Abstract: An application determines an emotional state of a user based on evaluating facial recognition data of the user captured from the user interacting with the application. The application receives a request from the user to initiate a support call. The request identifies the emotional state of the user. The application predicts, from a set of outcomes of support calls processed by support agents interacting with users having different emotional states, an emotional state that increases a likelihood of achieving a specified outcome for the support call based on the emotional state of the user. The application identifies an available support agent having the predicted emotional state, and assigns the user to interact with the identified support agent for the support call.

Abstract: A method and system parses natural language in a unique way, determining important words pertaining to a text corpus of a particular genre, such as tax preparation. Sentences extracted from instructions or forms pertaining to tax preparation, for example are parsed to determine word groups forming various parts of speech, and then are processed to exclude words on an exclusion list and word groups that don't meet predetermined criteria. From the resulting data, synonyms are replaced with a common functional operator and the resulting sentence text is analyzed against predetermined patterns to determine one or more functions to be used in a document preparation system.

Abstract: A system for account mapping includes functionality for obtaining more than one labeled accounts labeled by more than one accountant; pre-processing more than one labeled accounts using natural language processing, using the more than one pre-processed labeled accounts to train an account mapping model that performs multinomial classification; receiving an account name from an accounting application where the account name includes a text label for an account included in a chart of accounts; generating an account mapping by applying the account mapping model to the account name, where the account mapping includes a type of the account, a sub-type of the account, a code, and a series associated with an accounting form; returning the account mapping to the accounting application through an Application Programming Interface (API); and receiving a corrected account mapping from an accountant and using the corrected account mapping as a new text label to incrementally update the account mapping model.

Abstract: Disclosed methods and systems provide domain-specific incremental search results for search query terms with a customer self-service system for a financial management system, according to one embodiment. Incremental search results are search results that are generated for search query terms, as a user enters the search query terms as part of a search query to a customer self-service system, according to one embodiment. The customer self-service system receives search query terms (e.g., a partial search query) in a search text box, searches user content (e.g., business transactions, personal transactions, etc.) and customer self-service system content for content that is relevant to the search query terms, and provides incremental search results navigation links within a results text box to enable users to easily see search results (e.g., transaction records) and to enable users to navigate to transactions records or other user content, from the results text box, according to one embodiment.

Abstract: The present disclosure relates to processing domain-specific natural language commands. An example method generally includes receiving a natural language command. A command processor compares the received natural language command to a corpus of known commands to identify a probable matching command to the received natural language command, wherein the matching is based on an edition error and a recognition error. The corpus of known commands comprises a plurality of domain-specific commands, each of which is mapped to one or more domain-specific entities. Based on the comparison, the command processor identifies one or more entities in the received natural language command to perform an action on based on the mapping of the one or more domain-specific entities in the probable matching command and executes a domain-specific action included in the natural language command on the identified entity.

Abstract: Certain aspects of the present disclosure provide techniques for selecting a response to a self-support query. One example method generally includes receiving an audio stream query including spoken content from a user recorded by a mobile device and determining a set of paralinguistic features from the spoken content. The method further includes estimating an emotional state of the user based on the set of paralinguistic features and identifying subject matter of the spoken content in the audio stream query. The method further includes determining two or more query responses corresponding to the subject matter to present to the user and transmitting at least one query response to the mobile device.

Abstract: Techniques are disclosed for a bi-directional notification service. The techniques disclosed herein notify a user of an application in real-time based on a context of the application. A given notification may require a particular context before the notification can be presented to the user. The notification service also provides a channel for a user to provide feedback regarding the application. Multiple users may submit feedback messages each of which are enriched with contextual information related to the application at the time the message is submitted. Each feedback message is assigned to a category based on the content of the feedback message and contextual information of the notification. A heat map may be generated to visualize a number of feedback messages associated with one or more categories.

Abstract: A method may include generating a source transaction description, encoding, by an encoder model of a machine translation model executing on a computer processor, the source transaction description to create a context vector, decoding, by a decoder model of the machine translation model, the context vector to predict a target entity description, generating a transaction including the target entity description, detecting an acceptance, by a user, of an action performed on the transaction, in response to detecting the acceptance, updating a translation accuracy metric for the target entity description, determining that the updated translation accuracy metric satisfies a translation accuracy criterion, and in response to determining that the updated translation accuracy metric satisfies the translation accuracy criterion, adding the target entity description to golden entity descriptions.

Abstract: Field extraction from a form image includes identifying a target field of the form image, defining a patch from the form image based on the target field, and encoding the patch using a color encoding scheme to obtain an encoded patch. Field extraction further includes applying a trained classifier to the encoded patch to identify a relationship between a field value and a field identifier, and extracting the field value from the form image according to the relationship.

Abstract: Computer-implemented methods, systems and articles of manufacture detecting inconsistent electronic tax return data. In a tax return preparation application in which a tax logic agent and user interface controller share a data store, and the tax logic agent including or executing a cross-field validation engine is separate from or loosely coupled to a user interface controller, the tax logic agent reads runtime data of the electronic tax return from the data store, and the validation engine compares runtime data cross-field rule conditions. A cross-field rule having a condition that is not satisfied by the runtime data is identified, and a non-binding suggestion is generated by the tax logic agent and received by the user interface controller, which generates an interview screen concerning the detected inconsistency for display to the user.

Abstract: During an analysis technique, information associated with two entities may be compared to determine a similarity metric. For example, a string distance between the information may be computed. Then, an association metric between financial-transaction histories for the two entities may be calculated. This calculation may involve comparing the nodes and branches in a commerce graph that represents financial interrelationships among a set of entities, including inputs received by the set of entities, outputs provided by the set of entities, and financial transactions among the set of entities. Next, a determination of whether the entities are likely to be a same entity may be based on the similarity metric and/or the association metric. If the entities are likely to be the same entity, the entities may be combined in a data structure. Alternatively, if the entities are not likely to be the same, they may remain separate.

Abstract: Systems, methods and articles of manufacture for preparing a tax return in which tax questions may be generated and presented to a user in an ad hoc manner. A tax system includes a computing device executing a tax program. The tax program can display a plurality of user interface presentations, such as interview or question screens, with tax questions asking the user to provide tax data related to a taxpayer for preparing a tax return for the taxpayer. The tax program allows the user to skip a tax question on a user interface presentation, and proceed with subsequent user interface presentations. When the tax program determines that a response to the skipped tax questions is required, the tax program dynamically generates a skipped tax question user interface presentation with the skipped tax question and presents it to the user.

Abstract: Disclosed are techniques for automatic generation of software application code. These techniques identify or generate a data model in a first programming language. At least one rule is identified or generated from the data model. This at least one rule is tokenized into a plurality of subparts based at least in part upon a second programming language. A rule hierarchy at which the plurality of sub-parts is located is determined based in part or in whole upon a corresponding class hierarchy of a class with which the at least one rule is associated. A transformed data model is generated in the second programming language using at least the plurality of sub-parts and the rule hierarchy.

Abstract: This disclosure relates to cross-platform applications that include native and non-native components on mobile devices. An exemplary method generally includes receiving a first workflow step definition including a first set of widgets to be loaded into an application shell. A mobile shell identifies a type of each widget in the first set of widgets (e.g., native or platform-agnostic) and loads each widget into the mobile shell based on the widget type. For a platform-agnostic widget, the mobile shell creates a platform-agnostic widget proxy service, which provides a runtime environment. The platform-agnostic widget may be loaded into the platform-agnostic widget proxy service and executes in the runtime provided thereby.

Abstract: Systems of the present disclosure provide a versatile, reusable mock server to respond to Application-Programming-Interface (API) requests. The mock server receives an API request and a cookie associated with the API request. The API server identifies response instructions found in the cookie. The response instructions may include a static response value, a name of an API server for the mock server to imitate, or code for the mock server to execute in the process of generating a mock API response. The mock server generates a mock API response based on the response instructions and sends the mock API response in reply to the API request.

Abstract: Graphical user interfaces or visual representations of budgets and related systems and methods for visually presenting a budget and how consumer spending adheres to the budget. Budget elements or panels relate to budget categories, and respective sizes of budget elements indicate respective portions of income allocated to corresponding budget categories. Income amounts or desired savings amounts can be adjusted, in response to which sizes of the selected budget and savings elements are changed. After establishing a budget, the consumer purchases items. Corresponding budget elements are filled to indicate respective allocated amounts of income spent on respective budget elements by the consumer. An empty budget element indicates no money has been spent, a partially filled budget element indicates that some of the allocated income was spent, and a completely filled budget element indicates that all of the allocated income was spent. Overspending may be indicated by emphasizing the budget element.

Abstract: This disclosure relates to increasing database functionality in non-relational databases using an object relational mapper to provide updates to table field values across a plurality of separate tables and provide virtual table functionality through compound search operations and condensed table structures.

Abstract: The present disclosure relates to processing domain-specific natural language commands. An example method generally includes receiving a natural language command. A command processor compares the received natural language command to a corpus of known commands to identify a probable matching command in the corpus of known commands to the received natural language command. The corpus of known commands comprises a plurality of domain-specific commands, each of which is mapped to a domain-specific action. Based on the comparison, the command processor identifies the domain-specific action associated with the probable matching command to perform in response to the received command and executes the identified domain-specific action.

Abstract: During an image-analysis technique, the system calculates features by performing image analysis (such as optical character recognition) on a received image of a document. Using these features, as well as spatial and grammatical constraints, the system determines a layout of the document. For example, the layout may be determined using constraint-based optimization based on the spatial and the grammatical constraints. Note that the layout specifies locations of content in the document, and may be used to subsequently extract the content from the image and/or to allow a user to provide feedback on the extracted content by presenting the extracted content to the user in a context (i.e., the determined layout) that is familiar to the user.

Abstract: Embodiments presented herein provide systems and methods for creating and modifying a hash chain. A hash chain is created to track resource-privilege transfers between entities. A root node of the hash chain identifies the resource and specifies a digest of a possession token held by a first entity that initially possesses the privilege. A transfer of the privilege to a second entity can be recorded by adding an expansion node to the hash chain. If the second entity successfully reveals a possession token that a hashing function associated with the hash chain maps to the digest, an expansion node is linked to the root node. The expansion node indicates the possession token and a successor digest that is based on a successor possession token.