Abstract: A system for tracking and representing data science model runs includes a hub including a first computing device communicatively coupled with a data store. A runner including a second computing device having a cache is communicatively coupled with the hub through a telecommunications network. An end user computing device includes a display and is communicatively coupled with the runner and the hub. User interfaces displayed on the display include: a unique identifier identifying a data science model run performed by the runner; a list of input files used by the runner to perform the run; a list of output files output by the runner as a result of the run; and a diagram diagramming a process flow including a visual representation of the input files, a visual representation of the run, and a visual representation of the output files.

Abstract: Automated assistant responses may be assembled and/or evaluated to address privacy concerns. In various implementations, a free-form natural language input may be received from a first user and may include a request for information pertaining to a second user. Multiple data sources may be identified that are accessible by an automated assistant to retrieve data associated with the second user. The multiple data sources may collectively include sufficient data to formulate a natural language response to the request. Respective privacy scores associated with the multiple data sources may be used to determine an aggregate privacy score associated with responding to the request. The natural language response may then be output at a client device operated by the first user in response to a determination that the aggregate privacy score associated with the natural language response satisfies a privacy criterion established for the second user with respect to the first user.

Abstract: Methods, systems and computer-program products are directed to a Privacy Engine for evaluating initial electronic documents to identify document content categories for portions of content within the electronic documents, with respect to extracted document structures and document positions, that may include privacy information for possible redaction via visual modification. The Privacy Engine builds a content profile based on detecting information at respective portions of electronic document content that indicate one or more pre-defined categories and/or sub-categories. For each respective portion of electronic document content, the Privacy Engine applies a machine learning model that corresponds with the indicated category (or categories and sub-categories) to determine a probability value of whether the respective portion of content includes data considered likely to be privacy information.

Abstract: A method of displaying web pages, capable of using the same settings for all web pages, is provided. The method includes receiving a user request for opening a web page; selecting, from among a plurality of views, a view in which the web page is to be opened, based on metadata corresponding to the web page, wherein each view of the plurality of views has a corresponding set of viewing properties; and displaying the web page in a browser window based on the corresponding set of viewing properties of the selected view.

Abstract: A digital communications or content platform is provided on a desktop and mobile platform that is configured to provide a useful tool for generating visualizations or other user content that is operable to create a layer or skin that acts as a pane of glass over original web content. The digital communications or content platform is configured to maintain, in a data structure, a set of data linkages representative of a set of visual components whose visual locations and website locations are dynamically maintained despite changes to the underlying webpage. The digital communications or content platform is configured to cause a display to render a hybrid webpage including both the underlying webpage and the overlay.

Abstract: A user interface (UI) enhancement system modifies a computer user interface based on user interactions recorded over time. The UI enhancement system records user interactions and develops a corresponding heat map. The heat map represents frequencies of user actions such as links clicked, scrolling actions, buttons selected, jumps within a page, etc. Based on the heat map, the UI enhancement system modifies a user interface configuration file which in turn influences the appearance and/or behavior of elements within the user interface to improve the user experience. Some embodiments give greater prominence to a UI element associated with greater “heat” and less prominence to an element with less “heat”.

Abstract: Systems and techniques may be used for providing website change detection. An example technique may include processing snapshots of a website, for example on a first day and a second day to produce a set of features of the website. The technique may include comparing at least one of the set of features from the first day to at least one of the set of features from the second day, for example using a distance function. A change score may be determined, the change score corresponding to at least one of a deleted feature, an added feature, or an unchanged feature of the website from the first day to the second day. The technique may include outputting the change score (e.g., for display).

Abstract: Users of a database management engine may generate fillable digital documents by mapping interface elements onto form documents. When a user maps interface elements onto a form document, the user may accidentally overlap two or more interface elements. To rectify this, the database management engine may modify the position of one of interface elements based on a set of positioning rules. In addition, the database management engine may identify and suggest mappings to users based on similar documents that have been previously mapped. The database management engine identifies similar documents using information about the document, the user, and the mapping itself. The mapping associated with the most similar document may be provided to the user as a suggested mapping. The database management engine converts the form document and finalized mapping into a fillable digital document. The fillable digital document is sent to recipients, who complete the fillable digital document.

Abstract: A system enables tracking machine learning (“ML”) model data provenance. In particular, a computing device is configured to accept ML model code that, when executed, instantiates and trains an ML model, to parse the ML model code into a workflow intermediate representation (WIR), to semantically annotate the WIR to provide an annotated WIR, and to identify, based on the annotated WIR and ML API corresponding to the ML model code, data from at least one data source that is relied upon by the ML model code when training the ML model. A WIR may be generated from an abstract syntax tree (AST) based on the ML model code, generating provenance relationships (PRs) based at least in part on relationships between nodes of the AST, wherein a PR comprises one or more input variables, an operation, a caller, and one or more output variables.

Abstract: A system and method are provided for automatically proofing the patent specification of a granted patent issued on a patent application. The method comprises the steps of scanning amendments made to the patent application during prosecution; assembling the most recent claims, based on the scanned amendments, to create a test claim set of the claims; comparing the test claim set to the claims of the issued patent as published; and identifying any differences between the compared sets of claims and communicating the differences to a user.

Abstract: A system that improves productivity and computer access to various network sites by providing users with a single stop destination for consumption of information that is spread across multiple network sites is described. A user interface (UI) view is customizable in the sense that each user is enabled to define the type of information to be displayed, as well as the way the information is organized. To do this, the system is configured to expose various options for configuring the UI view. The options enable a user to create a widget, which is a mechanism configured to retrieve data from a target location and display the data in a tile of the UI view. Accordingly, the UI view acts as a container for multiple widgets configured at the user level, thereby providing a one stop solution for the network information that is of continued interest to a particular user.

Abstract: A content analysis system includes processor and memory hardware storing data analyzed content items and instructions for execution by the processor hardware. The instructions include, in response to a first intermediate content item being analyzed to generate a first text description, receiving the first intermediate content item and analyzing the first text description to generate a first reduced text description. The instructions include identifying a first set of tags by applying a tag model to the first text description and generating a first analyzed content item. The instructions include adding the first analyzed content item to the analyzed content database and, in response to a displayed content item being associated with at least one tag of the first set of tags, displaying a first user-selectable link corresponding to the first analyzed content item on a portion of a user interface of a user device displaying the displayed content item.

Abstract: An online system presents content to its users, in which the content includes links that launch a messaging application and initiate conversations via the application. The system receives information indicating that negative experiences occurred during the conversations and may use this information and attributes of entities participating in the conversations to train a model to predict a likelihood of an occurrence of a negative experience. Upon determining an opportunity to present a user with a link that launches the application and initiates a potential conversation with an entity via the application, the system applies the model to predict a likelihood of an occurrence of the negative experience by the user during the potential conversation based on the entity's attributes. Based on the predicted likelihood, the system adjusts a value associated with presenting the link and passes the adjusted value to a process that selects content for presentation to the user.

Abstract: Technologies are provided for the generating of optimal policies for bidding in auctions having unknown dynamics. In some embodiments, a computing system can configure many multi-armed bandit (MAB) models defining candidate directed contents for a sequence of pages. A particular MAB model of the many MAB models defines candidate directed contents for a particular page in the sequence of pages, where each arm in the particular MAB model corresponds to a candidate impression on the particular page. The computing system can then determine a solution to an optimization problem with respect to an objective function based on an expected long-term reward for a defined impression on the first page, a defined impression on the second page, and a defined impression on the third page. The solution results in respective directed content for presentation on the first, second, and third pages.

Abstract: Aspects of the present disclosure involve a system comprising a computer-readable storage medium storing a program and method for determining zone types of a webpage. The program and method provide for generating, for at least one node a first webpage, a vector including HTML content corresponding to the at least one node; providing the vector as input to a machine learning model configured to output a predicted node type based on the vector, the machine learning model having been trained with plural vectors including HTML content corresponding to plural nodes of second webpages; and determining, based on the output of the machine learning model, the predicted node type of the vector, to classify the at least one node.

Abstract: Systems and methods for generating a website having a decoupled composition, layout, and processing are disclosed. A method may include an event creator computer program: (1) creating an event having a plurality of attendees; (2) creating a plurality of attendee types for the plurality of attendees; (3) generating a communication template for each of the plurality of attendee types; (4) receiving an event specific welcome page for each of the plurality of attendee types; (5) receiving a website configuration file for each of the plurality of the attendee types; and (6) uploading the event specific welcome page and the website configuration file for each attendee type to a library, wherein a web site host computer program links one of the plurality of welcome pages and the website configuration file to each of the attendee types.

Abstract: A computing system having at least one sensor, at least one processor, and at least one memory including a plurality of instructions stored thereon that, in response to execution by the at least one processor, causes the computing system to receive one or more surgical parameters associated with a ligament balancing of a patient's joint, receive real-time sensor data generated by the at least one sensor and indicative of at least one characteristic of the patient's joint, and apply machine learning to determine a next ligament balancing step of the ligament balancing of the patient's joint based on the one or more surgical parameters and the real-time sensor data, wherein the next ligament balancing step is a step of one or more steps intended to result in a target state of the patient's joint identified by the machine learning.

Abstract: Methods, systems, and computer storage media are provided for tracking an advertisement based on the advertisement's context. When an ad event is received on a client-computing device, a single URL is determined to display an item and track a context of the ad event. A first parameter related to the ad event is encoded as a HTTP header, and a second parameter related to tracking the ad event is encoded as a query parameter appended to the URL. The URL with the HTTP header is called, causing a domain server named in the URL to extract the first parameter from the HTTP header and the second parameter from the query parameter in order to determine the context of the ad. The domain server asynchronously requests tracking of the ad based on the context. Additionally, content for a landing page is received from the domain server.

Abstract: Accelerated machine learning using an efficient preconditioner for Kernel Ridge Regression (KRR). A plurality of anchor points may be selected by: projecting an initial kernel onto a random matrix in a lower dimensional space to generate a randomized decomposition of the initial kernel, permuting the randomized decomposition to reorder its columns and/or rows to approximate the initial kernel, and selecting anchor points representing a subset of the columns and/or rows based on their permuted order. A reduced-rank approximation kernel may be generated comprising the subset of columns and/or rows represented by the selected anchor points. A KRR system may be preconditioned using a preconditioner generated based on the reduced-rank approximation kernel. The preconditioned KRR system may be solved to train the machine learning model. This KRR technique may be executed without generating the KRR kernel, reducing processor and memory consumption.

Abstract: A first machine learning model processes a set of inputs to generate a first set of results. Based on that first set of results, a quality control range is calculated. A second machine learning model calculates a mean accuracy of a second set of results, based on the set of inputs. A determination of whether the mean accuracy of the second set of results is within the quality control range is made, and a user is notified of that determination.