Abstract: Deep learning is applied to combined image and text analysis of messages that include images and text. A convolutional neural network is trained against the images and a recurrent neural network against the text. A classifier predicts human response to the message, including classifying reactions to the image, to the text, and overall to the message. Visualizations are provided of neural network analytic emphasis on parts of the images and text. Other types of media in messages can also be analyzed by a combination of specialized neural networks.

Abstract: Methods and apparatus, including computer program products, implementing and using techniques for providing a customized representation of a business process involving one or more organizational entities that are accessible through a network. A user interface that can display a customized representation of a business process is provided. User information, including user privileges associated with the business process, is stored in a central repository that is accessible through the network. A request to view the business process is received from a first user through the user interface. The user privileges for the first user are retrieved from the central repository. A customized representation of the business process is displayed with the user interface. The customized representation is in compliance with the retrieved user privileges for the first user.

Abstract: In accordance with embodiments, there are provided mechanisms and methods for selecting a synchronous or asynchronous process to determine a forecast. These mechanisms and methods for such synchronous/asynchronous process selection can enable embodiments to determine forecasts for multiple users (e.g. with hierarchical relationships, etc.) over an arbitrary time interval. The ability of embodiments to provide forecasts that involve such a large amount of data in an effective way can enable forecasting that was otherwise infeasible due to resource limitations.

Abstract: The technology disclosed describes staging and deploying major releases, updates and hot fixes to service clouds spanning data centers that include hardware stacks. User-specified builds of deployable installations are received as file sets, and the installations are staged to local storage at the data centers. User-specified deployment launch specifications that specify multiple already-staged deployable installations and service clouds to execute a deployment are received; and asynchronous, concurrent deployment of the multiple deployable installations by the service clouds based on the deployment launch specification are launched.

Abstract: In some implementations, a processing system of a database system may identify identification information associated with a user within a request, received from a client device, for data associated with the user, compare the identification information with user identity entries persisting on the memory device, determine, from the comparison, that a user identity entry persisting on the memory device is associated with the user, and query a backend database associated with the client device with one or more parameters included within the user identity entry to obtain the data associated with the user.

Abstract: A method of establishing personalized limits on a search responsive to a key word query in an enterprise search system is described that includes receiving an object types access history for a particular user. Applying this method, the object types access history includes records of object types selected from search results returning multiple object types and records of object types selected via interfaces other than search results. The method continues with determining and storing in computer readable memory a personalized scope of object types. The personalized scope of object types sets a limit on object types initially returned by an enterprise search system for the particular user in response to key word queries by the particular user that do not specify object types to search.

Abstract: Systems and methods for tracking engagement with an online application while multi-task among assignments are disclosed. The technology tracks time an agent actively spends on each work item by tracking how long the tab is in-focus in the service console. Agents log in, and work may be routed to them, based on their capacity. After the agent accepts the work items, the system tracks the time the agent stays on each open work tab. Each time the agent switches to a different tab, or back and forth, the time count stops for the previous tab and starts counting for the current tab. When the agent closes the tab, the total active time spent on the related tab is saved along with the agent's work record. If an agent logs out, the active time is saved for all of their open work tabs and subtabs.

Abstract: The technology disclosed relates to identifying and notifying a user of nearby attendees at a mega attendance event who are in user's social graph by comparing the user's social graph to a list of event attendees. The identified attendees can be stratified into social graph tags that annotate, categorize and prioritize other users in the user's social graph. The technology disclosed also relates to identifying and notifying the user of nearby attendees of sessions at the event who meet introduction preferences of the user by finding matches between introduction preference attributes specified by the user and attributes of the attendees provided by the list of event attendees.

Abstract: The technology disclosed provides a so-called “pointer sentinel mixture architecture” for neural network sequence models that has the ability to either reproduce a token from a recent context or produce a token from a predefined vocabulary. In one implementation, a pointer sentinel-LSTM architecture achieves state of the art language modeling performance of 70.9 perplexity on the Penn Treebank dataset, while using far fewer parameters than a standard softmax LSTM.

Abstract: Disclosed are methods, systems, and computer program products for presenting data on a mobile device in communication with an on-demand database system. In some implementations, a mobile device receives information identifying one or more data objects stored in the on-demand database system based on captured behavioral data for a user of the mobile device. The captured behavioral data may be based on interactions with the one or more data objects. The mobile device provides the information for display in a user interface for a mobile application. The mobile device provides for display a subset of a plurality of information items in the user interface, the plurality of information items provided from the on-demand database system, the subset determined based on the identified one or more data objects and based on user input received via the user interface. In some implementations, the user input comprises a selected type of data object.

Abstract: An index to be backed up in a computing environment is identified. A staged copy of the index to be backed up is created. The staged copy of the index is segmented. The segments are uploaded to a pre-selected location.

Abstract: Techniques for partition balancing. Organization identifiers (orgIDs) are acquired for one or more organizations within a multi-organization on-demand services environment having multiple nodes, each of which have one or more partitions. A mapping of the orgIDs to the partitions within the multi-organization on-demand services environment is determined. Resource consumption for the organizations is determined. Resource consumption for the partitions based on the resource consumption of the organization and the mapping of the organization to the partitions is determined. A score for the partitions based on the resource consumption is generated. The partitions are redistributed among the nodes based on the partition scores.

Abstract: The technology disclosed preserves the tenant specificity and user specificity of the tenant data by associating user IDs to complementary special IDs referred to as the integration user(s). In particular, it combines the traceability of user actions, the integration of security models and the flexibility of a service ID into one integration user(s).

Abstract: In some implementations, a system causes a set of main tabs to be displayed at a first portion of a display. The system is configured to receive a user selection of a selected main tab. In response, the system causes a set of secondary tabs to be displayed at a second portion of the display. The system is configured to receive user input to manipulate the secondary tabs causing a rearrangement of the secondary tabs into a new order and/or causing one or more of the secondary tabs to be hidden. Based on the user input, the system causes the manipulated set of secondary tabs to be displayed and stores information defining the manipulated set of secondary tabs. Upon a subsequent access by the user of a database object associated with the selected main tab, the system causes the manipulated set of secondary tabs to be displayed.

Abstract: The technology disclosed provides a novel and innovative technique for compact deployment of application code to stream processing systems. In particular, the technology disclosed relates to obviating the need of accompanying application code with its dependencies during deployment (i.e., creating fat jars) by operating a stream processing system within a container defined over worker nodes of whole machines and initializing the worker nodes with precompiled dependency libraries having precompiled classes. Accordingly, the application code is deployed to the container without its dependencies, and, once deployed, the application code is linked with the locally stored precompiled dependencies at runtime. In implementations, the application code is deployed to the container running the stream processing system between 300 milliseconds and 6 seconds. This is drastically faster than existing deployment techniques that take anywhere between 5 to 15 minutes for deployment.

Abstract: In accordance with embodiments, there are provided mechanisms and methods for publicly providing web content of a tenant using a multi-tenant on-demand database service. These mechanisms and methods for publicly providing web content of a tenant using a multi-tenant on-demand database service can allow the web content to be published by a tenant using the multi-tenant on-demand database service for use by non-tenants of the multi-tenant on-demand database service.

Abstract: Methods and apparatus, including computer program products, implementing and using techniques for orchestrating a business process in a computer network. A script describing a business process is received from a client machine over the computer network. The script is compiled into executable program instructions. A reference to the script and data to be operated on by the script is received from the client machine over the computer network. The data is processed in accordance with the script. The result of the data processing is returned to the client machine over the computer network.

Abstract: Systems and methods are provided for managing license objects to applications in an application platform database system. The method includes associating an LMA with an application installed to the application platform by a developer, notifying a license manager to which the license manager application is installed of the installation of the application to the application platform, and managing subscriber access to the application using the license manager application.