公开(公告)号：US10599449B1

公开(公告)日：2020-03-24

申请号：US15389102

申请日：2016-12-22

Applicant: Amazon Technologies, Inc.

Inventor： Nikolaos Chatzipanagiotis ,  Pragyana K. Mishra ,  Roopesh Ranjan

IPC: G06F9/451 ,  G06F17/22 ,  G06N5/02 ,  G06N3/04 ,  G06Q30/06

Abstract: A prediction model may be created to predict future actions likely to be performed by users while interacting with electronic content via user devices. The predictions may be used to streamline access to interface controls or other information to enable the users to facilitate or expedite performance of the predicted actions, while reducing computational demands on computing devices that provide the electronic content by, for example, reducing unnecessary intervening computing actions.

公开(公告)号：US11556773B1

公开(公告)日：2023-01-17

申请号：US16024434

申请日：2018-06-29

Applicant: Amazon Technologies, Inc.

Inventor： Shikha Aggarwal ,  Nikolaos Chatzipanagiotis ,  Shivani Matta ,  Pragyana K. Mishra ,  Anil Padia ,  Nikhil Raina

IPC: G06N3/08 ,  G06N7/00

Abstract: Aspects of the present disclosure relate to machine learning techniques for identifying the incremental impact of different past events on the likelihood that a target outcome will occur. The technology can use a recurrent neural network to analyze two different representations of an event sequence—one in which some particular event occurs, and another in which that particular event does not occur. The incremental impact of that particular event can be determined based on the calculated difference between the probabilities of the target outcome occurring after these two sequences.

公开(公告)号：US11468348B1

公开(公告)日：2022-10-11

申请号：US16788100

申请日：2020-02-11

Applicant: Amazon Technologies, Inc.

Inventor： Can Cui ,  Nikolaos Chatzipanagiotis ,  Tamal Krishna Kuila ,  Narayanan Sadagopan

IPC: G06N20/00 ,  G06N5/04 ,  G06K9/62 ,  G06N3/04

Abstract: Methods and apparatus for identifying features that may have a high potential impact on key application metrics. These methods rely on observational data to estimate the importance of application features, and use causal inference tools such as Double Machine Learning (double ML) or Recurrent Neural Networks (RNN) to estimate the impacts of treatment features on key metrics. These methods may allow developers to estimate the effectiveness of features without running online experiments. These methods may, for example, be used to effectively plan and prioritize online experiments. Results of the online experiments may be used to optimize key metrics of mobile applications, web applications, websites, and other web-based programs.

公开(公告)号：US11048530B1

公开(公告)日：2021-06-29

申请号：US16827471

申请日：2020-03-23

Applicant: Amazon Technologies, Inc.

Inventor： Nikolaos Chatzipanagiotis ,  Pragyana K. Mishra ,  Roopesh Ranjan

IPC: G06F9/451 ,  G06N5/02 ,  G06F40/197 ,  G06Q30/06 ,  G06N3/04

Abstract: A prediction model may be created to predict future actions likely to be performed by users while interacting with electronic content via user devices. The predictions may be used to streamline access to interface controls or other information to enable the users to facilitate or expedite performance of the predicted actions, while reducing computational demands on computing devices that provide the electronic content by, for example, reducing unnecessary intervening computing actions.