Showing papers by "Nils J. Nilsson published in 2000"
TL;DR: A set of distinguished scholars and practitioners who were involved in AI's formative stages to describe the most notable trend or controversy during AI's development are asked.
Abstract: The transition to the next millennium gives us an opportunity to reflect on the past and project the future. In this spirit, we have asked a set of distinguished scholars and practitioners who were involved in AI's formative stages to describe the most notable trend or controversy (or nontrend or noncontroversy) during AI's development. The responses provide an interesting characterization of AI-and, in many ways, of the people of AI. We gave our contributors a great deal of flexibility in the nature of their responses. Some provided grand summaries of the history of the field as a whole. Others commented insightfully on more focused topics. Some observed changes and changed along with them. Others are still making advances on research agendas articulated presciently long ago. Some are optimistic. Others are pessimistic. Despite the range, both individually and collectively they provide insights into where we have been and where we are going. Although each contribution is a unique expression of its author's glimpse back through AI's development, there is repetition of important themes that are at the discipline's core. The article serves as an interesting record of where AI is today, as well as setting the stage for what's to come.
31 citations
01 Jan 2000
TL;DR: It is asserted that such programs are PAC learnable and then described some techniques for learning them and the results of some preliminary learning experiments are described.
Abstract: Versatile robots will need to be programmed, of course. But beyond explicit programming by a programmer, they will need to be able to plan how to perform new tasks and how to perform old tasks under new circumstances. They will also need to be able to learn. In this article, I concentrate on two types of learning, namely supervised learning and reinforcement learning of robot control programs. I argue also that it would be useful for all of these programs, those explicitly programmed, those planned, and those learned, to be expressed in a common language. I propose what I think is a good candidate for such a language, namely the formalism of teleo-reactive (T-R) programs. Most of the article deals with the matter of learning T-R programs. I assert that such programs are PAC learnable and then describe some techniques for learning them and the results of some preliminary learning experiments. The work on learning T-R programs is in a very early stage, but I think enough has been started to warrant further development and experimentation. For that reason I make this article available on the web, but I caution readers about the tentative nature of this work. I solicit comments and suggestions at:
7 citations