one or more image transformations to the synthesized image; and storing the synthesized image in a collection of synthesized images to train a machine learning model.

Abstract: A method may include receiving, from a first client, a first message. The first message may be matched to a second user based on a similarity between a first keyword included in the first message and a second keyword included in a profile of a second user. The first keyword may be determined to be similar to the second keyword based on a distance between a first vector representation of the first keyword and a second vector representation of the second keyword not exceeding a threshold value. In response to the first message being matched with the second user, the first message may be sent to a second client associated with the second user. In response to receiving, from the second client, a second message responsive to the first message, the second message may be sent to the first client. Related systems and articles of manufacture are also provided.