公开(公告)号：US20180137121A1

公开(公告)日：2018-05-17

申请号：US14147002

申请日：2014-01-03

Applicant: Amazon Technologies, Inc.

Inventor： Tarun Agarwal

IPC: G06F17/30

CPC classification number: G06F16/437

Abstract: The present disclosure relates to content identification based on dynamic formation of group profiles. The group profile can be associated with some organization criteria, such as physical location, temporal attributes, etc., and can be configured with various preferences and restrictions. Subsequent to the initialization of the group profile(s), individual users can be associated with the group. Based on the addition (or subtraction) of users, the established group profiles are updated or modified according to one or more attributes of the individual profiles of the users associated with the group. Content of interest to the group can be identified using the group profile information. The identified content can also be provided to the individual users associated with the group. The group profile can be updated based on group membership change or feedback to identified content, and additional or alternative content recommendations can be provided.

公开(公告)号：US10013416B1

公开(公告)日：2018-07-03

申请号：US14974501

申请日：2015-12-18

Applicant: Amazon Technologies, Inc.

Inventor： Vikas Seshagiri Rao Bhardwaj ,  Jared Kane Kramer ,  Hung Chiu Phillip Keung ,  Kapil Chahal ,  Tarun Agarwal ,  Jason Sun ,  Gunashekar Jayaraman ,  Dhvani Yogesh Shah ,  Ariel Grob Dos Santos ,  Sheng Bi ,  Polly Po Yee Lee ,  Anirban Chakraborty

IPC: G06F17/27 ,  G06F17/28 ,  H04L29/06 ,  H04L29/08 ,  G06N99/00 ,  G10L15/18

CPC classification number: G10L15/18 ,  G06F17/2785 ,  G06N5/022 ,  G06N20/00 ,  G10L15/26

Abstract: Disclosed are various embodiments for an automated service agent. Natural language inputs are obtained from a client. The natural language inputs and other data are input to machine learning models to identify intents reflected in the natural language inputs. A state in a workflow is updated to reflect the identified intents. Responses may be communicated to the client to further progress in the workflow.

公开(公告)号：US09886670B2

公开(公告)日：2018-02-06

申请号：US14319880

申请日：2014-06-30

Applicant: Amazon Technologies, Inc.

Inventor： Leo Parker Dirac ,  Tarun Agarwal

IPC: G06N99/00

CPC classification number: G06N99/005

Abstract: A first representation of a feature processing recipe is received at a machine learning service. The recipe includes a section in which groups of variables on which common transformations are to be applied are defined, and a section in which a set of transformation operations are specified. The first representation of the recipe is validated based at least in part on a library of function definitions supported by the service, and an executable version of the recipe is generated. In response to a determination that the recipe is to be executed on a particular data set, a set of provider network resources is used to implement a transformation operation indicated in the recipe.

公开(公告)号：US10339465B2

公开(公告)日：2019-07-02

申请号：US14463434

申请日：2014-08-19

Applicant: Amazon Technologies, Inc.

Inventor： Robert Matthias Steele ,  Tarun Agarwal ,  Leo Parker Dirac ,  Jun Qian

IPC: G06N20/00 ,  G06N5/02

Abstract: During a training phase of a machine learning model, representations of at least some nodes of a decision tree are generated and stored on persistent storage in depth-first order. A respective predictive utility metric (PUM) value is determined for one or more nodes, indicating expected contributions of the nodes to a prediction of the model. A particular node is selected for removal from the tree based at least partly on its PUM value. A modified version of the tree, with the particular node removed, is stored for obtaining a prediction.