Self-play

Definition and motivation

In multi-agent reinforcement learning experiments, researchers try to optimize the performance of a learning agent on a given task, in cooperation or competition with one or more agents. These agents learn by trial-and-error, and researchers may choose to have the learning algorithm play the role of two or more of the different agents. When successfully executed, this technique has a double advantage:

1. It provides a straightforward way to determine the actions of the other agents, resulting in a meaningful challenge.
2. It increases the amount of experience that can be used to improve the policy, by a factor of two or more, since the viewpoints of each of the different agents can be used for learning.

Czarnecki et al argue that most of the games that people play for fun are "Games of Skill", meaning games whose space of all possible strategies looks like a spinning top. In more detail, we can partition the space of strategies into sets ${\displaystyle L_1,L_2,...,L_n}$, such that any ${\displaystyle i<j,{\pi }_i\in L_i,{\pi }_j\in L_j}$, the strategy ${\displaystyle {\pi }_j}$ beats the strategy ${\displaystyle {\pi }_i}$. Then, in population-based self-play, if the population is larger than ${\displaystyle \underset{i}{max}|L_i|}$, then the algorithm would converge to the best possible strategy.