公开(公告)号：US12141214B2

公开(公告)日：2024-11-12

申请号：US17995248

申请日：2020-03-30

Applicant: GOOGLE LLC

Inventor： Michael Bendersky ,  Przemysław Gajda ,  Sergey Novikov ,  Marc Alexander Najork ,  Shuguang Han

IPC: G06F16/951 ,  G06F18/214 ,  G06Q30/0207 ,  G06Q30/0601

Abstract: Techniques of generating recrawl policies for commercial offer pages include generating a multiple strategy approach using a number of different strategies. In some implementations, each strategy is an arm of a K-armed adversarial bandits algorithm with reinforcement learning. Moreover, in some implementations, the multiple strategy approach also uses a machine learning algorithm to estimate parameters such as a click rate, impression rate, and likelihood of price change, i.e., change rate, which was assumed known in the conventional approaches.

公开(公告)号：US20250068679A1

公开(公告)日：2025-02-27

申请号：US18936579

申请日：2024-11-04

Applicant: GOOGLE LLC

Inventor： Michael Bendersky ,  Przemyslaw Gajda ,  Sergey Novikov ,  Marc Alexander Najork ,  Shuguang Han

IPC: G06F16/951 ,  G06F18/214 ,  G06Q30/0207 ,  G06Q30/0601

Abstract: Techniques of generating recrawl policies for commercial offer pages include generating a multiple strategy approach using a number of different strategies. In some implementations, each strategy is an arm of a K-armed adversarial bandits algorithm with reinforcement learning. Moreover, in some implementations, the multiple strategy approach also uses a machine learning algorithm to estimate parameters such as a click rate, impression rate, and likelihood of price change, i.e., change rate, which was assumed known in the conventional approaches.

公开(公告)号：US20230169128A1

公开(公告)日：2023-06-01

申请号：US17995248

申请日：2020-03-30

Applicant: GOOGLE LLC

Inventor： Michael Bendersky ,  Przemyslaw Gajda ,  Sergey Novikov ,  Marc Alexander Najork ,  Shuguang Han

IPC: G06F16/951 ,  G06Q30/0207

CPC classification number: G06F16/951 ,  G06Q30/0239

Abstract: Techniques of generating recrawl policies for commercial offer pages include generating a multiple strategy approach using a number of different strategies. In some implementations, each strategy is an arm of a K-armed adversarial bandits algorithm with reinforcement learning. Moreover, in some implementations, the multiple strategy approach also uses a machine learning algorithm to estimate parameters such as a click rate, impression rate, and likelihood of price change, i.e., change rate, which was assumed known in the conventional approaches.