Abstract: A machine learning model is trained with a training dataset, where the machine learning model comprises a plurality of layers. During training, values of a plurality of coefficients of one or more layers are monitored. In response to detecting a change of a given coefficient by more than a threshold during a given training run, a given reference to a given input dataset of the given training run is stored. In response to detecting an output error of a trained version of the machine learning model, the given reference to the given input dataset is retrieved if the given coefficient is located on a backward path providing more than a threshold contribution to the output error. Next, the given reference is provided to an application analyzing the trained version of the machine learning model in order to determine a cause of the output error.

Abstract: A machine learning model is trained with a training dataset, where the machine learning model comprises a plurality of layers. During training, values of a plurality of coefficients of one or more layers are monitored. In response to detecting a change of a given coefficient by more than a threshold during a given training run, a given reference to a given input dataset of the given training run is stored. In response to detecting an output error of a trained version of the machine learning model, the given reference to the given input dataset is retrieved if the given coefficient is located on a backward path providing more than a threshold contribution to the output error. Next, the given reference is provided to an application analyzing the trained version of the machine learning model in order to determine a cause of the output error.

Abstract: A set of locational information items are provided, each of which includes a set of coordinate values. These locational information items can be represented on a map including various levels of detail. A mapping application is provided which is intended to display these locational information items on various levels of detail of the map. The locational information items are clustered on certain levels of detail of the map. To cluster the locational information items, a level of detail for each LOD can be specified; a Delaunay Triangulation can be created over the set of the locational information items provided; a list of the edges associated with the Delaunay Triangulation can be generated and sorted by length; a cluster tree can be built; and the locational information items are displayed in clusters at each LOD.

Abstract: The invention relates to an efficient system for user rights in a semantic digital network, whereby users are arranged in the same semantic network as the information objects. The rights are thus derived from the semantic relations between users and information objects in a common semantic network.

Abstract: A set of locational information items are provided, each of which includes a set of coordinate values. These locational information items can be represented on a map including various levels of detail. A mapping application is provided which is intended to display these locational information items on various levels of detail of the map. The locational information items are clustered on certain levels of detail of the map. To cluster the locational information items, a level of detail for each LOD can be specified; a Delaunay Triangulation can be created over the set of the locational information items provided; a list of the edges associated with the Delaunay Triangulation can be generated and sorted by length; a cluster tree can be built; and the locational information items are displayed in clusters at each LOD.

Abstract: Various embodiments for hit testing of visual objects are described herein. Data of visual objects is generated in a two-dimensional clip space. The data of visual objects includes two-dimensional projections of the visual objects. Cursor coordinates are transformed into the clip space and bounding box calculations are performed using the transformed cursor coordinates and the generated data. Hit testing is performed when there is a hit on a bounding box of at least one of the visual objects in the clip space. The hit testing is performed in a three-dimensional space. A result is then presented on a user interface based on the hit testing.

Abstract: Various embodiments for hit testing of visual objects are described herein. Data of visual objects is generated in a two-dimensional clip space. The data of visual objects includes two-dimensional projections of the visual objects. Cursor coordinates are transformed into the clip space and bounding box calculations are performed using the transformed cursor coordinates and the generated data. Hit testing is performed when there is a hit on a bounding box of at least one of the visual objects in the clip space. The hit testing is performed in a three-dimensional space. A result is then presented on a user interface based on the hit testing.

Abstract: A system may include presentation of a user interface including fields associated with a business object, reception, via the user interface, of an instruction to add an extension field to the business object, reception, via the user interface, of an indication of a reference field of the business object, determination of a binding of the reference field to the business object, and binding of the extension field to the business object based on the determined binding of the reference field to the business object.

Abstract: What is described is a method and system for development of interactive user interfaces. The interfaces provide visualization environment and the opportunity for creating visual applications with minimal coding. A scene is selected from a scene catalogue and a visualization object is selected and displayed on the scene. A data source is selected from backend data. A parameter from the content of the data source is connected to a property of the visualization object to create a script to modify the property. The script is customized and the processed visualization object is displayed on a GUI.

Abstract: The invention relates to an efficient system for user rights in a semantic digital network, whereby users are arranged in the same semantic network as the information objects. The rights are thus derived from the semantic relations between users and information objects in a common semantic network.

Abstract: Systems and methods are provided for managing content. In one implementation a system includes software constructs, each being associated with a context defining a class of content modules included in a respective construct. In the system, a context is selected and context data for a plurality of content modules is retrieved. The context data includes a primary context identifier that identifies the context in which each content module was created or modified. In addition, the context data includes at least one secondary context identifier identifying additional contexts in which the content module is permitted to be included. Based upon the context identifiers, it is determined whether a content module is permitted in the selected context and a collection of modules is generated including the content modules permitted in the selected context based on the determination.

Abstract: A method and apparatus for iterative price zone analysis includes steps of generating auxiliary data for a geographic region that includes a plurality of retail locations, the regions being dividable into a plurality of geographic zones. The method and apparatus further includes modeling of sales information for the retail locations and thereby generate demand model information based on the set zones. The method and apparatus further includes providing the demand model information and the retail locations having the demand model information associated therewith to a price zoning display of the geographic region. Additionally the auxiliary data is provided for display relative to the model information and retail locations, where the display allows for the iterative selection of various factors and recalculations of the modeling information on the visual display.

Abstract: In a business environment, a method for executing context-driven transactions includes defining one or more relations among two or more objects in a semantic network, and defining one or more rules relating to the relations. The “transaction” refers to a request or a sequence of information exchanges for one or more objects. The method includes monitoring the semantic network for changes in any of the objects and relations, and upon detecting a change, automatically triggering a transaction based upon a rule in the system.

Abstract: The invention relates to an efficient system for user rights in a semantic digital network, whereby users are arranged in the same semantic network as the information objects. The rights are thus derived from the semantic relations between users and information objects in a common semantic network.