Abstract: A continuous meta-learning system operating at a server is described. In one aspect, a computer-implemented method includes accessing, at the server, a deployed machine learning model, metadata, and new prediction data for the deployed machine learning model, detecting a novel trend in the deployed machine learning model based on the new prediction data, generating label suggestions for the novel trend using metadata, querying a plurality of users to verify the label suggestions, detecting a consensus of the plurality of users, the consensus based on a rate of responses from the plurality of users that confirm the label suggestions, and in response to detecting the consensus, training a new machine learning model based on the new prediction data and the consensus of the plurality of users. The computer-implemented method also includes deploying the new machine learning model at the server.

Abstract: A system for private and secure data portal is described. A method includes receiving, from a first client device, a permission request and a data request for a first dataset that is stored at a second client device, providing the permission request and the data request to the second client device, the second client device configured to generate, in response to the permission request and the data request, a data usage approval document and an encrypted synthesized dataset corresponding to the data usage approval document, the encrypted synthesized dataset includes a synthetic second dataset representative of the first dataset, receiving, from the second client device, the data usage approval document and the encrypted synthesized dataset, performing, at a server, a computation on the encrypted synthesized dataset based on the data request, and providing the data usage approval document and results of the computation to the first client device.

Abstract: A post-deployment drift detection monitoring of a predictive model is described. The method includes accessing a predictive performance metric of a machine learning model that is deployed at a server, the machine learning model being trained with an initial set of training data containing historical data, the predictive performance metric being based on the initial set of training data and an additional set of training data, the additional set of training data containing training data collected since training the machine learning model, detecting a drift based on the predictive performance metric exceeding a drift detection threshold, generating a drift warning notification to a client device, the drift warning notification indicating that the predictive performance metric exceeds the drift detection threshold, receiving a user feedback from the client device, and adjusting one of the machine learning model or the drift detection threshold based on the user feedback.

Abstract: A machine learning platform operating at a server is described. The machine learning platform accesses a dataset from a datastore. A task that identifies a target of a machine learning algorithm from the machine learning platform is defined. The machine learning algorithm forms a machine learning model based on the dataset and the task. The machine learning platform deploys the machine learning model and monitors a performance of the machine learning model after deployment. The machine learning platform updates the machine learning model based on the monitoring.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for utilizing machine-learning models to determine take-over scores and intelligently provide or limit access to account features. In particular, in one or more embodiments, the disclosed systems can train and utilize digital security machine-learning models to generate a take-over score indicating a likelihood that the request to access the secure digital account is unauthorized activity. Based on the determining that the take-over score satisfies a take-over threshold, the disclosed systems can allow access to the secure digital account by providing secure account information but prohibit access to a subset of account features.

Abstract: A machine learning platform operating at a server is described. The machine learning platform accesses a dataset from a datastore. A task that identifies a target of a machine learning algorithm from the machine learning platform is defined. The machine learning algorithm forms a machine learning model based on the dataset and the task. The machine learning platform deploys the machine learning model and monitors a performance of the machine learning model after deployment. The machine learning platform updates the machine learning model based on the monitoring.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for assigning a limited access state to an account based on a triggering event and causing a client device to provide full navigable access to a client application. For example, in one or more embodiments, the disclosed system detects a triggering event corresponding with activity of an account. In particular, based on the detected triggering event, the disclosed system assigns the limited access state to the account and provides a notification to the account in the limited access state. Furthermore, while in the limited access state the client device associated with the account in the limited access state provides full navigable access to the client application, detects restricted activity based on restriction rules, and provides graphical elements to restore the account from the limited access state to a full access state.

Abstract: A head-mounted device (HMD) of a first user has a transparent display. The HMD determines location information of a second user relative to the HMD of the first user. The second user is located within a predefined distance of the HMD. The location information identifies a distance and a direction of the second user relative to the HMD. The HMD receives audio content from the second user, generates augmented reality (AR) content based on the audio content, and displays the AR content in the transparent display based on the location information of the second user. The AR content appears coupled to the second user.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for utilizing machine-learning models to determine take-over scores and intelligently provide or limit access to account features. In particular, in one or more embodiments, the disclosed systems can train and utilize digital security machine-learning models to generate a take-over score indicating a likelihood that the request to access the secure digital account is unauthorized activity. Based on the determining that the take-over score satisfies a take-over threshold, the disclosed systems can allow access to the secure digital account by providing secure account information but prohibit access to a subset of account features.

Abstract: A method, apparatus and computer program product are provided for calculating closing metrics regarding a contract between a promotion service and provider. A promotional system may calculate a probability of closing, and an estimated time to close. The promotion service may offer a promotion to consumers for a discounted product or service, to be honored by the provider. A category, lead source, historical data, stage in sales, and/or size of the provider may be used in calculating a probability of close and/or time to close. An example method may comprise supplying a classifying model with a dataset, wherein the dataset comprises an identification of a provider and attributes corresponding to the provider and identifying a class of the provider in accordance with the plurality of corresponding attributes, wherein the identification is determined based on one or more patterns determinative of a return rate by the classifying model.

Abstract: A machine learning platform operating at a server is described. The machine learning platform accesses a dataset from a datastore. A task that identifies a target of a machine learning algorithm from the machine learning platform is defined. The machine learning algorithm forms a machine learning model based on the dataset and the task. The machine learning platform deploys the machine learning model and monitors a performance of the machine learning model after deployment. The machine learning platform updates the machine learning model based on the monitoring.

Abstract: A mountable support and vice device for field use in small work areas includes a fastener, which is attached to a first terminus of a rod. The fastener is selectively attachable to an element of a structure, such as a scaffold, to removably attach the rod to the element of the structure. A pipe is nested around the rod so that the pipe is selectively extensible from, and selectively rotatable relative to, the rod. A coupler attached to the pipe and is positioned to selectively couple to the rod so that the pipe is fixedly positioned relative to the rod and the element of the structure. A vice is attached to the pipe proximate to a first end of the pipe and can selectively grip a substrate, such as a pipe elbow. The substrate thus is fixedly positioned and can be manipulated by a user.

Abstract: A method, apparatus and computer program product are provided for calculating closing metrics regarding a contract between a promotion service and provider. A promotional system may calculate a probability of closing, and an estimated time to close. The promotion service may offer a promotion to consumers for a discounted product or service, to be honored by the provider. A category, lead source, historical data, stage in sales, and/or size of the provider may be used in calculating a probability of close and/or time to close. An example method may comprise supplying a classifying model with a dataset, wherein the dataset comprises an identification of a provider and attributes corresponding to the provider and identifying a class of the provider in accordance with the plurality of corresponding attributes, wherein the identification is determined based on one or more patterns determinative of a return rate by the classifying model.

Abstract: Intelligent application of reserves to transactions is described. In an example, server(s) associated with a payment processing service receives first transaction data associated with a plurality of first transactions, wherein the first transaction data includes indications, for an individual first transaction, of whether a chargeback request occurred and item(s) and a customer associated with the transaction. In an example, the servers process the plurality of first transactions. A predictive model is trained based at least in part on the first transaction data. Second transaction data is received from a POS device of a second merchant associated with a second transaction. A value reflective of a level of risk for the second transaction is determined based on inputting an item or customer associated with the second transaction into the predictive model.

Abstract: Systems and methods for generation and aggregation of point-of-sale account data and generation of a marketplace platform are disclosed. Transaction information may be utilized to value one or more businesses, and those valuations may be presented to potential buyers in a generated marketplace platform. A remote system may facilitate a sales agreement between a buying merchant and a seller merchant by, for example, transferring funds from a remote system account to the selling merchant's account and accepting repayment from the buying merchant over time. Merging of transaction information and/or user accounts pursuant to the sales agreement may also be performed.

Abstract: Intelligent application of reserves to transactions are described. In an example, server(s) associated with a payment processing service receive, from a point-of-sale (POS) device operated by a merchant, transaction data associated with a payment transaction between the merchant and a customer. Based on a predictive model, the server(s) can determine a level of risk associated with the merchant and/or the payment transaction and can determine a portion of the transaction to withhold from a settlement amount of the payment transaction based at least in part on the level of risk. The portion can be deposited into a reserves account associated with the payment processing service to satisfy costs associated with chargeback requests.

Abstract: A system and method for generating a virtual content to a physical object is described. The system accesses a video that was captured by a camera. The video depicts a user performing a physical operation in relation to a physical object. The system determines metadata identifying positions of the camera relative to the physical object during the capturing of the video, identifies feature points of the physical object in the video, and retrieves a virtual object based on the feature points of the physical object. The system generates virtual content based on the video, the metadata, and the virtual object, and stores the object identifier and the virtual content relationally in a data storage that stores multiple sets of virtual content associated with multiple physical objects and their corresponding object identifiers.

Abstract: A system and method for generating a virtual content to a physical object is described. A processor includes an augmented reality application. The augmented reality application creates virtual content at the head mounted device, and associates the virtual content with predefined conditions based on data from sensors embedded in the head mounted device at a time of creation of the virtual content. The virtual content is displayed in a display of the head mounted device in response to sensor data satisfying the predefined conditions.

Abstract: A method, apparatus and computer program product are provided for calculating closing metrics regarding a contract between a promotion service and provider. A promotional system may calculate a probability of closing, and an estimated time to close. The promotion service may offer a promotion to consumers for a discounted product or service, to be honored by the provider. A category, lead source, historical data, stage in sales, and/or size of the provider may be used in calculating a probability of close and/or time to close. An example method may comprise supplying a classifying model with a dataset, wherein the dataset comprises an identification of a provider and attributes corresponding to the provider and identifying a class of the provider in accordance with the plurality of corresponding attributes, wherein the identification is determined based on one or more patterns determinative of a return rate by the classifying model.

Abstract: The disclosure describes embodiments of systems, methods, and non-transitory computer readable storage media that utilize a transaction request fraud detection model with a harvesting threshold to detect account number harvesting from network transaction requests. For example, the disclosed systems can identify network transaction requests from a transaction facilitator computing device and transaction request response codes for the network transaction requests. Subsequently, utilizing a transaction request fraud detection model, the disclosed systems can determine that the number of network transaction requests that include declined transaction request response codes satisfies a harvesting threshold indicating an account number harvesting event. In response to the indicated account number harvesting, the disclosed systems can send a selected transaction request response for responses to additional declined network transaction requests instead of one or more original transaction request response codes.