ARCHAI, a new fully funded interdisciplinary research project of the VUB, introduces a new research framework that treats archaeological prospection as a sequential decision-making process explored through computational simulation and deep multi-objective reinforcement learning. The project will investigate a simulation platform that reproduces the structure of real prospection activities by integrating multiple, heterogeneous geospatial and archaeological data layers. Within this environment, we will design and evaluate new reinforcement learning algorithms capable of operating in large, partially observed spatial domains to infer efficient, interpretable strategies for estimating archaeological potential, that capture distinct criteria including scientific yield and cost effectiveness. These strategies will then be tested in real field settings (e.g. Turkey, Cyprus and Flanders) to assess how well simulation-trained decision policies transfer to practical archaeological contexts.

Plus d’informations :
[Website Vrije Universiteit Brussel]