# US 2019/0244137 A1 · Using machine learning to estimate or forecast resource use with time-varying demand in gaming platforms

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0244137 A1
- Document ID: US20190244137A1
- Status: Published application
- Kind: A1
- Application number: 15/892,148
- Filed: 2018-02-08
- Published: 2019-08-08
- Earliest claimed: 2018-02-08
- Assignees: Roblox Corporation
- Applicants: Roblox Corporation
- Inventors: Arthur Remy Malan
- Web page: https://patentgrep.com/patent/US20190244137A1

## CPC

- G06N 3/084: Backpropagation, e.g. using gradient descent
- A63F 13/352: involving special game server arrangements, e.g. regional servers connected to a national server or a plurality of servers managing partitions of the game world
- A63F 13/67: adaptively or by learning from player actions, e.g. skill level adjustment or by storing successful combat sequences for re-use
- A63F 13/75: Enforcing rules, e.g. detecting foul play or generating lists of cheating players
- G06N 3/0499: Feedforward networks
- G06N 3/09: Supervised learning
- G06N 20/00: Machine learning
- G06Q 10/40: Business processes related to social networking or social networking services
- H04L 67/535: {Tracking the activity of the user (network monitoring arrangements H04L43/00; recording of computer activity G06F11/34)}
- A63F 13/79: involving player-related data, e.g. identities, accounts, preferences or play histories
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- A63F: CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A method and system is disclosed for training a machine learning model by generating first training input that includes a first number of reports at a first point in time. The reports are submitted by the users of the gaming platform and identify incidents where content of the gaming platform violates a policy of use associated with the gaming platform. The method and system generates second training input including a number of resources active at the first point in time. The method and system generates first target output identifies a number of resources sufficient to evaluate the target percentage of the first number of reports. The method and system provide the training data to train the machine learning model on (i) a set of training inputs comprising the first training input and the second training input, and (ii) a set of target outputs comprising the first target output.

## Claims

### Claim 1

A method for training a machine learning model to estimate resources sufficient to evaluate a target percentage of reports submitted by users of a gaming platform where a number of the reports vary over time, the method comprising:
generating training data for the machine learning model, wherein generating the training data comprises:
generating first training input, the first training input comprising a first number of reports at a first point in time, wherein the reports are submitted by the users of the gaming platform and identify incidents where content of the gaming platform violates a policy of use associated with the gaming platform;
generating second training input, the second training input comprising a number of resources active at the first point in time;
generating first target output for the first training input and the second training input, wherein the first target output identifies a number of resources sufficient to evaluate the target percentage of the first number of reports; and
providing the training data to train the machine learning model on (i) a set of training inputs comprising the first training input and the second training input, and (ii) a set of target outputs comprising the first target output.

### Claim 2 (depends on 1)

The method of claim 1, wherein generating the training data further comprises:
generating third training input, the third training input comprising a percentage of the first number of reports that are being evaluated at the first point in time by the number of resources active at the first point in time; and
wherein the set of training inputs comprising the first, the second, and the third training input.

### Claim 3 (depends on 2)

The method of claim 2, wherein generating the training data further comprises:
generating fourth training input, the fourth training input comprising an indication of a rate of change between the first number of reports at the first point in time and another number of reports at a previous point in time; and
wherein the set of training inputs comprising the first, the second, the third, and the fourth training input.

### Claim 4 (depends on 3)

The method of claim 3, wherein generating the training data further comprises:
generating fifth training input, the fifth training input comprising an identifier of a category for each of the first number of reports; and
wherein the set of training inputs comprising the first, the second, the third, the fourth, and the fifth training input.

### Claim 5 (depends on 1)

The method of claim 1, wherein generating the training data further comprises:
generating sixth training input, the sixth training input comprising a second number of reports at a second point in time;
generating seventh training input, the seventh training input comprising a number of resources active at the second point in time;
generating second target output for the sixth training input and the seventh training input, wherein the second target output identifies a number of resources sufficient to evaluate the target percentage of the second number reports; and
wherein the set of training inputs comprising the first, the second, the sixth and the seventh training input, and wherein set of target outputs comprising the first target output and the second target output.

### Claim 6 (depends on 1)

The method of claim 1, wherein the content of the gaming platform identified in the reports as violating the policy of use associated with the gaming platform comprises user-generated content.

### Claim 7 (depends on 1)

The method of claim 1, wherein the resources comprise at least one of server resources or data store resources.

### Claim 8 (depends on 1)

The method of claim 1, wherein a number of the reports varies over time and the target percentage of reports to evaluate at any point in time is constant percentage, and wherein a number of resources active varies over the time.

### Claim 9 (depends on 1)

The method of claim 1, wherein each training input of the set of training inputs is mapped to the target output of the set of target outputs.

### Claim 10

A method for using a trained machine learning model to estimate resources to use in order to evaluate a target percentage of reports submitted by users of a gaming platform where a number of the reports submitted vary over time, comprising:
providing to the trained machine learning model a first input comprising an indication of a time period during which the target percentage of a first number of reports is to be evaluated, wherein the first number reports are submitted by the users of the gaming platform and identify incidents where content of the gaming platform violates a policy of use associated with the gaming platform; and
obtaining, from the trained machine learning model, one or more outputs identifying (i) a number of resources to evaluate the target percentage of the number of reports during the time period, and (ii) a level of confidence that the number of resources is sufficient to evaluate the target percentage of the number reports during the time period.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
generating a forecast that estimates the number of resources to use in order to evaluate the target percentage of the number of reports during the time period in view of the level of confidence.

### Claim 12 (depends on 11)

The method of claim 11, where generating the forecast that estimates the number of resources to use in order to evaluate the target percentage of the number of reports during the time period in view of the level of confidence comprises:
determining whether the level of confidence that the number of resources is sufficient to evaluate the target percentage of the number of reports during the time period exceeds a threshold level of confidence; and
responsive to determining that the level of confidence that the number of resources is sufficient to evaluate the target percentage of the number of reports during the time period exceeds the threshold level, generating the forecast.

### Claim 13 (depends on 10)

The method of claim 10, further comprising:
providing to the trained machine learning model second input comprising the number of reports at a first point in time of the time period.

### Claim 14 (depends on 13)

The method of claim 13, further comprising:
providing to the trained machine learning model a third input comprising the number of resources active at the first point in time of the time period, a fourth input comprising a percentage of the number of reports at the first point in time that are being evaluated by a number resources active at the first point in time, and a fifth input comprising an indication of a rate of change between the first number of reports at the first point in time and a second number of reports at a previous point in time.

### Claim 15 (depends on 10)

The method of claim 10, wherein the first number of the reports to evaluate during the time period is an unknown number.

### Claim 16 (depends on 10)

The method of claim 10, wherein the trained machine learning model was trained using a first training input comprising a first known number of reports at a second point in time, a second training input comprising a number of resources active at the second point in time, and a first target output for the first training input and the second training input, the first target output identifying a number of resources sufficient to evaluate the target percentage of the first known number of reports.

### Claim 17

A system for using a trained machine learning model to estimate resources to use in order to evaluate a target percentage of reports submitted by users of a platform where a number of the reports submitted vary over time, the system comprising:
a memory; and
a processing device, coupled to the memory, the processing device to:
provide to the trained machine learning model a first input comprising an indication of a time period during which the target percentage of a first number of reports is to be evaluated, wherein the first number reports are submitted by the users of the platform and identify incidents where content of the platform violates a policy of use associated with the platform; and
obtain, from the trained machine learning model, one or more outputs identifying (i) a number of resources to evaluate the target percentage of the number of reports during the time period, and (ii) a level of confidence that the number of resources is sufficient to evaluate the target percentage of the number reports during the time period.

### Claim 18 (depends on 17)

The system of claim 17, the processing device further to:
generate an estimate of the number of resources to use in order to evaluate the target percentage of the number of reports during the time period in view of the level of confidence.

### Claim 19

A system for training a machine learning model to estimate resources sufficient to evaluate a target percentage of reports submitted by users of a platform where a number of the reports vary over time, the system comprising:
a memory; and
a processing device, coupled to the memory, the processing device to:
generate training data for the machine learning model, wherein generating the training data comprises:
generate first training input, the first training input comprising a first number of reports at a first point in time, wherein the reports are submitted by the users of the platform and identify incidents where content of the platform violates a policy of use associated with the platform;
generate second training input, the second training input comprising a number of resources active at the first point in time;
generate first target output for the first training input and the second training input, wherein the first target output identifies a number of resources sufficient to evaluate the target percentage of the first number of reports; and
provide the training data to train the machine learning model on (i) a set of training inputs comprising the first training input and the second training input, and (ii) a set of target outputs comprising the first target output.

### Claim 20 (depends on 19)

The system of claim 19, wherein to generate training data for the machine learning model, the processing device further to:
generate third training input, the third training input comprising a percentage of the first number of reports that are being evaluated at the first point in time by the number of resources active at the first point in time; and
wherein the set of training inputs comprising the first, the second, and the third training input.
