Botball is a national program in which teams of middle and high school students design, build, and program small autonomous mobile robots to compete in a highly charged interactive (but nondestructive) tournament. Botball students learn to program in c, construct feedback and control loops, create electromechanical systems, and integrate it all together while they work on a team. Botball takes place in regional tournaments across the country and culminates in a National Botball Tournament traditionally hosted by the Association for the Advancement of Artificial Intelligence at its annual conference. This program puts reusable equipment into schools and, at the Botball Teacher Workshops, trains teachers in robotics and the integration of robotics into their curriculum.

(IKS 2002).

The AAAI Press - Distributed by The MIT Press Massachusetts Institute of Technology, 5 Cambridge Center, Cambridge, Massachusetts 02142 To order, call toll free: (800) 356-0343 or (617) 625-8569. SPRING 2002 5 first time that AAAI's National conference has been held in Canada--a In addition, the program chairs are experimenting with a new format for AAAI.

The debut of the AI in the News column elsewhere in this issue of AI Magazine created a good opportunity to introduce the professional community to the AI Topics web site, home of the AI in the news virtual page. Although AI Topics is designed for the lay public, it serves a much larger audience.

Botball is a national program in which teams of middle and high school students design, build, and program small autonomous mobile robots to compete in a highly charged interactive (but nondestructive) tournament. Botball students learn to program in c, construct feedback and control loops, create electromechanical systems, and integrate it all together while they work on a team. Botball takes place in regional tournaments across the country and culminates in a National Botball Tournament traditionally hosted by the Association for the Advancement of Artificial Intelligence at its annual conference. This program puts reusable equipment into schools and, at the Botball Teacher Workshops, trains teachers in robotics and the integration of robotics into their curriculum. Botball appeals to a wide variety of students and brings out the best in each, challenging them to solve realworld problems in a dynamic environment at their own level.

"Neats and scruffies alike were mesmerized by the animal-like responses of the robots demonstrated there," says Bonasso. "At the end of "This won't be a slick, polished competition. Over the AAAI Mobile Robot Competition and Exhibition years, the event and AI Magazine have served as was born. The event has endured to a venue for this and several other intellectual become the oldest AIcentric robotics competition debates, including sensing versus modeling, in the world. As we near the end of our color-based versus shape-based object recognition, first decade, it seems worthwhile to reflect on and reactive control versus symbolic what the origins of the event were, how it has planning for robot navigation (Balch et al. evolved, and where it is headed. The contest immediately took on two important but apparently conflicting roles: First, it provided a target for research in AI and robotics; in Pete Bonasso's words, the event was cast "in the spirit of trying to develop as animate, responsive, and intelligent robot behavior as possible" (Dean and Bonasso 1993).

We present an expressive agent design language for reinforcement learning that allows the user to constrain the policies considered by the learning process.The language includes standard features such as parameterized subroutines, temporary interrupts, aborts, and memory variables, but also allows for unspecified choices in the agent program. For learning that which isn't specified, we present provably convergent learning algorithms. We demonstrate by example that agent programs written in the language are concise as well as modular. This facilitates state abstraction and the transferability of learned skills. 1 Introduction The field of reinforcement learning has recently adopted the idea that the application of prior knowledge may allow much faster learning and may indeed be essential if realworld environments are to be addressed. For learning behaviors, the most obvious form of prior knowledge provides a partial description of desired behaviors. Several languages for partial descriptions have been proposed, including Hierarchical Abstract Machines (HAMs) [8], semi-Markov options [12], and the MAXQ framework [4]. This paper describes extensions to the HAM language that substantially increase its expressive power, using constructs borrowed from programming languages. Obviously, increasing expressiveness makes it easier for the user to supply whatever prior knowledge is available, and to do so more concisely.

We present an expressive agent design language for reinforcement learning that allows the user to constrain the policies considered by the learning process.The language includes standard features such as parameterized subroutines, temporary interrupts, aborts, and memory variables, but also allows for unspecified choices in the agent program. For learning that which isn't specified, we present provably convergent learning algorithms. We demonstrate by example that agent programs written in the language are concise as well as modular. This facilitates state abstraction and the transferability of learned skills. 1 Introduction The field of reinforcement learning has recently adopted the idea that the application of prior knowledge may allow much faster learning and may indeed be essential if realworld environments are to be addressed. For learning behaviors, the most obvious form of prior knowledge provides a partial description of desired behaviors. Several languages for partial descriptions have been proposed, including Hierarchical Abstract Machines (HAMs) [8], semi-Markov options [12], and the MAXQ framework [4]. This paper describes extensions to the HAM language that substantially increase its expressive power, using constructs borrowed from programming languages. Obviously, increasing expressiveness makes it easier for the user to supply whatever prior knowledge is available, and to do so more concisely.

Igor V. Cadez Information and Computer Science University of California Irvine, CA 92697-3425, U.S.A. PadhraicSmyth Information and Computer Science University of California Irvine, CA 92697-3425, U.S.A. Abstract We investigate a general characteristic of the tradeoff in learning problems between goodness-of-fit and model complexity. Specifically wecharacterize a general class of learning problems where the goodness-of-fit function can be shown to be convex within firstorder asa function of model complexity. This general property of "diminishing returns" is illustrated on a number of real data sets and learning problems, including finite mixture modeling and multivariate linear regression. 1 Introduction, Motivation, and Related Work Assume we have a data set D Such learning tasks can typically be characterized by the existence of a model and a loss function. A fitted model of complexity k is a function of the data points D and depends on a specific set of fitted parameters B. The loss function (goodnessof-fit) isa functional of the model and maps each specific model to a scalar used to evaluate the model, e.g., likelihood for density estimation or sum-of-squares for regression. Figure 1 illustrates a typical empirical curve for loss function versus complexity, for mixtures of Markov models fitted to a large data set of 900,000 sequences.

We present an expressive agent design language for reinforcement learning thatallows the user to constrain the policies considered by the learning process.Thelanguage includes standard features such as parameterized subroutines,temporary interrupts, aborts, and memory variables, but also allows for unspecified choices in the agent program. For learning that which isn't specified, we present provably convergent learning algorithms. Wedemonstrate by example that agent programs written in the language are concise as well as modular. This facilitates state abstraction and the transferability of learned skills. 1 Introduction The field of reinforcement learning has recently adopted the idea that the application of prior knowledge may allow much faster learning and may indeed be essential if realworld environmentsare to be addressed. For learning behaviors, the most obvious form of prior knowledge provides a partial description of desired behaviors. Several languages for partial descriptions have been proposed, including Hierarchical Abstract Machines (HAMs) [8], semi-Markov options [12], and the MAXQ framework [4]. This paper describes extensions to the HAM language that substantially increase its expressive power,using constructs borrowed from programming languages. Obviously, increasing expressivenessmakes it easier for the user to supply whatever prior knowledge is available, and to do so more concisely.