Machine-learned trust scoring for player matchmaking

1. A method, comprising:

training a machine learning model using training data to obtain a trained machine learning model;

accessing, by a computing system, data associated with a plurality of user accounts registered with a video game service;

determining trust scores associated with the plurality of user accounts by:

identifying a first user account of the plurality of user accounts that is associated with at least one of a professional player or an employee of a service provider of the video game service;

proactively assigning a predetermined trust score to the first user account without using the trained machine learning model;

providing the data associated with the plurality of user accounts other than the first user account as input to the trained machine learning model; and

generating, as output from the trained machine learning model, additional trust scores associated with the plurality of user accounts other than the first user account, the additional trust scores relating to probabilities of players associated with the plurality of user accounts other than the first user account cheating, or not cheating, while playing one or more video games in a multiplayer mode;

receiving, from a plurality of client machines, information indicating logged-in user accounts, of the plurality of user accounts, that are currently logged into a client application that is executing a video game on each client machine of the plurality of client machines;

defining, by the computing system, multiple matches into which players associated with the logged-in user accounts are to be grouped for playing the video game in the multiplayer mode, the multiple matches comprising at least a first match and a second match;

assigning, by the computing system, a first subset of the logged-in user accounts to the first match based at least in part on the trust scores associated with the first subset;

assigning, by the computing system, a second subset of the logged-in user accounts to the second match based at least in part on the trust scores associated with the second subset; and

causing, by the computing system, the client application executing the video game on each client machine to initiate one of the first match or the second match based at least in part on a user account, of the logged-in user accounts, that is associated with the client machine.

2. The method of claim 1 , wherein the training data includes labels for each user account of a sampled set of user accounts indicating whether the user account is associated with a player who has cheated while playing at least one video game.

3. The method of claim 1 , further comprising retraining the machine learning model using updated training data to obtain a newly trained machine learning model that is adapted to recent player behaviors.

4. A method, comprising:

training a machine learning model using training data to obtain a trained machine learning model;

determining, by a computing system, scores for a plurality of user accounts registered with a video game service, wherein the scores are determined by:

identifying a first user account of the plurality of user accounts that is associated with at least one of a professional player or an employee of a service provider of the video game service;

proactively assigning a predetermined score to the first user account without using the trained machine learning model;

accessing data associated with an individual user account of the plurality of user accounts other than the first user account;

providing the data as input to the trained machine learning model; and

generating, as output from the trained machine learning model, a score associated with the individual user account, the score indicative of a probability of a player associated with the individual user account cheating, or not cheating, while playing one or more video games in a multiplayer mode;

receiving, by the computing system, information from a plurality of client machines, the information indicating logged-in user accounts, of the plurality of user accounts, that are logged into a client application executing a video game;

defining, by the computing system, multiple matches into which players associated with the logged-in user accounts are to be grouped for playing the video game in the multiplayer mode, the multiple matches comprising at least a first match and a second match;

assigning, by the computing system, a first subset of the logged-in user accounts to the first match and a second subset of the logged-in user accounts to the second match based at least in part on the scores determined for the logged-in user accounts;

causing, by the computing system, a first subset of the plurality of client machines associated with the first subset of the logged-in user accounts to execute the video game in the first match; and

causing, by the computing system, a second subset of the plurality of client machines associated with the second subset of the logged-in user accounts to execute the video game in the second match.

5. The method of claim 4 , wherein the training of the machine learning model comprises:

accessing, by the computing system, the training data, the training data associated with a sampled set of user accounts registered with the video game service; and

labeling each user account of the sampled set of user accounts with a label that indicates whether the user account is associated with a player who has cheated while playing at least one video game.

6. The method of claim 4 , wherein the training of the machine learning model comprises setting weights for at least one of a set of features derived from the training data or parameters internal to the machine learning model.

7. The method of claim 4 , wherein at least some of the training data was generated as a result of players playing one or multiple video games on a platform provided by the video game service.

8. The method of claim 7 , wherein the at least some of the training data represents match history data of the players who have played the one or multiple video games in the multiplayer mode by participating in matches.

9. The method of claim 4 , further comprising retraining the machine learning model using updated training data to obtain a newly trained machine learning model that is adapted to recent player behaviors.

10. The method of claim 5 , wherein:

the label indicates, for each user account, whether the user account is associated with a player who has been banned from playing, via the video game service, the at least one video game as a consequence of having been determined to have cheated while playing the at least one video game.

11. The method of claim 4 , wherein the assigning of the first subset of the logged-in user accounts to the first match and the second subset of the logged-in user accounts to the second match comprises:

determining that the scores associated with the first subset of the logged-in user accounts are less than a threshold score; and

determining that the scores associated with the second subset of the logged-in user accounts are equal to or greater than the threshold score.

12. The method of claim 4 , wherein the assigning of the first subset of the logged-in user accounts to the first match and the second subset of the logged-in user accounts to the second match is further based on at least one of:

skill levels of players associated with the logged-in user accounts;

amounts of time the logged-in user accounts have been waiting to be placed into one of the multiple matches; or

geographic regions associated with the logged-in user accounts.

13. A system, comprising:

one or more processors; and

memory storing computer-executable instructions that, when executed by the one or more processors, cause the system to:

train a machine learning model using training data to obtain a trained machine learning model;

determine trust scores for a plurality of user accounts registered with a video game service, wherein the trust scores are determined by:

identifying a first user account of the plurality of user accounts that is associated with at least one of a professional player or an employee of a service provider of the video game service;

proactively assigning a predetermined score to the first user account without using the trained machine learning model;

accessing data associated with an individual user account of the plurality of user accounts other than the first user account;

providing the data as input to the trained machine learning model; and

generating, as output from the trained machine learning model, a trust score associated with the individual user account, the trust score relating to a probability of a player associated with the individual user account cheating, or not cheating, while playing one or more video games in a multiplayer mode;

receive information from a plurality of client machines, the information indicating logged-in user accounts, of the plurality of user accounts, that are logged into a client application executing a video game;

define multiple matches into which players associated with the logged-in user accounts are to be grouped for playing the video game in the multiplayer mode, the multiple matches comprising at least a first match and a second match;

assign a first subset of the logged-in user accounts to the first match and a second subset of the logged-in user accounts to the second match based at least in part on the trust scores determined for the logged-in user accounts;

cause a first subset of the plurality of client machines associated with the first subset of the logged-in user accounts to execute the video game in the first match; and

cause a second subset of the plurality of client machines associated with the second subset of the logged-in user accounts to execute the video game in the second match.

14. The system of claim 13 , wherein the computer-executable instructions, when executed by the one or more processors, further cause the system to, prior to determining the trust scores:

access the training data, the training data associated with a sampled set of user accounts; and

label each user account of the sampled set of user accounts with a label that indicates whether the user account is associated with a player who has cheated while playing at least one video game.

15. The system of claim 14 , wherein:

the label indicates, for each user account, whether the user account is associated with a player who has been banned from playing, via the video game service, the at least one video game as a consequence of having been determined to have cheated while playing the at least one video game.

16. The system of claim 13 , wherein at least some of the training data was generated as a result of players playing one or multiple video games on a platform provided by the video game service.

17. The system of claim 13 , wherein the computer-executable instructions, when executed by the one or more processors, further cause the system to, prior to determining the trust scores, determine that players associated with the plurality of user accounts have accrued a level of experience in an individual play mode.

18. The method of claim 2 , wherein each of the labels indicates whether the user account is associated with a player who has been banned from playing, via the video game service, the at least one video game as a consequence of having been determined to have cheated while playing the at least one video game.

19. The method of claim 4 , further comprising, prior to the determining of the scores, determining that players associated with the plurality of user accounts have accrued a level of experience in an individual play mode.

20. The method of claim 1 , further comprising, prior to the determining of the trust scores, determining that players associated with the plurality of user accounts have accrued a level of experience in an individual play mode.