Science, AI, and (Video)games: An approach through decision-making research

  • LECTURER: Florian Richoux
  • AFFILIATION: National Institute of Advanced Industrial Science and Technology, Tokyo

Outline

Games are wonderful environments to develop and improve AI methods: They can propose a rich, dynamic environment in which decision-making is challenging, raising numerous scientifically pertinent questions and problems.

In this lecture, we will start by exposing some arguments for choosing games as AI test beds. We will then focus on decision-making problems through the lens of combinatorial optimization, and after a quick introduction of Constraint Programming, we will study some decision-making problems that arise from real-time strategy games like StarCraft. We will start by some simple problems, to finish by some more complex ones dealing with the uncertain nature of the game environment. Finally, we will conclude the lecture by giving a lead for people willing to start doing research in Game AI.

Assessment Method

Attendance and an on-line questionnaire.

Lective hours

2

Timetable

  • 15 December, 15:00-17:00

Lecture Theatre

  • Salón de Actos (building 1, ground floor)

Tuition Language

English.