User click modelling in search queries

Patent metadata

Field	Value
Publication number	US-11853309-B2
Application number	US-202117514522-A
Country	US
Kind code	B2
Filing date	Oct 29, 2021
Priority date	Oct 30, 2020
Publication date	Dec 26, 2023
Grant date	Dec 26, 2023

Abstract

A method for ranking documents in search results includes defining a first training data set, the first training data set including, for each of a plurality of user queries, information respective of a document selected by a user from results responsive to the query and information respective of one or more documents within an observation window after the selected document in the results, and defining a second training data set, the second training data set including, for each of the plurality of user queries, information respective of the selected document. The method further includes training a first machine learning model with the first training data set, training a second machine learning model with the second training data set, and ranking documents of a further search result set according to the output of the first machine learning model and the output of the second machine learning model.

First claim

What is claimed is:

1. A method for ranking documents in search results, the method comprising:
defining a first training data set, the first training data set comprising, for each of a plurality of user queries, first information respective of a document selected by a user from results responsive to the query and second information respective of one or more documents within an observation window after the selected document in the results, the one or more documents within the observation window being different than the selected document;
defining a second training data set, the second training data set comprising, for each of the plurality of user queries, the first information respective of the selected document;
training a first machine learning model with the first training data set, the first machine learning model configured to output a first predicted user document selection;
training a second machine learning model with the second training data set, the second machine learning model configured to output a second predicted user document selection; and
ranking documents of a further search result set according to the output of the first machine learning model and the output of the second machine learning model.

2. The method of claim 1 , wherein training the first machine learning model and training the second machine learning model comprises:
conducting reinforcement learning on the first machine learning model and the second machine learning model to maximize a reward, the reward comprising a combination of a prediction accuracy of the first machine learning model and a prediction accuracy of the second machine learning model.

3. The method of claim 1 , wherein the observation window includes between one and three documents after the selected document.

4. The method of claim 1 , wherein:
the plurality of user queries comprises:
a plurality of previous user queries; and
one or more current user queries; and

training the first machine learning model with the first training data set and training the second machine learning model with the second training data set comprises:
batch training the first machine learning model and the second machine learning model according to the previous user queries; and
conducting reinforcement learning on the first machine learning model and the second machine learning model according to the one or more current user queries.

5. The method of claim 4 , wherein conducting reinforcement learning comprises maximizing a reward, the reward comprising a combination of a prediction accuracy of the first machine learning model and a prediction accuracy of the second machine learning model.

6. The method of claim 1 , further comprising:
receiving the documents of the further search result set, the further search result set responsive to a further search query; and
inputting information respective of the documents to the first machine learning model and the second machine learning model;
wherein ranking documents of a further search result set according to the output of the first machine learning model and the output of the second machine learning model comprises ranking the documents of the further search result set according to a mathematical combination of the output of the first machine learning model and the output of the second machine learning model.

7. The method of claim 1 , further comprising:
displaying the ranked further search result set to a user.

8. A method for ranking documents in search results, the method comprising:
defining a first training data set, the first training data set comprising, for each of a plurality of user queries, first information respective of a document selected by a user from results responsive to the query and second information respective of one or more documents within an observation window surrounding the selected document in the results, the one or more documents within the observation window being different than the selected document;
defining a second training data set, the second training data set comprising, for each of the plurality of user queries, the first information respective of the selected document;
training at least one machine learning model with the first training data set and the second training data set, the at least one machine learning model configured to output a predicted user document selection; and
ranking documents of a further search result set according to the output of the at least one machine learning model.

9. The method of claim 8 , wherein training the at least one machine learning model comprises:
conducting reinforcement learning on the at least one machine learning model to maximize a reward, the reward comprising a prediction accuracy of the at least one machine learning model.

10. The method of claim 8 , wherein the observation window includes between one and three documents after the selected document.

11. The method of claim 8 , wherein:
the plurality of user queries comprises:
a plurality of previous user queries; and
one or more current user queries; and

training the at least one machine learning model comprises:
batch training the at least one machine learning model according to the previous user queries; and
conducting reinforcement learning on the at least one machine learning model according to the one or more current user queries.

12. The method of claim 11 , wherein conducting reinforcement learning comprises maximizing a reward, the reward comprising a prediction accuracy of the at least one machine learning model.

13. The method of claim 8 , further comprising:
receiving the documents of the further search result set, the further search result set responsive to a further search query; and
inputting information respective of the documents to the at least one machine learning model.

14. The method of claim 8 , further comprising:
displaying the ranked further search result set to a user.

15. A system comprising:
a non-transitory, computer-readable medium storing instructions; and
a processor configured to execute the instructions to:
define a first training data set, the first training data set comprising, for each of a plurality of user queries, first information respective of a document selected by a user from results responsive to the query and second information respective of one or more documents within an observation window surrounding the selected document in the results, the one or more documents within the observation window being different than the selected document;
define a second training data set, the second training data set comprising, for each of the plurality of user queries, the first information respective of the selected document;
train at least one machine learning model with the first training data set and the second training data set, the at least one machine learning model configured to output a predicted user document selection; and
rank documents of a further search result set according to the output of the at least one machine learning model.

16. The system of claim 15 , wherein training the at least one machine learning model comprises:
conducting reinforcement learning on the at least one machine learning model to maximize a reward, the reward comprising a prediction accuracy of the at least one machine learning model.

17. The system of claim 15 , wherein the observation window includes between one and three documents after the selected document.

18. The system of claim 15 , wherein:
the plurality of user queries comprises:
a plurality of previous us

Assignees

• Home Depot Product Authority Llc

Inventors

• Zhou Jianghong
• Zahiri Sayyed
• Hughes Simon
• Kallumadi Surya
• Al Jadda Khalifeh
• Agichtein Eugene

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06F16/24578 Primary

  using ranking · CPC title

• G06F16/248

  Presentation of query results · CPC title