Meta Platforms Technologies Products Meta Horizon Worlds World Recommendations Notifications AI system

UPDATED APR 9, 2025
The content you see on your Meta Platforms Technologies Products Meta Horizon Worlds World Recommendations Notifications is selected, ranked and delivered to you by an artificial intelligence (AI) system. Within one AI system, multiple machine learning models work together to deliver your experience. These models and their input signals are dynamic and they change frequently as the system learns and improves over time.

Overview of Meta Platforms Technologies Products Meta Horizon Worlds World Recommendations Notifications

When you view and interact with Meta Horizon Worlds, one of the underlying AI systems will send you World recommendations from the Meta Horizon Worlds App that might interest you.
How Meta Platforms Technologies Products Meta Horizon Worlds World Recommendations Notifications works
The AI system behind Meta Horizon Worlds Notifications automatically finds Worlds you might be interested in and sends you notifications to join. Here’s how it works:
  1. Gather Inventory
    First, the system gathers content eligible to be recommended. This includes Worlds that might interest you, based on Worlds you have previously visited, your Worlds that might be popular with other users, and how experienced you are in Meta Horizon Worlds or in Virtual Reality.
  2. Apply Filters
    Next, the system applies integrity processes to remove content that is non-recommendable for quality reasons or that may go against our Code of Conduct for Virtual Experiences. We also remove worlds rated 18+ for certain age groups.
  3. Score Worlds
    After filtering Worlds, the system individually scores the Worlds that have not yet been sent to you. Scores are produced based on a variety of factors, such as previous World visits, similarity to other items, how well the World matches others you tend to interact with and how many people are in this World.
  4. Rank Worlds by score
    After scoring the Worlds, the system orders all the Worlds to pick the one with the best score to recommend to you.
  5. Build Notification
    After ranking the Worlds by score, the system picks which World to recommend and generates a candidate notification. This notification attaches a recommendation message to the world recommendation.
  6. Filter Notifications
    Finally, we filter out candidate notifications based on a variety of criteria, including user-level notification settings and previously sent notifications.
How to customize what you see
Your experience on Meta Platforms Technologies Products Meta Horizon Worlds World Recommendations Notifications is personalized based on your activity, and you have options to control or customize what you see. Below, we describe how to do this with different in-product features. Options shown here may not be available to everyone.
Follow
You can follow other users in Meta Horizon Worlds and see which Worlds they are currently active at if they choose to make their presence visible. This information is available on the shelf with Friends Active Worlds. The system also uses this signal to rank the Worlds that are suggested to you. When you unfollow a user, your recommendations may change.
Save
You can save a World to your own collection. This can help the system show you similar content in the future.
Dismiss
You can dismiss any World notification you’re not interested in. This helps to ensure we won’t recommend similar Worlds to you.
Report
You can report other users in the world if you think something goes against our Community Guidelines for violating the Code of Conduct for Virtual Experiences and the Meta Horizon Worlds Worlds Mature and Prohibited Worlds Policy and report a World as a whole.
Visit
You can visit a World to see recommendations of similar Worlds.
Like
You can like a World. This can help the system show you similar content in the future.
How the AI delivers content to you
We want you to see content you enjoy and find interesting. To achieve this, the AI system has models that help it make predictions about content you'll find most relevant and valuable. These prediction models use underlying input signals to help select content you're most likely to engage with.
Below are some of the significant predictions–and input signals that inform them–that we use in this AI system.
How likely you are to click on the World notification
Signals influencing this prediction include:
  • The number of encounters and high-fives you have made with other users across all Worlds
  • The number of clicks, visits and saves you have made across all Worlds, and the amount of time you have spent in each World
  • The number of clicks, visits and saves other Meta Horizon Worlds users have made across all Worlds
  • How long you have been using Horizon Worlds
  • The genre, description, and other basic information of all Worlds
How likely you are to spend more than 5 minutes in the World
Signals influencing this prediction include:
  • The number of times you see World recommendations within Meta Mixed Reality (MR) Horizon Feed and Meta Horizon Worlds
  • The genre, description, and other basic information of all Worlds
  • The number of encounters and high-fives you have made with other users across all Worlds
  • The number of clicks, visits, time spent and saves you have made across all Worlds and the number of clicks, visits and saves other Worlds you have made across all Worlds
  • How long you have been using Horizon Worlds