Recently a new group of layered planar and quasi-planar metamaterials has emerged which promise unique electromagnetic properties. Layered metallic microstructures could play a special role in future technology, as they can be manufactured on a sub-optical wavelength scale and can be fabricated using established microelectronics technologies. We have begun a predominantly experimental study of planar and quasi-planar metallic microstructures, a new generation of metamaterials for optical applications. We will concentrate on various chiral planar metamaterials fabricated on the optical wavelength scale.
This project is concerned with the development and application of optimisation methods for machine learning algorithms. Many modern machine learning algorithms can be viewed as optimising bounds on the generalisation error derived in learning theory. Modern tools from mathematical programming such as second-order cone and semi-definite programs will be adapted to the optimisation problems arising in machine learning. The resulting methods will be tested on benchmark data and - whenever possible - on suitable real-world data sets.
Learning systems based on kernels are a powerful class of algorithms that includes Support Vector Machines and Gaussian Processes. These systems have become a major part of current research into and applications of adaptive systems. Despite this fact very little is known about when we can expect these systems to perform well. There has even been the assumption made that they provide a universal learning methodology. The proposed project will address this in order to:
This project invetsigates the application of adaptive numerical modelling techniques for the process optimisation of aluminium alloys.
This project investigates the problem of classifying and predicting fatigue crack initiation sites, through microstructure quantification in Austempered Ductile Iron. The aim of this work is to build data driven classifiers that provide enhanced understanding of a system through the ability to visualise input/output relationships, as well as providing good predictive performance for a set of imbalanced data.
This research programme is aimed at extending current Southampton research on adaptive multi-spectral imaging techniques in 3D (2 spatial, 1 spectral), to hyper-spectral data fusion which allows dynamic re-configuration for specific target/background conditions enabling optimal performance to be achieved in terms of target detection and tracking, whilst minimising the number of spectral windows.
In this research endeavour, we aim to develop flexible and robust methods for managing decentralised data fusion. We will be developing an agent-based control system for data fusion that:
We will be using market-based approaches to view management of data fusion activities from an economic point of view and investigate market design for structuring marketplace to achieve various properties such as Pareto optimality, fairness and stability. In order to maximise their individual utility in such markets, strategies will be designed for agents while keeping in mind the overall protocol of the marketplace. These strategies can be augmented through adaptive behaviour (for example though some form of Q-Learning) that aims to utilise knowledge gained from past interactions.
Through this research project, we aim to develop novel market-based control algorithms (together with a simple demonstrator) that evaluate the effectiveness of decentralised control using market-based techniques. This research will also provide a theoretical analysis of the marketplace design to determine its effectiveness, efficiency and predictability and a systematic evaluation of the system's operational performance.
LAVA is a 3 year EC funded Research and Technology Development project in the Information Society Technologies programme of the 5th Framework. Xerox Research Centre Europe is the co-ordinating partner in this project. The LAVA project began in May 2002. The main objective of the project is to devise machine learning technologies:
We are investigating the use of biometric data fusion to provide for secure identity verification. This is an extension of our existing research programmes in automatic gait recognition and lies within the Defence Technology Centre's theme research on data fusion.