| Abstract |
In order to facilitate information sharing in a coalition operation, unstructured data from various sensors in the environment must be analysed effectively. Current work within the Distributed Analytics and Information Sciences International Technology Alliance has developed the notion of generative policy models capable of learning over strings and text based data in tabular formats. However, in order to fully utilise generative pol- icy models in coalition environments, it is vital that unstructured data can be analysed alongside tabular data, potentially at the edge of the network. This paper performs a deep-dive into the emerging field of neural-symbolic machine learning with a view towards enabling future neural-symbolic generative policy models that are capable of analysing both structured and unstructured data present in a coalition operation, whilst providing full transparency and enabling edge of network reasoning capability. Specifically, a technique called DeepProbLog is investigated and applied to a coalition scenario, highlighting key research challenges and questions to be addressed in order to advance the capability of generative policy models for coalition operations. |