A Multimodal Architecture for Endpoint Position Prediction in Team-based Multiplayer Games

TL;DR

This paper introduces a multimodal, U-Net-based architecture utilizing multi-head attention to predict future player locations in multiplayer games, integrating diverse data types for improved accuracy and enabling advanced game analytics.

Contribution

The paper presents a novel multimodal architecture that effectively combines heterogeneous game data for accurate endpoint position prediction in complex multiplayer environments.

Findings

Achieved accurate future player location predictions.

Effectively integrated image, numerical, and categorical data.

Enabled downstream tasks like bot behavior modeling and anomaly detection.

Abstract

Understanding and predicting player movement in multiplayer games is crucial for achieving use cases such as player-mimicking bot navigation, preemptive bot control, strategy recommendation, and real-time player behavior analytics. However, the complex environments allow for a high degree of navigational freedom, and the interactions and team-play between players require models that make effective use of the available heterogeneous input data. This paper presents a multimodal architecture for predicting future player locations on a dynamic time horizon, using a U-Net-based approach for calculating endpoint location probability heatmaps, conditioned using a multimodal feature encoder. The application of a multi-head attention mechanism for different groups of features allows for communication between agents. In doing so, the architecture makes efficient use of the multimodal game state including image inputs, numerical and categorical features, as well as dynamic game data. Consequently, the presented technique lays the foundation for various downstream tasks that rely on future player positions such as the creation of player-predictive bot behavior or player anomaly detection.