AI Researchers, Video Games Are Your Friends!
If you are an artificial intelligence researcher, you should look to video games as ideal testbeds for the work you do. If you are a video game developer, you should look to AI for the technology that makes completely new types of games possible. This cha
- PDF / 194,548 Bytes
- 16 Pages / 439.37 x 666.142 pts Page_size
- 88 Downloads / 187 Views
Abstract. If you are an artificial intelligence researcher, you should look to video games as ideal testbeds for the work you do. If you are a video game developer, you should look to AI for the technology that makes completely new types of games possible. This chapter lays out the case for both of these propositions. It asks the question “what can video games do for AI”, and discusses how in particular general video game playing is the ideal testbed for artificial general intelligence research. It then asks the question “what can AI do for video games”, and lays out a vision for what video games might look like if we had significantly more advanced AI at our disposal. The chapter is based on my keynote at IJCCI 2015, and is written in an attempt to be accessible to a broad audience. Keywords: Artificial intelligence gence
1
·
Games
·
Artificial general intelli-
Introduction
Video games and artificial intelligence are two of my favorite topics. Both as work and hobby. The great thing is that they go together so well: there is a great need for video games in artificial intelligence and for artificial intelligence in video games. In this chapter, I discuss what video games can do for AI and what AI can do for video games. In the first part, I discuss the need for benchmarks in AI research and how games have historically been used as AI benchmarks. I then argue the advantages of video games over classic board games as AI benchmarks, and in particular the advantages of general video game playing. I present the general video game playing competition and benchmark, and the vision of having games both generated and played automatically. I discuss how this fits into the idea of artificial general intelligence, the idea of developing AI that is good not only at a single things but at all things, or at least most of them. In the second part of the chapter, I discuss what AI can do in and for games. Lots of things, it turns out—playing them is what most people think of first, and it is true that there is a need for skilled and interesting adversaries and other non-player characters in many games—but perhaps even more exciting is all the possibilities that AI offers for modeling players, generating levels and perhaps even whole games, adapting games to suit players, and assisting game designers. The second section is structured as a vision of what playing an open-world game c Springer International Publishing AG 2017 J.J. Merelo et al. (eds.), Computational Intelligence, Studies in Computational Intelligence 669, DOI 10.1007/978-3-319-48506-5 1
4
J. Togelius
might be like in a future where we have the AI technologies to truly make the game we like, followed by a brief description of some of the research challenges involved in getting there. It is important to note that this paper does not go into any technical depth on any particular topic, nor is it a comprehensive survey of the field. It is instead meant as an accessible, informal and inspirational introduction as well as a long-form argument. It is equal parts propaganda and
Data Loading...
