Bid value determination for a first-price auction

Patent metadata

Field	Value
Publication number	US-12223405-B2
Application number	US-202318199953-A
Country	US
Kind code	B2
Filing date	May 21, 2023
Priority date	Aug 17, 2020
Publication date	Feb 11, 2025
Grant date	Feb 11, 2025

Abstract

Shaded bid values may be determined and/or submitted to one or more auction modules for participation in auctions. Auction information including at least one of impression indications associated with the auctions, sets of features associated with the auctions, the shaded bid values associated with the auctions, etc. may be stored in a database. A machine learning model may be trained using the auction information to generate a first machine learning model with feature parameters associated with features. A bid request, indicative of a second set of features, may be received. The first machine learning model may be used to determine win probabilities and/or expected bid surpluses associated with multiple shaded bid values based upon one or more feature parameters, of the feature parameters, associated with the second set of features. A shaded bid value for submission may be determined based upon the win probabilities and/or the expected bid surpluses.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
training a machine learning model using a plurality of sets of auction information to generate a first machine learning model comprising a plurality of feature parameters associated with a plurality of features of the plurality of sets of auction information, wherein the first machine learning model comprises (i) a first bid parameter comprising a first bid weight and (ii) a first bias parameter comprising a first bias weight;
loading the first machine learning model onto a bid shading module of a demand-side platform (DSP);
receiving, by the DSP, a bid request, wherein:
the bid request is associated with a request for content associated with a client device;
the bid request is indicative of a set of features comprising one or more features associated with the request for content;
a set of feature parameters, of the plurality of feature parameters, are associated with the set of features;
a first feature parameter, of the set of feature parameters, is associated with a first feature of the set of features;
a set of weights, of the set of feature parameters, is associated with the set of features; and
the first feature parameter comprises a first weight, of the set of weights, associated with the first feature;

determining a bid value associated with a content item;
inputting, into the bid shading module of the DSP, the bid value;
determining, based upon the set of feature parameters and using the first bid parameter and the first bias parameter of the first machine learning model loaded onto the bid shading module of the DSP, a plurality of win probabilities associated with a plurality of shaded bid values, wherein:
each shaded bid value of the plurality of shaded bid values does not exceed the bid value; and
a first win probability of the plurality of win probabilities is associated with a first shaded bid value of the plurality of shaded bid values and corresponds to a probability that the content item wins an auction associated with the request for content responsive to submitting the first shaded bid value to an auction module associated with the request for content;

determining, based upon the plurality of win probabilities associated with the plurality of shaded bid values, a second shaded bid value, wherein the determining the second shaded bid value is performed based upon (i) the set of weights, (ii) the first bid weight and (iii) the first bias weight; and
submitting the second shaded bid value to a first auction module for participation in a first auction associated with the request for content,
wherein one or more content items are provided for presentation on the client device associated with the request for content based upon a determination that the one or more content items are a winner of the auction.

2. The method of claim 1 , wherein:
the determining the second shaded bid value comprises selecting, based upon the plurality of win probabilities associated with the plurality of shaded bid values, the second shaded bid value from the plurality of shaded bid values.

3. The method of claim 1 , wherein:
the auction is a first-price auction.

4. The method of claim 1 , wherein:
the auction is an open first-price auction.

5. The method of claim 1 , wherein:
the set of features comprises at least one of:
a internet resource associated with the request for content;
a time of day associated with the request for content;
a day of week associated with the request for content; or
a location associated with the client device.

6. The method of claim 1 , wherein:
the auction is a closed first-price auction.

7. A computing device comprising:
a processor; and
memory comprising processor-executable instructions that when executed by the processor cause performance of operations, the operations comprising:
training a machine learning model using a plurality of sets of auction information to generate a first machine learning model comprising a plurality of feature parameters associated with a plurality of features of the plurality of sets of auction information, wherein the first machine learning model comprises (i) a first bid parameter comprising a first bid weight and (ii) a first bias parameter comprising a first bias weight;
loading the first machine learning model onto a bid shading module of a demand-side platform (DSP);
receiving, by the DSP, a bid request, wherein:
the bid request is associated with a request for content associated with a client device;
the bid request is indicative of a set of features comprising one or more features associated with the request for content;
a set of feature parameters, of the plurality of feature parameters, are associated with the set of features;
a first feature parameter, of the set of feature parameters, is associated with a first feature of the set of features;
a set of weights, of the set of feature parameters, is associated with the set of features; and
the first feature parameter comprises a first weight, of the set of weights, associated with the first feature;

determining a bid value associated with a content item;
inputting, into the bid shading module of the DSP, the bid value;
determining, based upon the set of feature parameters and using the first bid parameter and the first bias parameter of the first machine learning model loaded onto the bid shading module of the DSP, a plurality of expected bid surpluses associated with a plurality of shaded bid values, wherein:
each shaded bid value of the plurality of shaded bid values does not exceed the bid value; and
a first expected bid surplus of the plurality of expected bid surpluses is associated with a first shaded bid value of the plurality of shaded bid values;

determining, based upon the plurality of expected bid surpluses and the plurality of shaded bid values, a second shaded bid value, wherein the determining the second shaded bid value is performed based upon (i) the set of weights, (ii) the first bid weight and (iii) the first bias weight; and
submitting the second shaded bid value to an auction module for participation in an auction associated with the request for content,
wherein one or more content items are provided for presentation on the client device associated with the request for content based upon a determination that the one or more content items are a winner of the auction.

8. The computing device of claim 7 , wherein:
the determining the second shaded bid value comprises selecting, based upon the plurality of expected bid surpluses associated with the plurality of shaded bid values, the second shaded bid value from the plurality of shaded bid values.

9. The computing device of claim 8 , wherein:
the selecting the second shaded bid value from the plurality of shaded bid values is performed based upon a determination that the second shaded bid value is associated with a maximum expected bid surplus of the plurality of expected bid surpluses.

10. The computing device of claim 7 , wherein:
the auction is a first-price auction.

11. The computing device of claim 7 , wherein:
the set of features comprises at least one of:
a internet resource associated with the request for content;
a time of day associated with the request for content;
a day of week associated with the request for content; or
a location associated with the client device.

12. The computing device of claim 7 , wherein:
the auction is an open first-price auction.

13. The computing device of claim 7 , wherein:
the auction is a closed first-price auction.

14. A non-transitory machine readable medium

Assignees

• Yahoo Ad Tech Llc

Inventors

• Pan Shengjun
• Zhou Tian
• Kitts Brendan
• He Hao
• Shetty Bharatbhushan
• Gligorijevic Djordje
• Pan Junwei
• Mao Tingyu
• Gultekin San
• Paladugu Balaji Srinivasa Rao
• Zhang Jianlong
• Thomas Sneha
• Flores Aaron

Classifications

• G06Q30/0275 Primary

  Auctions · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N20/00 Primary

  Machine learning · CPC title