Policy gradient methods have been widely applied in reinforcement learning. For reasons of safety and cost, learning is often conducted using a simulator. However, learning in simulation does not traditionally utilise the opportunity to improve learning by adjusting certain environment variables – state features that are randomly determined by the environment in a physical setting but controlla...

Bayesian optimisation has been successfully applied to a variety of reinforcement learning problems. However, the traditional approach for learning optimal policies in simulators does not utilise the opportunity to improve learning by adjusting certain environment variables – state features that are randomly determined by the environment in a physical setting but are controllable in a simulator...

We consider the estimation of the extremes of the metocean climate, in particular those of the univariate and joint distributions of wave height, wave period and wind speed. This is of importance in the design of oil rigs and other marine structures which must be able to withstand extreme environmental loadings. Such loadings are often functions of two or more metocean variables and the problem...

Usage of activity theory has emerged in other applications in computer science, including human computer interfaces. This paper applies activity theory as a theoretical framework for the analysis of DSS for extreme events. An example, based on computer support on the web for hurricane response, is used to illustrate how activity theory might be applied to extreme events in a decision support sy...

Bayesian optimisation has been successfully applied to a variety of reinforcement learning problems. However, the traditional approach for learning optimal policies in simulators does not utilise the opportunity to improve learning by adjusting certain environment variables: state features that are unobservable and randomly determined by the environment in a physical setting but are controllabl...

Nowadays, many real problems can be formalized as Distributed CSPs. A distributed constraint satisfaction problem (DisCSP) is a CSP in which variables and constraints are distributed among multiple automated agents. Many researchers assume for simplicity that each agent has exactly one variable. For real distributed problem these techniques require a large amount of messages passed among agents...

Per tone equalization was recently proposed as an alternative equalization structure for multicarrier systems. In this paper we propose two NLMS based variable length per tone equalization algorithms. In the first algorithm, the number of equalizer taps used in every subchannel is determined according to channel conditions which results in a reduction in equalization complexity. While the secon...

(k, s)-SAT is the propositional satisfiability problem restricted to instances where each clause has exactly k distinct literals and every variable occurs at most s times. It is known that there exists an exponential function f such that for s ≤ f(k) all (k, s)-SAT instances are satisfiable, but (k, f(k)+1)-SAT is already NP-complete (k ≥ 3). Exact values of f are only known for k = 3 and k = 4...

Assessing entity (e.g., person) risk from entity-related events requires appropriate techniques to address the relevance of events (individually and/or in aggregate) relative to a prevailing temporal frame of reference—for continuous risk monitoring, a running time point representing “the present.” We describe two classes of temporal relevance techniques we have used towards insider threat dete...