What Data Scientists Should Know About Game Theory: Types of Games

Game theory is one of the foundational building blocks of our behavior as social beings as well as many of our behavioral patterns. In the context of artificial intelligence(AI) and deep learning systems, game theory is essential to enable some of the key capabilities required in multi-agent environments in which different AI programs need to interact or compete in order to accomplish a goal.

In the past, I’ve written about the implications of game theory in AI systems. The famous Nash equilibrium popularized by the movie “A Wonderful Mind” is the cornerstone of many AI interactions in modern systems. However, modeling an AI universe using the principles of game theory many times goes beyond the Nash equilibrium. A good place to start understanding the implications architecting AI systems using principles of game theory is to understand the different types of games that we typically encountered in our social or economic interactions.

Any days, we participate in hundreds of interactions that are based on game dynamics. However, the architecture of those gamified environments are completely different and so are the incentives and goals of the participant. How to apply some of principles to the modeling of AI agents? Well, the first step is to identify the nature of the game we are trying to create:

5 Types of Games that Rule the AI World

Suppose that we are modeling an AI system that involves multiple agents that will interact and compete to accomplish a specific goal. That’s a classic example of game theory. Since its inception in 194, game theory has focused on modeling the most common interaction patterns that now we are seeing every day in multi-agent AI systems. Here is a taxonomy that might help you identify some of the most relevant types of games that have an equivalent in the AI world :

Symmetric vs. Asymmetric.

One of the simplest classifications of games is based on their symmetry. A symmetric game describes an environment in which each player has the same goals and the results will only depend on the strategies involved. Chess is a classic example of a symmetric game. Many of the situations we encountered in the real world lack the mathematical elegance of symmetry as participants often have different and even conflicting goals. A business negotiation is an example of asymmetric game in which each party has different goals and evaluates the results from a different perspective (ex: winning a contract vs. minimizing an investment).

Perfect vs. Imperfect Information

Another important categorization of games is based on the type of information available. A perfect information game refers to an environment in which each player can see the other player’s moves. Chess, again, is an example of a perfect information game. Many modern interactions are based on environments in which the moves from each player are hidden from other players and game theory classifies those scenarios as imperfect information games. From card games like poker to self-driving car scenarios, imperfect information games are all around us.

Cooperative vs. Non-Cooperative

A cooperative game environment is one in which the different participants can establish alliances in order to maximize the end result. Contractual negotiations are often modeled as cooperative games. Non-cooperative scenarios describe environments in which players are forbidden from forming alliances. Wars are the ultimate example of non-cooperative games.

Simultaneous vs. Sequential

A sequential game takes place in an environment in which each player has information about the other player earlier actions. Board games are mostly sequential in nature. Simultaneous games represent scenarios in which both players can take concurrent actions. Securities trading is an example of simultaneous games.

Zero-Sum vs. Non-Zero-Sum

A zero-sum game refers to a scenario in which the gains or one player always come translate into looses for other players. Board games are examples of zero-sum games. Non-zero-sum games are often encountered in scenarios in which multiple players can benefit from the actions of one players. Economic interactions in which multiple participants collaborate to increase the size of the market is an example of a non-zero-sum game.

Source: Deep Learning on Medium