The Generalist Approach to Frame Problems
Shohei Hidaka, Neeraj Kashyap · Institutional Repositories DataBase (IRDB) · 2014
The frame problem commonly appears in the field of artificial intelligence and in related philosophical literature. It pertains to the difficulty of making meaningful inferences in dynamic contexts. We discuss the frame problem and its manifestations in the practical design of artificial agents, and we set out some criteria against which the effectiveness of such an agent may be judged. Finally, we claim that technology currently exists in the form of a new clustering technique - dimensional clustering - which can enable an agent to satisfy these criteria. We justify this claim with three case studies where we apply dimensional clustering to data sets of varied nature.