| Home | People | Publications | Projects | Journal Club |
People
Prof Alan Murray |
|||||
|
Neural Computation Algorithms, Architectures and Applications of Artificial Neural Networks - with particular interests in Spiking Computation and Probabilistic Computation.
| |||||
| Visiting Researcher (left) | Dr Evangelos Delivopoulos |
||||
|
COMPUTATIONAL INTELLIGENCE Learning from data: Neural Network Algorithms, Gaussian fitting, Classification, Probabilistic methods of learning
NEURON-SILICON INTERFACING
| |||||
Dr Zhijun Yang |
|||||
|
NEURAL ENGINEERING Biologically inspired algorithms and signal processing for single neuron and neuron populations, including, e.g., spiking neuron models, visual cortex signal processing and development of underlying learning mechanisms. Distributed and parallel computation in sensorimotor and somatomotor regime, including theory, algorithms, and applications. Complex system models, as neural networks and related models, including bioinformatic and neurophysiologic aspects, as ionic functionalities in synaptic signal transmission dynamics. Statistical time series analysis, pattern recognition and classification, and related neural network model development. Neuromorphic very large scale integrated (VLSI) circuit implementation of biological models.
| |||||
Dr Katherine Cameron |
|||||
|
NEUROMORPHIC VLSI Bio-inspired engineering solutions to analogue computation imperfections, mixed-signal VLSI design, and neural computation. In particular the ability of spike timing systems to adapt out the effects of transistor mismatch.
| |||||
Mr Andrew Cogman |
|||||
|
Probabilistic Neural Architectures and Deep Sub Micron Devices The characteristics of future deep sub micron MOSFET devices will render traditional device and circuit design techniques obsolete. Innovative circuit design and complex modeling in previous technology nodes has perpetuated the scaled silicon trend, the next generation of devices will have different problems which are unlikely to be solved using the same techniques. The need for a step change is apparent. A neural network design approach offers salvation from many of the device characteristics preventing continued scaling, such as device mismatch and high intrinsic noise levels. The use of simple stochastic processing units in a hierarchical structure results in an architecture which can be used in many signal processing and classification tasks. My research is centered on looking at using probabilistic architectures with deep sub micron devices to offer a new computing paradigm. Using the facilities offered through the NanoCMOS project, I hope to develop simulation models of these circuits and determine their potential for future computing systems.
| |||||
Dr Tong Boon Tang |
|||||
|
Probabilistic computation I am interested in probabilistic neural networks, especially those which are hardware-amenable. Applications: distributed sensor networks and biomedical devices.
| |||||
Mr Cheng-Kai Lu |
|||||
Mr Hugo A P Monteiro |
|||||
| Postgraduate (left) | Mr Vasin Boonsobhak |
||||
|
NEURAL ENGINEERING Vasin's current research aims at implementing in VLSI a depth-from-motion algorithm in a standard CMOS technology. This algorithm is based on an artificial retina with radially arranged neurons. During forward motion, a given object-trace will traverse exactly along one retinal radius, successively exciting its neurons. The temporal delay between two subsequently excited neurons directly encodes the distance to the stimulating object part. Stream of depth information can then be used for creating a three-dimensional map within a scene of interest. The focus of his work is on the implementation of a specialised silicon retina that provides locally processed depth information. Such a device could be integrated into many systems such as self-navigating robots and driver-assistant systems.
| |||||
| Postgraduate (left) | Miss Juan Huo |
||||
| Postgraduate (left) | Mr Alexandros Kourkoulas-Chondrorizos |
||||
|
Computational Neuroscience Investigating the effect of noise in spiking neural networks and particularly the dynamics of information in an attempt to clarify whether noise improves their information processing abilities.
| |||||
Dr Leena Patel |
|||||
|
BROAD INTERESTS Intelligent systems engineering with particular emphasis on biomimetic solutions, artificial intelligence (AI), evolutionary computing, cognitive science - particularly visual perception, image processing and object recognition, developing "usable" technology, robotics and renewable energy. More recently, Leena;s interests have focussed on intelligent solutions for eco-initiative projects. This includes bio-inspired control of renewable energy devices and simulation and optimisation of carbon capture processes.
BIO-INSPIRED CPG CONTROL Recent preparatory work involved optimising the lamprey CPG responsible for swimming control using Genetic Algorithm (GA) techniques. Improved controllers demonstrate greater operative ranges of locomotive speed, oscillation frequency and phase lags between segments of the fish's body, whilst being simpler in design. Improved efficiency and optimisation of a simple Wave Energy Converter (WEC) coupled to a lamprey CPG has also been successful.
BRAIN & BEHAVIOUR
| |||||
| Home | People | Publications | Projects | Journal Club |