AI-Olympics: Exploring the Generalization of Agents through Open Competitions

TL;DR

AI-Olympics hosted from 2021 to 2023 provided a platform for testing AI agents' ability to generalize across diverse sports tasks in a competitive environment, offering insights into multi-agent decision-making.

Contribution

This paper introduces a series of competitions that evaluate the generalization of AI agents in multi-agent sports scenarios, highlighting engineering approaches and key findings.

Findings

Agents demonstrated varying levels of generalization across tasks

Engineering efforts improved agent robustness and adaptability

The competition series revealed challenges in multi-agent coordination

Abstract

Between 2021 and 2023, AI-Olympics, a series of online AI competitions was hosted by the online evaluation platform Jidi in collaboration with the IJCAI committee. In these competitions, an agent is required to accomplish diverse sports tasks in a two-dimensional continuous world, while competing against an opponent. This paper provides a brief overview of the competition series and highlights notable findings. We aim to contribute insights to the field of multi-agent decision-making and explore the generalization of agents through engineering efforts.