Granted Patent B2

US 12,493,898 · App. 17/566,035 · Granted Dec 9, 2025

Inventors: Oren Shlomo Somekh (Cfar-Neter, IL); Alex Shtoff (Haifa, IL); Avi Shahar (Ramat-Gan, IL); Tomer Shadi (New York, NY); Yair Koren (Haifa, IL); Anna Itzhaki (New York, NY); Yohay Kaplan (Haifa, IL); Tal Cohen (Ramat-Gan, IL); Boris Trayvas (New York, NY)

Assignee: YAHOO AD TECH LLC

G06Q30/0255G06Q30/0246G06Q30/0275G06Q30/0276G06N20/00

Abstract

Claims (108)

1 . A method implemented on at least one processor, a memory, and a communication platform for displaying ads, comprising:

receiving, at an explore/exploit layer (EEL) associated with one of a plurality of frontend ad serving engines, from an ad recommendation backend server, and based on a selection criteria configuration that maps each of subsets of combination distributions with respect to a plurality of ads to a corresponding one of the plurality of frontend ad serving engines, one of the subsets of combination distributions,

wherein each of the plurality of ads has a plurality of attributes, each of the plurality of attributes is associated with multiple assets, and each of the multiple assets can be used for rendering the attribute,

wherein different subsets of the combination distributions are received by different ones of the plurality of frontend ad serving engines, and

wherein each of the combination distributions represents a probability for each of a plurality of different combinations of assets being selected by users sharing one or more common characteristics;

sending, by the frontend ad serving engine to the ad recommendation backend server, a request for a recommendation of an ad to be used for bidding an ad display opportunity in a slot of a webpage viewed by a user on a user device;

receiving, at the frontend ad serving engine from the ad recommendation backend server, an ad recommended based on a display environment related to the user and the webpage, wherein the ad recommended is one of the plurality of ads;

upon a successful auction of the ad,

obtaining, by the frontend ad serving engine, a segment key representing a combination of a feature of the user and a feature of the user device, and

determining, by the frontend ad serving engine, whether combination distributions for the ad are available in the EEL;

in response to a determination that the combination distributions for the ad are available in the EEL, drawing, by the frontend ad serving engine, directly from the EEL without communicating with the ad recommendation backend server, a combination from the plurality of combinations based on the segment key and the combination distribution associated with the ad in the EEL; and

transmitting, by the frontend ad serving engine to the user device, the combination to render the ad to the user in the slot of the webpage on the user device.

2 . The method of claim 1 , wherein the plurality of attributes of each of the plurality of ads include at least some of:

a title of the ad;

an image that visually conveys information about the ad; and

a description that textually summarizes content of the ad, wherein

any of the multiple assets associated with each of the plurality of attributes of the ad can be used to render the attribute, and

the combination includes a plurality of assets, each for one of the plurality of attributes of the ad and is drawn to maximize a return of displaying the ad according to a predetermined criterion.

3 . The method of claim 2 , wherein the predetermined criterion includes:

a click through rate (CTR); or

a conversion rate (CVR).

4 . The method of claim 1 , wherein combination distributions associated with each of the plurality ads provide different ways to render the ad, wherein each combination in the combination distributions

represents one way to render the ad with respect to the display environment; and

is provided with an indication of a likelihood to be drawn to render the ad;

wherein the indication of the likelihood is determined based on a predicted performance of the combination in the display environment estimated by a prediction model; and

wherein the prediction model is obtained via machine learning based on training data to maximize a return of the ad according to a predetermined criterion.

5 . The method of claim 4 , wherein

the indication of the likelihood corresponds to a probability;

combinations related to an ad in the plurality of ads under each display environment form a distribution with the probabilities associated therewith adding up to one; and

the step of drawing is performed in accordance with the distributed probabilities for the combinations.

6 . The method of claim 1 , wherein the display environment is characterized based on information related to at least one of:

the user;

the user device;

the slot for displaying an ad;

the webpage where the slot resides; and

a traffic segment associated with the display environment.

7 . The method of claim 1 , wherein the combination distributions stored in the EEL are regularly updated so that the combination is drawn from the updated combination distributions in the EEL to maximize a return of displaying the ad in the display environment in a manner adapted to a change in the marketplace.

8 . Machine readable and non-transitory medium having information recorded thereon for displaying ads, wherein the information, once read by the machine, causes the machine to perform the following steps:

receiving, at an explore/exploit layer (EEL) associated with one of a plurality of frontend ad serving engines, an ad recommendation backend server, and based on a selection criteria configuration that maps each of subsets of combination distributions with respect to a plurality of ads to a corresponding one of the plurality of frontend ad serving engines, one of the subsets of combination distributions,

wherein each of the plurality of ads has a plurality of attributes, each of the plurality of attributes is associated with multiple assets, and each of the multiple assets can be used for rendering the attribute,

wherein different subsets of the combination distributions are received by different ones of the plurality of frontend ad serving engines, and

wherein each of the combination distributions represents a probability for each of a plurality of different combinations of assets being selected by users sharing one or more common characteristics;

sending, by the frontend ad serving engine to the ad recommendation backend server, a request for a recommendation of an ad to be used for bidding an ad display opportunity in a slot of a webpage viewed by a user on a user device;

receiving, at the frontend ad serving engine from the ad recommendation backend server, an ad recommended based on a display environment related to the user and the webpage, wherein the ad recommended is one of the plurality of ads;

upon a successful auction of the ad,

obtaining, by the frontend ad serving engine, a segment key representing a combination of a feature of the user and a feature of the user device, and

determining, by the frontend ad serving engine, whether combination distributions for the ad are available in the EEL;

in response to a determination that the combination distributions for the ad are available in the EEL, drawing, by the frontend ad serving engine, directly from the EEL without communicating with the ad recommendation backend server, a combination from the plurality of combinations based on the segment key and the combination distribution associated with the ad in the EEL; and

transmitting, by the frontend ad serving engine to the user device, the combination to render the ad to the user in the slot of the webpage on the user device.

9 . The medium of claim 8 , wherein the plurality of attributes of each of the plurality of ads include at least some of:

a title of the ad;

an image that visually conveys information about the ad; and

a description that textually summarizes content of the ad, wherein

any of the multiple assets associated with each of the plurality of attributes of the ad can be used to render the attribute, and

the combination includes a plurality of assets, each for one of the plurality of attributes of the ad and is drawn to maximize a return of displaying the ad according to a predetermined criterion.

10 . The medium of claim 9 , wherein the predetermined criterion includes:

a click through rate (CTR); or

a conversion rate (CVR).

11 . The medium of claim 8 , wherein combination distributions associated with each of the plurality ads provide different ways to render the ad, wherein each combination in the combination distributions

represents one way to render the ad with respect to the display environment; and

is provided with an indication of a likelihood to be drawn to render the ad;

wherein the indication of the likelihood is determined based on a predicted performance of the combination in the display environment estimated by a prediction model; and

wherein the prediction model is obtained via machine learning based on training data to maximize a return of the ad according to a predetermined criterion.

12 . The medium of claim 11 , wherein

the indication of the likelihood corresponds to a probability;

combinations related to an ad in the plurality of ads under each display environment form a distribution with the probabilities associated therewith adding up to one; and

the step of drawing is performed in accordance with the distributed probabilities for the combinations.

13 . The medium of claim 8 , wherein the display environment is characterized based on information related to at least one of:

the user;

the user device;

the slot for displaying an ad;

the webpage where the slot resides; and

a traffic segment associated with the display environment.

14 . The medium of claim 8 , wherein the combination distributions stored in the EEL are regularly updated so that the combination is drawn from the updated combination distributions in the EEL to maximize a return of displaying the ad in the display environment in a manner adapted to a change in the marketplace.

15 . A system for displaying ads comprises:

explore/exploit layers (EELs) each associated with a corresponding one of a plurality of frontend ad serving engines, each configured for receiving, from an ad recommendation backend server, and based on a selection criteria configuration that maps each of subsets of combination distributions with respect to a plurality of ads to a corresponding one of the plurality of frontend ad serving engines, one of the subsets of combination distributions,

wherein each of the plurality of ads has a plurality of attributes, each of the plurality of attributes is associated with multiple assets, and each of the multiple assets can be used for rendering the attribute,

wherein different subsets of the combination distributions are received by different ones of the plurality of frontend ad serving engines, and

wherein each of the combination distributions represents a probability for each of a plurality of different combinations of assets being selected by users sharing one or more common characteristics; and

the plurality of frontend ad serving engine each associated with a corresponding one of the EELs and configured for

sending, to the ad recommendation backend server, a request for a recommendation of an ad to be used for bidding an ad display opportunity in a slot of a webpage viewed by a user on a user device,

receiving, from the ad recommendation backend server, an ad recommended based on a display environment related to the user and the webpage, wherein the ad recommended is one of the plurality of ads,

upon a successful auction of the ad,

obtaining a segment key representing a combination of a feature of the user and a feature of the user device, and

determining whether combination distributions for the ad are available in the EEL,

in response to a determination that the combination distributions for the ad are available in the EEL, drawing, directly from the EEL without communicating with the ad recommendation backend server, a combination from the plurality of combinations based on the segment key and the combination distribution associated with the ad stored in the EEL, and

transmitting the combination to the user device to render the ad to the user in the slot of the webpage on the user device.

16 . The system of claim 15 , wherein the plurality of attributes of each of the plurality of ads include at least some of:

a title of the ad;

an image that visually conveys information about the ad; and

a description that textually summarizes content of the ad, wherein

any of the multiple assets associated with each of the plurality of attributes of the ad can be used to render the attribute, and

the combination includes a plurality of assets, each for one of the plurality of attributes of the ad and is drawn to maximize a return of displaying the ad according to a predetermined criterion.

17 . The system of claim 16 , wherein the predetermined criterion includes:

a click through rate (CTR); or

a conversion rate (CVR).

18 . The system of claim 15 , wherein combination distributions associated with each of the plurality ads provide different ways to render the ad, wherein each combination in the combination distributions

represents one way to render the ad with respect to the display environment; and

is provided with an indication of a likelihood to be drawn to render the ad;

wherein the indication of the likelihood is determined based on a predicted performance of the combination in the display environment estimated by a prediction model; and

wherein the prediction model is obtained via machine learning based on training data to maximize a return of the ad according to a predetermined criterion.

19 . The system of claim 15 , wherein the display environment is characterized based on information related to at least one of:

the user;

the user device;

the slot for displaying an ad;

the webpage where the slot resides; and

a traffic segment associated with the display environment.

20 . The system of claim 15 , wherein the combination distributions stored in the EEL are regularly updated so that the combination is drawn from the updated combination distributions in the EEL to maximize a return of displaying the ad in the display environment in a manner adapted to a change in the marketplace.

Assignments (2)

Continuity (1)

References Cited (22)