Other

Apple Inc.

Yahoo! Inc.

Subash Sundaresan

Ramesh R Sarukkai

Thomas M Alsina

David T Wilson

Sagar A Joshi

Sam P Hamilton

Shubhasheesh Anand

Julie Wan-Chi Cheng

P Sam Hamilton

George R Dicker

Michael R Siracusa

Ryan S Mehlmauer

Todd Michael Teresi

Chandrasekaran Jagadeesan

Julie W Cheng

Kuriakose Sony Theakanath

Anthony J Tarlano

Nikhil A Desai

A method for sharing an incrementally trained machine learning (ML) model from an edge device to other edge devices that are associated with a peer to peer network is provided. The method includes (i) receiving event data at an edge device from among one or more edge devices of the peer to peer network upon the occurrence of an event at the edge device, (ii) incrementally training a base version of a ML model at the edge device based on the received event data, (iii) generating, a unique persistent file format of the incrementally trained ML model at the edge device, (iv) optimizing a payload associated with the unique persistent file format to include one or more parameters with associated weights changing beyond a predetermined configurable threshold and (v) encapsulating the unique persistent file format with a unique metadata.

System and method for sharing an incrementally trained machine learning (ML) model from an edge device to one or more other edge devices in a peer to peer network

A method and system for incremental training of machine learning (ML) models on edge devices, is disclosed. A base version of ML model is received by a first device of the peer to peer network of devices. The base version of the ML model is incremental trained in real-time by updating weights associated with the parameters of the ML model during a predetermined window of time. The ML model are devoid of an underlying event data used to generate the incremental version of the ML model. The updated weights associated with parameters of the incrementally trained ML model and/or the respective parameters are shared by each edge device with other edge devices. The edge devices update their respective ML models based on the received updated weights and/or parameters upon subsequent events at the edge devices and the updated version of the ML models are further incrementally trained.

Method and system for incremental training of machine learning models on edge devices

A method and a system for maintaining network integrity for incrementally training machine learning (ML) models at edge devices is provided. The method includes registering, by a certifying node, one or more edge devices with a peer to peer network. Upon registration, an incrementally updated ML model is received from a first registered device at the certifying node. The certifying node accepts the incrementally updated ML model if a contribution of the first edge device is within a predetermined threshold, and else rejects the updated ML model if the contribution is beyond the predetermined threshold. Limiting the contribution by each edge device enables prevention of skew by any of the edge devices at the certifying node. Upon accepting the updated ML model, the certifying node certifies the updated ML model and transfers the certified ML model to one or more other edge devices in the peer to peer network.

System and method for maintaining network integrity for incrementally training machine learning models at edge devices of a peer to peer network

A method for concurrent training and updating a machine learning (ML) model at an edge node in a peer to peer network using a certifying node is provided. The method includes performing an incremental training of the ML model by a first edge node based on one or more data points associated with a data event. The first edge node ML transfers the incrementally trained ML model to the certifying node. While awaiting a response form the certifying node, the first edge node receives a certified ML model from the certifying node. The certified ML model originates from a second edge node. The first edge node performs an incremental training of the received certified ML model based on one or more data points by re-applying the data points associated with the event data. The first edge node transfers the incrementally trained ML model to the certifying node for certification.

System and method for concurrent training and updating of machine learning models at edge nodes in a peer to peer network

There is provided a method of operating a certifying node to certify incremental trained machine learning (ML) models of one or more edge devices associated with a peer to peer network. The method includes (i) generating a predictive outcome value for a test data set by executing a candidate ML model against the test data set available to the certifying node; (ii) determine a measure of quality of the candidate ML model by matching the predictive outcome value of the candidate ML model with an actual outcome value of the test data set; and (iii) certify the candidate ML model by comparing the measure of quality of the candidate ML model against a threshold error value, for use in real time incremental training by the one or more edge devices of the peer to peer network.

System and method of certification for incremental training of machine learning models at edge devices in a peer to peer network

Embodiments described herein provide a predictive failure analysis that enables design-time error and exception handling techniques to be supplemented or assisted by a predictive failure analysis system. One embodiment provides an electronic device, comprising a non-transitory machine-readable medium to store instructions; one or more processors to execute the instructions; and a memory coupled to the one or more processors, the memory to store the instructions which, when executed by the one or more processors, cause the one or more processors to receive injection of dynamic error detection logic into the instructions, the dynamic error handling logic including an error handling update to indicate a response to a predicted failure; receive a set of events indicative of the predicted failure; and respond to the set of events according to the error handling update.

Method and system for predicting failure events

Systems, methods, and computer-readable storage media for targeting segments of customers. A system maps users to targeting segments based on a respective purchase history of the users, and receives a segment identifier from a server, the segment identifier to be associated with a targeting segment from the targeting segments for anonymously identifying the targeting segment. Next, the system associates the segment identifier with the targeting segment and users mapped to the targeting segment, and sends, to the server, the segment identifier and unique user identifiers of the users mapped to the targeting segment associated with the segment identifier, the unique user identifiers masking respective identities of the users from the server and the segment identifier masking an identity of the targeting segment associated with the segment identifier. The system then instructs the server to deliver selected invitational content to at least one unique user identifier associated with the segment identifier.

Targeting customer segments

Systems, methods, and computer-readable medium are provided for managing electronic content requests. For example, a user device may collect usage information associated with use of the user device. The user device may transmit data corresponding to a first category of usage information to a service provider. In some examples, a file that comprises a plurality of electronic content items can be received. The file may also contain logic for determining which of the content items to display. The logic may be executed to determine a particular item of the content items to display based at least in part on additional data corresponding to a second category of the usage information. In some examples, the user device may prepare the particular content item for display.

Electronic content requests

Methods and arrangements allow merchants to set up advertising campaigns that distribute special offers or discounts and allows users to forward the special offers to friends. The forwarding of the offers could expose merchants to offering more discounts than budgeted and this is a concern. The invention addresses this by having the forwarded offer check with the merchant server or advertising server to see if there is enough budget in the campaign left to allow the issuance of the forwarded offer, and if so, the server issues a new identifier. If not, the user receiving the forwarded offer receives a message saying the campaign has expired. Since every offer has its own identifier, the forwarding of the offers can be tracked, as well as the redemption, which allows for social metric collection.

Social referrals of promotional content

A system is described for targeting data to users on mobile devices. The system may include a processor, a memory and an interface. The memory may be operatively connected to the processor and the interface and may store a request, a mobile identifier, a session identifier, a user behavior dataset and a data. The interface may be operatively connected to the memory and the processor and may communicate with the mobile device. The processor may be operatively connected to the interface and the memory and may receive a request from a user on a mobile device via the interface. The processor may extract a mobile identifier from the request and map the mobile identifier to a session identifier. The processor may use the session identifier to collect a user behavior dataset and may communicate data to the user based on the user behavior dataset via the interface.

System for targeting data to users on mobile devices

A server for delivering advertisements to mobile devices includes an application programming interface executable by a processor and configured to: receive an ad call over a network from an application of a mobile device to deliver an advertisement to an ad position on a small screen thereof, the ad position located at a fixed location of the application, the fixed location having a fixed height and a fixed width as viewed by a user of the mobile device throughout use of the application; detect the ad position, including the fixed height and width of the ad position within the fixed location; retrieve from the storage memory an advertisement having an ad format best matching the fixed height and the fixed width for filling the ad call at the fixed location; and deliver the advertisement to the mobile device for display in the ad position during the execution of the application.

Advertisements on mobile devices using integrations with mobile applications

A system is described for serving targeted advertisements over mobile messaging services. The system may include a processor, a memory and an interface. The memory may store a request for an advertisement and an advertisement. The interface may be operatively connected to the memory and the processor and may communicate with mobile network operators. The mobile network operators may be capable of providing mobile services users. The processor may be operatively connected to the interface and the memory and may receive a request for an advertisement from a mobile network operator via the interface. The request for an advertisement may be related to a mobile message sent from a first user to a second user via the mobile network operator. The processor may retrieve an advertisement based on the request for an advertisement. The processor may communicate the advertisement to the mobile network operator via the interface.

System for serving targeted advertisements over mobile messaging services

A method for delivering advertisements to mobile devices includes integrating an ad position within an application executable from a mobile device, wherein the ad position remains substantially constant in a location on a screen of the mobile device as viewed by a user throughout use of the application; including an ad call associated with the ad position; in response to executing the application, causing the ad call to retrieve an advertisement from an ad server; and receiving the advertisement into the ad position by the mobile device for display to the user.

Growing community of inventors

Fremont, CA, United States of America

Subash Sundaresan