Application of adaptive game-based learning in image interpretation
This paper presents adaptive e-learning to a map-based serious game in aerial image interpretation which trains both geographic knowledge as well as the identification of relevant objects. We attach an interoperable adaptivity framework, called "E-Learning A.I." (ELAI) which adjusts the game's difficulty and the provided amount of assistance to match the learners need. The underlying software architecture uses the Experience API (xAPI) to collect, store and analyze the captured usage data. The scientific research questions affect the possible usages of the collected interaction data and how to manifest adaptivity in serious games. A study has been conducted to answer questions regarding the acceptance and effects of adaptivity in map-based seek-and-find games for aerial image interpretation. The three main hypotheses cover the questions on the acceptance of adaptivity mechanisms, on the users' motivation and on the learning outcome effectivity. Although the adaptivity has been recognized, the dynamically inserted virtual agent was rather seen as a disturbance to the game flow. The study verifies the feasibility of the adaptivity framework and reveals issues on the study design of similar adaptive systems. The ELAI software architecture consists of generic and specific components. The most specific component is the so called ELAI-adapter which basically captures and adapts attached games or computer simulations. Obviously this adapter has to be specific to the actual game, genre, characters etc. In our application example we have implemented the ELAI-adapter for the Unity game engine. The xAPI protocol allows to easily connect other xAPI-compliant serious games or e-learning course software. The ELAI-controller interprets the collected usage data and computes influence strategies for the attached games or simulations, which in turn are then realized in the games itself by the ELAI-adapter.