Groupon, Inc.

Amazon Technologies, Inc.

Amber Roy Chowdhury

Sri Subramaniam

Vijay A Kumar

Abhaya Parthy

Boris Lerner

Emma Sawin

Daniel M Roarty

Tarun Bhatia

Valerie Galluzzi Liptak

Lawrence Lee Wai

Roger Henry Castillo

Bhupesh Bansal

Greyson Gregory

Sridatta Viswanath

Ming-Chi Tsai

Alexander Y Kalinin

Soo-Min Pantel

Saurabh Sohoney

Hassan Haider Malik

Karthik Ram Srinivasan

Sanjay Kumar

Sourav Kumar Agarwal

Anand Vishwanath Suvarnkar

Pawel Matykiewicz

Sayan Putatunda

Govind

Tong Zhao

Guokai Zeng

Lei Tang

Ming-Jun Chen

Ravi Garg

George Henry Forman

Sihui Zhang

Steven Casey

Uday S Sandhar

Ming Ma

Techniques for improving last mile transportation and planning execution are described. In an example, a description of an address is received. Whether the address is not associated with a known geocode is determined. Using an ML model, an embedding vector of the address in an embedding space is generated, the ML model trained to embed addresses and geocodes in the embedding space. Based on the embedding vector, one or more known geocodes is determined. The address is then associated with the one or more known geocodes.

Machine learning model-based resolution of location information

Techniques for triggering a mitigation action based on an artificial intelligence (AI) model-based delivery defect prediction are described. In an example, a computer system determines contextual information associated with a stage of a delivery flow of an item to a delivery location. The computer system generates an input based at least in part on the contextual information and provides the input to an AI model trained to predict delivery defect information at delivery flow stages. The computer system determines, based at least in part on an output of the artificial intelligence model, a predicted delivery defect associated with the stage of the delivery flow. The computer system determining a mitigation action based at least in part on the predicted delivery defect and causing the mitigation action to be performed.

Triggering a mitigation action based on an artificial intelligence model-based delivery defect prediction

Techniques are described for predicting a future defect and triggering a mitigation action. In an example, a system generates a first input to a causal model based on geographical data associated with a location and area data associated with one or more areas that include the location, and determines a first output of the causal model. The first output indicates a first prediction of a cause of a past defect for a service associated with the location. The system also generates a second input to a predictive model based on the first input and the first output, and determines a second output of the predictive model. The second output indicates a second prediction of a future defect associated with the service. Based on the second prediction, a mitigation action can be performed such that the future defect is prevented.

Triggering a service mitigation action based on a causal-predictive system

Devices and techniques are generally described for machine learning personalization as a service for speech processing applications. In various examples, a first request for machine learning prediction for a first speech processing skill. First skill data schema data may be received that describes content of the first speech processing skill. A first machine learning model for the first speech processing skill may be determined. A first feature definition describing a first aspect of the content may be determined. A second feature definition describing user profile data may be determined. A prediction request may be received from the first speech processing skill. First feature data may be generated according to the first feature definition and second feature data may be generated according to the second feature definition based at least in part on the prediction request. The first machine learning model may generate prediction data based at least in part on the prediction request, the first feature data, and the second feature data.

Personalization for speech processing applications

Techniques for a geofence generation feature are described herein. Geospatial vector data for a geographical area and seed points for delivery locations within the geographical area may be obtained. A plurality of polygons that represent the geographical area may be determined based on an algorithm that uses the seed points and the geospatial vector data. A boundary point for each polygon may be determined based on coordinates for the plurality of polygons where each boundary point corresponds to a single seed point. A geofence of a first size may be generated, for each polygon, based on the coordinates for the polygons, the seed point, and the boundary point. Historical scans of GPS data may be obtained from user devices during historical visits to the geographical area. The geofence may be modified to a second size or a different shape based on the historical scans of the GPS data.

Neighborhood aware geofence generation

In general, embodiments of the present invention provide systems, methods and computer readable media for dynamically augmenting relevance rankings using data from external ratings sources.

Dynamic augmenting relevance rankings using data from external ratings sources

Techniques for integrating freshness and relevance within a programmatic relevance service are discussed herein. Some embodiments may include a relevance service system, including one or more batch processors and one or more online processors. The one or more batch processors may be configured to track message recipient behavior from electronic clickstream data, and generate models of functions that predict access probability for individual users. The one or more online processors may use the access probability calculated in batch to facilitate on demand deal ranking adjustments that account for freshness as informed by the predicted access probability for communications sent to the user.

Dynamic freshness for relevance rankings

In general, embodiments of the present invention provide systems, methods and computer readable media for a universal relevance service framework for ranking and personalizing items.

Ranked relevance results using multi-feature scoring returned from a universal relevance service framework

Enforcing diversity in ranked relevance results returned from a universal relevance service framework

An apparatus, method, and computer program product are provided to filter and modify option data objects and weighted values associated with option data objects through the application of specific rule sets based on the relative density of option data objects within a particularized area. In some example implementations, option data objects and related parameters are parsed to identify locations associated with the option data object and a weighted value, such as a weighted value generated by a predictive model. Based at least in part on the location associated with the option data object, a determined location of a user of a mobile device, and location-specific distance criteria, the weighted value associated with the option data object may be modified to reflect distance-related option election probabilities.

Apparatus and method for distance-based option data object filtering and modification

Systems and related methods of providing promotional offers to consumers are discussed herein. Some embodiments may provide for an apparatus including circuitry configured to provide promotional offers to consumers based on dimensions representing criteria by which promotions may be deemed relevant to a consumer. Some examples of dimensions may include location, time, environment, price, and/or consumer preference. Based on receiving signals from the consumer device, among other sources, indicating associated times, locations, and other characteristics of consumer activity, the apparatus may recognize patterns or trends in consumer behavior, and use such information to predict or influence future consumer behavior.

Systems and methods for providing dimensional promotional offers

Universal relevance service framework

Growing community of inventors

Bellevue, WA, United States of America

Amber Roy Chowdhury