The sensemaking task in investigative analysis generates stories that connect entities and events in an input stream of data. The Stab system represents crime stories as hierarchical scripts with goals and states. It generates multiple stories as explanatory hypotheses for an input data stream containing interleaved sequences of events, recognizes intent in a specific event sequence, and calculates confidence values for the generated hypotheses. In this report, we describe Stab2, a new interactive version of the knowledge‐based Stab system. Stab2 contains a story editor that enables users to enter and edit crime stories. We illustrate Stab2 with examples from the IEEE VAST contest datasets.

We examine the use of teleological metareasoning for self-adaptation in game-playing software agents. The goal of our work is to develop an interactive environment in which the game designer generates requirements for a new version of a game, and the legacy software agents from previous versions of the game adapt themselves to the new game requirements in cooperation with the human designer, who provides guidance where automation is not possible or not implemented. We are developing and testing our metareasoning technique for adapting a mature program in the domain of turn-based, multi-player strategy games, specifically FreeCiv (www.freeciv.wikia.com). In this paper, we first present an analysis of adaptations to FreeCiv, next describe our general approach, then describe a specific adaptation scenario, and finally discuss our plans for future work.

We examine the problem of self-adaptation in game-playing agents as the game requirements evolve incrementally. The goal of our current work is to develop an interactive environment in which the game designer generates requirements for a new version of a game, and the legacy software agents from previous versions of the game adapt themselves to the new game requirements. We are developing and testing our metareasoning technique for adapting a game-playing agents in Freeciv, a mature program in the domain of turn-based, multi-player strategy games. In this paper, we first present an analysis of adaptations to FreeCiv, next describe our general approach, and then describe a specific adaptation scenario.

We examine the use of teleological metareasoning for self-adaptation in game-playing software agents. The goal of our work is to develop an interactive environment in which the game designer generates requirements for a new version of a game, and the legacy software agents from previous versions of the game adapt themselves to the new game requirements. We are developing and testing our metareasoning technique for adapting game-playing agents in Freeciv, a mature program in the domain of turn-based, multi-player strategy games (www.freeciv.wikia.com). In this paper, we first present an analysis of adaptations to FreeCiv, next describe our general approach, and then describe a specific adaptation scenario.

As the task environment of a software artifact evolves, so must its design. For example, as the task environment in a computer game evolves, so must the design of the software agent that plays the game (or the agent’s behavior is likely to become more suboptimal than before). We are exploring how a software artifact may adapt itself as it’s task environment evolves incrementally. In particular, we are investigating how a game-playing agent may adapt itself as the percepts, actions, rules and constraints of its environment evolve from one version of the game to the next. A core research question in our work is what must an agent know about its design so that it can identify and make the right self-modifications to meet the needs of the new task environment? Our hypothesis is that the agent’s self-knowledge of its teleology (i.e., the mechanisms by which it’s design achieves its functions) may support the process of self-adaptation. In this paper, we describe the preliminary design of an interactive environment called GAIA in which a human game engineer and a game-playing software agent cooperatively adapt the agent’s software design and program code. As the game-playing agent uses it’s self-knowledge of it’s teleology to identify modifications to it’s design and code, the game engineer may (or may not) accept specific modifications and thus guide the process of self-adaptation. We also illustrate a first, simple example from FreeCiv, an interactive turn-based strategy game, at a high-level of specification. In Proceedings of the Third International Conference on Design Science Research in Information Systems and Technology. (V. Vaishnavi & R. Baskerville, Eds). May 7-9, 2008, Atlanta, Georgia: Georgia State University. c ©The Authors 2008. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee. 1 Background, Motivation and Goals The design of a long-living software artifact evolves through many versions. Changes in the requirements from one version to the next typically are incremental and sometimes quite small (deltas). A software engineer (or a team of software engineers) formulates the requirements of a new version, adapts the design of the previous versions to meet the new requirements, implements and evaluates the modified design. Of course, the ordering of these tasks is not necessarily linear; the requirements, for example, may evolve during the design episode, and if the proposed design fails in the evaluation task, it may need to be redesigned. Thus, adaptive1 design includes both proactive adaptation (adapting a design to meet new requirements) and retrospective adaptation (redesigning a proposed design). In previous work, we have addressed both retrospective [1] [2] [3] and proactive adaptations [4] [5] [6] [7]. The earlier work had led to a knowledge representation language (called TMKL) for capturing a software agent’s knowledge of its teleology, and a knowledge-based engine (called REM) for adapting the agent’s design.2 The ongoing work described here began with a case study of changes made to an open-source game [8]. The current project focuses on proactive adaptations to the design of game-playing software agents as their game environment evolves incrementally, e.g., incremental changes to the percepts and actions in the game or to the rules and constraints of the game. The earlier work on REM and TMKL also leads to our research hypothesis for the current project: teleology, the explicit connection of functions to goals, is a basic organizational principle of adaptive software design. 2 Game-Playing Domain The domain for initial experimentation is computer-based strategy games, and the first case study examines FreeCiv3. FreeCiv is an open source variant of a class of Civilization games with similar properties. The aim in these games is to build an empire in a competitive environment. The major tasks in this endeavor are exploration of the randomly initialized game environment, resource allocation and development, and warfare that may at times be either offensive or defensive in nature. Winning the game is achieved most directly by destroying the civilizations of all opponents, but can also be achieved through more peaceful 1 The term adaptation has several meanings. In the Software Engineering community, it is largely synonymous with porting to a new hardware platform or software environment. In the Artificial Intelligence community, however, it normally means changes made to alter functionality. We will try to make clear by context which of the two interpretations we intend. 2 Agent is an AI term referring to software viewed as receiving percepts from an environment and taking actions in that environment; thus, the term connotes a perspective on a software system. 3 http://www.freeciv.org

In this paper, we examine the use of teleological metarea- soning for self-adaptation in game-playing software agents. The goal of our current work is to develop an interactive environment in which the game designer generates require- ments for a new version of a game, and the legacy software agents from previous versions of the game adapt themselves to the new game requirements in cooperation with the hu- man designer, who provides guidance where automation is not possible or not implemented. We are developing and testing our metareasoning technique for adapting a mature program in the domain of turn-based, multi-player strategy games, specifically FreeCiv (www.freeciv.wikia.com). In this paper, we first present an analysis of adaptations to FreeCiv, next describe our general approach, then describe a specific adaptation scenario, and finally discuss our plans for future work.