Acquiring visibly intelligent behavior with example-guided neuroevolution

Much of artificial intelligence research is focused on devis-ing optimal solutions for challenging and well-defined but highly constrained problems. However, as we begin creat-ing autonomous agents to operate in the rich environments of modern videogames and computer simulations, it becomes important to devise agent behaviors that display the visible attributes of intelligence, rather than simply performing op-timally. Such visibly intelligent behavior is difficult to spec-ify with rules or characterize in terms of quantifiable objec-tive functions, but it is possible to utilize human intuitions to directly guide a learning system toward the desired sorts of behavior. Policy induction from human-generated examples is a promising approach to training such agents. In this pa-per, such a method is developed and tested using Lamarckian neuroevolution. Artificial neural networks are evolved to con-trol autonomous agents in a strategy game. The evolution is guided by human-generated examples of play, and the system effectively learns the policies that were used by the player to generate the examples. I.e., the agents learn visibly intelli-gent behavior. In the future, such methods are likely to play a central role in creating autonomous agents for complex en-vironments, making it possible to generate rich behaviors de-rived from nothing more formal than the intuitively generated examples of designers, players, or subject-matter experts