|
My research interests are in statistical modeling for speech and
language processing. The goal of this work is to develop
modeling techniques needed for human-to-machine communication
and for the processing of speech and written text by autonomous
systems. Specific areas of interest are novel search and
estimation algorithms for speech recognition, all aspects of
statistical machine translation, and applications to problems
such as spoken document retrieval and speech translation.
|
|
Current interests include 3D shape recovery from uncalibrated images;
machine learning for object detection, segmentation and
recognition; computer vision for novel man-machine interfaces
and the visual guidance of robots.
|
|
Research interests include applications of real-time computer vision
in two areas:
i) Human Computer Interfaces (HCI) - with a particular focus on
Augmented Reality (AR)
ii) Visually Guided Robotics - including flexible automation for
manufacturing and control of autonomous Miniature Aerial Vehicles
(MAVs).
|
|
Current research interests include Bayesian
approaches to machine learning, artificial intelligence,
statistics, information retrieval, bioinformatics, and
computational motor control. Statistics provides the mathematical
foundations for handling uncertainty, making decisions, and
designing learning systems. I have recently worked on Gaussian
processes, non-parametric Bayesian methods, clustering,
approximate inference algorithms, graphical models, Monte Carlo
methods, and semi-supervised learning.
|
|
Current research interests are in the area of communication theory, information theory, coding theory, digital modulation, multiple-input multiple-output detection and signal processing techniques for communications.
|
|
Current research interests relate to multi-scale image and video processing methods and their application to tasks such as object recognition, image classification, object tracking and scene interpretation. In particular we are developing techniques based on dual-tree complex wavelets, which provides a multi-scale directional filter bank that closely imitates the behaviour of the human V1 visual cortex and is highly efficient to compute. We are studying what processes need to follow this front-end in order to perform reliable feature-point detection and description, and hence good robust methods of object recognition in the presence of all common sources of variation of views and lighting of an object. Various new local phase-based signal processing methods are being developed to achieve this within acceptable levels of computational complexity.
|
|
The brain has a remarkable capacity to learn continuously about
the environment and to use this knowledge flexibly to make
predictions and guide its future decisions. I study learning and
memory from computational, algorithmic/representational and
neurobiological viewpoints. Computationally and algorithmically,
I use ideas from Bayesian approaches to statistical inference to
characterize the goals and mechanisms of learning in terms of
normative principles and behavioral results. I also perform
dynamical systems analyses of reduced biophysical models to
understand the mapping of these mechanisms into cellular and
network models. I collaborate very closely with experimental
neuroscience groups, doing in vitro intracellular recordings,
multi-unit recordings in behaving animals, in both cases
associated with the hippocampus, and human psychophysical
experiments on forms of structure learning in perception.
|
|
Current research interests [relevant to cognitive systems] include
manufacturing & service applications of software agents and distributed
decision making, RFID system, Bayesian approaches to valueing industrial
information, probablistic methods for locating and tracking industrial
objects, distributed industrial information architectures.
|
|
My research interests primarily lie in understanding how
information associated with products can/should be managed and used
throughout their lifecycle so as to improve the effectiveness with which
decisions associated with them are made. In particular, I am investigating
the use of Bayesian Networks as a means to model product lifecycle decisions
and to quantify the value of information. The value of information problem
is of particular interest since it dictates the feasibility of using
emerging technologies such as RFID in product management.
|
|
I work broadly on probabilistic inference in machine learning, including
supervised, unsupervised and reinforcement learning. I'm particularly
interested in solving computationally challenging problems of representing and
manipulating uncertain knowledge in complex models using approximation
techniques including MCMC. Recently, I have worked extensively with Gaussian process models, see also my
recent book Gaussian Processes
for Machine Learning (with Chris Williams).
|
|
Current research interests include the development of user interfaces
for augmented reality , wearable computing, ubiquitous computing
environments and the integration of these. Current directions of work
are computer vision techniques for localisation and tracking of mobile
augmented reality devices and for interaction methods in intelligent
room environments.
|
|
We use theoretical, computational and experimental studies to investigate
the computational principles underlying the brains control of movement. We
have developed a research programme that uses computational techniques from
machine learning, control theory and signal processing together with novel
experimental techniques that include robotic interfaces and virtual reality
systems that allow for precise experimental control over sensory inputs and
task variables.
|
|
Interested in man-machine interfaces and machine learning; with
applications in speech understanding and dialogue management
(http://mi.eng.cam.ac.uk/research/dialogue/). Also interested in
user simulation, emotion detection and generation, speech
synthesis, and voice conversion.
|