We want to understand how much information can be transferred in secret for the resource 
constrained, short packet, low latency, inference applications most relevant to IoT.  Then we 
wanted to develop approaches that approach these rates.  This involves research in information 
theory, communications, signal processing, and inference.  We were also interesed in understanding 
the possible attacks on IoT inference systems, their impact and how to protect against them. We 
hoped to obtain fundamental theory, limits and protection approaches in the distributed processing 
environments typical in IoT.  Many important attacks on IoT and CPS systems, like electrical grids, 
vehicle networks and gas networks, can be described as changing the topology of a graph.  Due to 
this, we also hoped to perform fundamental research on detecting topology changes in graphs.  
Another, important topic is attacks on machine learning and reinforcement learning and we hoped to 
make contributions on that topic.   To augment research on attack detection and protection, it is also 
important to investigate authentication, especially at the physical layer where much remains uninvestigated.