In a coalition network, data is collected from sensing systems such as drones, satellites or robotic ground vehicles, and then processed by various computational techniques including signal processing and machine learning. The processed data is then delivered to human decision-makers in forms they can understand, such as natural language or pictorial visualisations. The DAIS ITA programme takes a unique perspective on the coalition network, viewing it as a “distributed brain”: re-imagining the traditional data analytic pipeline – from sensors to analytics to human – in terms of a neural network architecture used to simulate human cognition. This makes the system highly resilient and adaptable to changes in the network environment that often happen in rapidly-evolving emergency situations.
Principal Investigator at Cardiff: Ian Taylor
Collaborators: Cardiff University, Purdue University, IBM, US Army Research Laboratory (ARL), UK Defence Science and Technology Laboratory (Dstl)
Anticipatory Situational Understanding
Advances in sensing technologies and data analytics – especially machine learning – mean that computer systems are increasingly able to help people maintain awareness of an evolving situation and predict possible future events. However, current machine learning systems have two major limitations: (1) they are typically unable to explain their predictions in ways that are meaningful to humans; and (2) they need a huge amount of data to learn from, and many important kinds of event are relatively rare, meaning that computers need to be “told” rather than “taught”. This project is the first to study these problems – tellability and explainability – in the context of coalition operations, leading to novel ways of linking human-machine communication, reasoning, and learning.
Principal Investigator at Cardiff: Alun Preece
Collaborators: Cardiff University, University College London, University of California at Los Angeles, Airbus, BAE Systems, IBM, US Army Research Laboratory (ARL), UK Defence Science and Technology Laboratory (Dstl)
Evolution of Human Systems
Successful multi-agency operations often involve understanding complex situations involving multiple groups with varying aims and capabilities. DAIS ITA is undertaking basic research into mutability of group-based human systems, and it can be modelled. Early work has identified five critical features of social groups that govern mutability, relating to social norms, individual standing, status rivalry, ingroup bias and cooperation. All these features lend themselves to computer modelling. The research also considers the important role of online social media in relation to understanding the mutability of groups. This can play an active role in supporting collective behaviour.
Principal Investigator at Cardiff: Roger Whitaker
Collaborators: Cardiff University, University of Massachusetts at Amherst, Pennsylvania State University, University of Southampton, Yale, BAE Systems, IBM, US Army Research Laboratory (ARL), UK Defence Science and Technology Laboratory (Dstl)