# US 2020/0391118 A1 · Predictive data preloading

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

- Number: US 2020/0391118 A1
- Document ID: US20200391118A1
- Status: Published application
- Kind: A1
- Application number: 16/442,328
- Filed: 2019-06-14
- Published: 2020-12-17
- Earliest claimed: 2019-06-14
- Assignees: Roblox Corporation
- Applicants: Roblox Corporation
- Inventors: John Kerr
- Web page: https://patentgrep.com/patent/US20200391118A1

## CPC

- A63F 13/79: involving player-related data, e.g. identities, accounts, preferences or play histories
- H04L 67/306: User profiles
- A63F 13/358: Adapting the game course according to the network or server load, e.g. for reducing latency due to different connection speeds between clients
- A63F 13/45: Controlling the progress of the video game
- A63F 13/48: Starting a game, e.g. activating a game device or waiting for other players to join a multiplayer session
- A63F 13/67: adaptively or by learning from player actions, e.g. skill level adjustment or by storing successful combat sequences for re-use
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/0895: Weakly supervised learning, e.g. semi-supervised or self-supervised learning
- G06N 3/09: Supervised learning
- G06N 3/092: Reinforcement learning
- A63F 13/533: for prompting the player, e.g. by displaying a game menu
- A63F 13/63: by the player, e.g. authoring using a level editor
- A63F 13/77: involving data related to game devices or game servers, e.g. configuration data, software version or amount of memory
- A63F 2300/534: for network load management, e.g. bandwidth optimization, latency reduction
- A63F 2300/5546: using player registration data, e.g. identification, account, preferences, game history
- A63F 2300/636: involving process of starting or resuming a game
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- H04L 67/131: Protocols for games, networked simulations or virtual reality
- H04L 67/34: {involving the movement of software or configuration parameters (network booting or remote initial program loading [RIPL] G06F9/4416)}
- A63F: CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Methods, systems, and computer readable media for predictive preloading of game data are described. In some implementations, a method can include using a machine learning model of player cohort behavior to predict which games (or game-related data assets) to begin preloading prior to a user making a game selection. Once the user makes a selection, any preloading not related to the selection can be terminated. Thus, during the time period when the user is viewing available game selections, a given number of selections can begin to be preloaded based on predictions from the machine learning model. The preloading can help reduce latency from game selection by the user to game start time.

## Claims

### Claim 1

A computer-implemented method to train a machine-learning model to predict game options to preload in a gaming platform, the machine-learning model including a neural network, the method comprising:
providing training data that includes player data corresponding to a player and previous selection data corresponding to one or more game options previously selected by the player in the gaming platform, wherein the one or more previously selected game options was selected from one or more available game options; and
training the neural network to predict game options to preload, wherein the training comprises:
providing the player data and the one or more available game options as input to the neural network;
obtaining, as output of the neural network, one or more predicted game options to preload, wherein the neural network comprises one or more of: weighted nodes or weighted connections;
comparing the one or more predicted game options with the one or more previously selected game options in the previous selection data that corresponds to the one or more available game options; and
adjusting one or more parameters of the neural network based on the comparing, wherein the preloading is performed based upon one or more of: available memory of a client-device or network capacity between the client-device and the gaming platform.

### Claim 2

(canceled)

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein the training is performed until a loss function meets a predetermined criterion.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein the player data includes one or more of player gender, player age, player geographic region, player language, player experience on the gaming platform, or player purchase history on the gaming platform.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, further comprising:
determining a cohort that the player is associated with; and
training the neural network to predict one or more game options to preload based on the cohort that the player is associated with.

### Claim 6

(canceled)

### Claim 7

(canceled)

### Claim 8

A system, comprising:
one or more processors coupled to a nontransitory computer readable medium having stored thereon software instructions that, when executed by the one or more processors, cause the one or more processors to perform operations including:
providing training data that includes player data corresponding to a player and previous selection data corresponding to one or more game options previously selected by the player in a gaming platform, wherein the one or more previously selected game options was selected from one or more available game options; and
training a neural network to predict game options to preload, wherein the training comprises:
providing the player data and the one or more available game options as input to the neural network;
obtaining, as output of the neural network, one or more predicted game options to preload;
comparing the one or more predicted game options with the one or more previously selected game options in the previous selection data that corresponds to the one or more available game options; and
adjusting one or more parameters of the neural network based on the comparing, wherein the preloading is performed on one or more of: a client-device or a server-device.

### Claim 9 (depends on 8)

The system of claim 8, wherein the player data includes one or more of player gender, player age, player geographic region, player language, player experience on the gaming platform, or player purchase history on the gaming platform.

### Claim 10 (depends on 8)

The system of claim 8, wherein the operations further comprise:
determining a cohort that the player is associated with; and
training the neural network to predict one or more game options to preload based on the cohort that the player is associated with.

### Claim 11

A computer-implemented method to preload one or more game options in a gaming platform, the method comprising:
providing player data and one or more available game options as input to a machine-learning model;
training the machine learning model using the player data from the gaming platform;
obtaining, as output of the trained machine learning model, one or more game options to preload in the gaming platform; and
preloading at least one of the one or more game options prior to receiving user selection input of a particular game, wherein the preloading is performed during one or more of: a low processor utilization state or a low communications system utilization state.

### Claim 12 (depends on 11)

The computer-implemented method of claim 11, wherein the preloading comprises:
causing preloading to start at a player system, wherein the preloading includes preloading one or more resources corresponding to the one or more game options to preload; and
terminating the preloading of the one or more game options in response to a termination condition, wherein the terminating comprises release of memory resources used for the preloading.

### Claim 13 (depends on 12)

The computer-implemented method of claim 12, wherein the termination condition includes receiving a game option selection from the player system, and wherein the terminating includes termination of preloading of game options that do not match the game option selection.

### Claim 14 (depends on 12)

The computer-implemented method of claim 12, wherein the termination condition includes receiving user input indicative of a change in the available game options.

### Claim 15 (depends on 12)

The computer-implemented method of claim 12, wherein the termination condition includes receiving a new gaming platform search query.

### Claim 16

(canceled)

### Claim 17

(canceled)

### Claim 18 (depends on 11)

The computer-implemented method of claim 11, wherein the one or more available game options include one or more available games selectable by a player within the gaming platform.

### Claim 19 (depends on 11)

The computer-implemented method of claim 11, wherein the one or more available game options include one or more available game resources selectable by a player within a particular game in the gaming platform.

### Claim 20 (depends on 11)

The computer-implemented method of claim 11, wherein the player data includes one or more of player gender, player age, player geographic region, player language, player experience on the gaming platform, or player purchase history on the gaming platform.

### Claim 21 (depends on 12)

The computer-implemented method of claim 12, wherein the termination condition is based upon at least one or more of: preloading of a threshold number of the one or more resources, preloading of a predetermined number of the one or more game options, consumption of a predetermined amount of memory, or consumption of a threshold amount of a network resource.

### Claim 22 (depends on 1)

The computer-implemented method of claim 1, wherein the training comprises one or more of: supervised learning, unsupervised learning, or semi-supervised learning techniques.

### Claim 23 (depends on 1)

The computer-implemented method of claim 1, wherein the training data comprises one or more of: a multidimensional vector, or a feature vector further comprising one or more of: player demographic information, player cohort data, player history, player location, or player game level, and wherein the neural network is trained using one or more of the player data or the cohort data.

### Claim 24 (depends on 8)

The system of claim 8, wherein the preloading is performed on a player system in a background prior to the player making a new game option selection, and wherein the preloading includes one or more of: retrieving one or more resources from the gaming platform via a communications network, loading the one or more resources in a random access memory of the player system, or caching the one or more resources at the gaming platform.

### Claim 25 (depends on 8)

The system of claim 8, wherein the one or more available game options include one or more of: games available to play, items included within the games, resources within the games, movement options available within the games, or action options available within the games.
