Alien game launched!

“When you play a game — If you learn to be good at it — you find what it is you should have been thinking about.” — John Conway

Since human beings have the ability to learn and “find out”, humans are much better than computers to learn how to solve complex problems and e.g. become very good at a computer game.

But, we still know surprisingly little about the strategies that humans use, in order to learn how to solve a new problem. If we know the nature of the problem, it is just about getting from A to B, in a straight line, of course, but how do people cope with not having (enough) information about where B, i.e. the goal, is?

One of the most intuitive metaphors in problem-solving, put forward by the pioneer of cognitive science and artificial intelligence, Simon Herbert, illuminates the issue: Simon likened problem-solving with a search through a landscape with a hidden treasure (Simon 1983). You don’t know where the treasure is, but each move gives you a bit more information and allows you to make decisions about where to move next. If it is a very easy problem you can easily skip from location to location, quickly getting better and better and making your way towards the final goal (more or less following the A-B straight line, even if you don’t know where B is!). However, for more complex problems, this might not be the optimal strategy, because you will get ambiguous feedback. What do we mean by ambiguous feedback? There’s an old story of three blind men and an elephant. Each of them touches a different part of the elephant and reaches wildly different conclusions: it’s a brush, a pillar or a plough. Before you laugh, think about this: It’s not trivial when you touch something that resembles a brush to figure out that you are really dealing with the tip of the tail of an elephant!

People have different approaches with respect to how they rely on feedback and past experience to generate new solutions. For instance, when people learn, they not only “record” feedback, but they are also able to interpret it, code it, and sometimes discard information.

We know for example that:

• people often use abstraction or analogies when attempting to solve novel problems (e.g.Veloso 1992)
• their strategy choice is often influenced by emotions elicited by positive/negative feedback (Spering et al 2005)
• and even that they tend to default to their most common solutions when they are hungry! (Linder et al 2014)

Still, despite decades of research, we lack a proper understanding of how humans really go about solving problems and we could use your help. In the end, by solving the complex Alien game, you can help us solve a really complex problem.

PLAY NOW !

Herbert A. Simon, Search and Reasoning in Problem Solving. Artif. Intell. 21(1-2): 7-29 (1983)
Linder, J. A., et al. (2014). “Time of Day and the Decision to Prescribe Antibiotics.” JAMA 174(12): 2029-2031.
Spering, Miriam, Peter A. Frensch, and Joachim Funke. “The role of emotions in complex problem-solving.” Cognition and Emotion 19 (2005): 1252-1261.
Veloso, Manuela M. Learning by analogical reasoning in general problem solving. No. CMU-CS-92-174. CARNEGIE-MELLON UNIV PITTSBURGH PA DEPT OF COMPUTER SCIENCE, 1992.