 Hi everyone, this is Alice Gao. In this video, I'll continue to introduce game theory and mechanism design. In the previous video, I showed you a few examples of game theory. Now let's talk about these two concepts more specifically. So what kind of problems are we really trying to solve with these two areas? Well, for game theory, we're usually given the rules of a game and then we're asking the question of how would the agent's players in the game play this game? How would they behave? So again, the direction is that we're given the rules and then we're trying to analyze and figure out the agents, how would agents play the game? What strategies are the agents going to use? Now for mechanism design, the question is a little bit reversed. So mechanism design says we want to design a game in a way such that the agents will behave in a particular way. So given that we want the agents to behave in a particular way, how should we design the rules of the game? So you can see that the two are a little bit reversed of each other. Often, mechanism design is referred to as reverse game theory. Both game theory and mechanism design are actually quite broad. There are a lot of small topics in each of these. So in this course, we only have time to cover a tiny bit of game theory and that's all. We have no time to go into mechanism design, unfortunately. Let's now make our multi-agent framework or environment more specific. So in this multi-agent framework, how should we think about it? Well, there are multiple agents in this environment and each agent is trying to decide on what strategies they should use to play this game. But how can they make their decision? What do their decisions based on? Well, their decision will be based on the information they have about the environment. This is the thing that the passive things, things that don't change in the environment. They also have information about other agents out there. These are the other intelligent agents that will make their own decisions and choose their own strategies. And finally, the agent will have their utility functions. So their decisions are going to be based on all of these information. And then once each agent chooses what they want to do, what strategy they want to play, then the outcome of the entire game is not going to depend on only one agent. In fact, the outcome is going to depend on the actions or the strategies of all of the agents. So all of the agent strategies will jointly determine what will happen. This is like life, right? In life, there are many people acting in this world and then the outcome, the welfare of the world, the welfare of everyone will depend on how people jointly act. Now, when you think about a multi-agent environment with multiple agents acting, you might be tempted to think that all of the agents are competing with each other. That's not necessarily the case, right? So in the multi-agent scenario, all we know is that each agent has their own interest, has their own goal, and then they are trying to act in a way to achieve their own goal to maximize their own utility. However, the relationship between the utility functions of the multiple agent, there can be many types of such relationships. It could be that the agents have conflicting goals, in that case, they are competitive. But it could also be that all of the agents have the same goal, in that case, they are cooperative. Or in the third case, there could be somewhere in between. So we could have an environment where we have multiple teams of agents within the team. The agents are cooperating with each other, but between teams, teams of agents are competing with another team of agents. If I were to use a board game analogy to illustrate this, most of the board games out there are probably competitive. A few of them are cooperative, such as Hanabi, Ghost Stories, Pandemic. And there are some board games out there are in between. So one example is code names. Code names you have teams of people within a team you are cooperating, and then between teams you are competing. Alright, this is the end of my short introduction to game theory. In the next video, I'm going to start talking about two-person normal form games. Thank you very much for watching. I will see you in the next video. Bye for now.