Amazon Technologies, Inc.

Canku Alp Calargun

Prajwal Yadapadithaya

Gustavo Eduardo Lopez

Zhaoqing Ma

Ilya Levner

Chih-Jen Huang

Scott Kerns Windsor

Siddharth Vivek Joshi

Brett E Lounsbury

James Robert Blair

Scott Southwood

Venkata Krishnan Ramamoorthy

Konstantinos Boulis

Gurbinder Gill

Vishal Bethur Sathyamurthy

Dharmil Satishbhai Chandarana

David Aronchick

Hyma Srinivasa Murthy

Riley James Williams

Ankit Garg

Vivek Chaganti

Prama Anand

Various embodiments provide processes for identification of embedded unauthorized data within document data. Such processes may include evaluating document data properties for indications of closing headers, comparing document data sizes against reference sizes, and comparing document data sizes against other document data sizes. Certain embodiments evaluating image data may evaluate image properties, such as histograms or gradient images, to determine random or otherwise unauthorized data. Upon detection of unauthorized data, the document data may be flagged for later processing, such as removal from a storage server.

System and method for evaluating data files

Techniques for improved image classification are provided. Face embeddings are generated for each face depicted in a collection of images, and the face embeddings are clustered based on the individual whose face is depicted. Based on these clusters, each embedding is assigned a label reflecting the cluster assignments. Some or all of the face embeddings are then used to train a classifier model to generate cluster labels for new input images. This classifier model can then be used to process new images in an efficient manner, and classify them into appropriate clusters.

Incremental clustering for face recognition systems

Categorizing images may include training a first neural network to cluster a plurality of images to obtain a first image embedding space, wherein a vector representation is determined for each of the plurality of images based on the training, determining a vector norm value corresponding to each of the plurality of images based on the vector representation for each of the plurality of images, and identifying a first subset of the images for which a corresponding vector norm value satisfies a predetermined vector norm quality threshold. Then, a second neural network may be trained using the first subset of images to obtain a second image embedding space, and the second image embedding space may be used to categorize additional images.

Vector norm algorithmic subsystems for improving clustering solutions

Customers may place orders for items for delivery, where merchants provide the items, and a service provider receives the orders and facilitates the delivery of the items. Upon receiving multiple orders, the orders may be prioritized based on users scores associated with users that placed the orders. An order objective for the orders may be determined and multiple fulfillment plans for each order may be generated. For a particular order, a fulfillment plan of the multiple fulfillment plans may be selected based on the order objective for that order, where the selected fulfillment plan may govern how and when the items included in the order are to be delivered to the user that placed the order. The service provider may cause the items associated with the order to be delivered to the user pursuant to the selected fulfillment plan.

System and methods for prioritization of items for delivery

Tracking waiting lists associated with merchants and presenting deals to users based on adding the users to waiting lists is described. A service provider may enable merchants to track waiting lists associated with their own services and services of other merchants. Users may be automatically added to waiting lists of merchants based at least in part on a service provider determining that a user is at a geographic location of a merchant. Based at least in part on adding a user to a waiting list for a particular merchant, the service provider may access and select deals from one or more alternative merchants. The one or more alternative merchants may not have waiting lists or may have waiting lists with fewer groups than the waiting list for the particular merchant. The service provider may offer the deals to the user via a user device.

Computerized waiting list tracking system

A meal ordering service is configured to receive takeout and delivery orders from customers and to place the orders with selected restaurants. Past information regarding a customer is collected and analyzed to create a predictive model of the customer's meal ordering habits. On any given day, the predictive model may be used, in light of the current and scheduled status of the customer, to predict a time of day that the customer may want to receive a meal and to also determine a restaurant from which the customer may want to receive the meal. The meal ordering service may be configured to automatically order a meal for the customer from the determined restaurant so that the meal arrives at or before the predicted time of day.

Predictive restaurant ordering

Growing community of inventors

Redmond, WA, United States of America

Canku Alp Calargun