Istituto di Scienza e Tecnologie dell'Informazione     
Kuruoglu E. E., Lasenby J., Cemgil A. T., Kokaram A. C., Morris R. D., (eds.) .. Special Issue in Honour of William J. (Bill) Fitzgerald. In: Digital Signal Processing: a review journal. Editorial, vol. 47 (12) pp. 1 - 2. Elsevier, 2015.
We all know that at times there can be too narrow a bandwidth for communication between different research fields. This lack of communication can be a bottleneck restricting the pace of progress. Anecdotes abound of discoveries and re-discoveries in many disciplines, where appropriate tools for solving seemingly hard problems are found to have been developed elsewhere, sometimes decades earlier, but escaped the attention or comprehension of individuals who are just deeply focused in their own pocket of research. Good recent examples include the resurgence of interest in Neural Networks, and the important impact of engineering disciplines on the Biomedical Sciences. Fortunately, there are multilingual scientists who can see across fields, like storytellers who spread the word from town to town. Their interest is broader; their priority is not to focus on one particular problem and solve it, but to build up a global picture of the world and the place of humans in it. They move from field to field, from problem to problem with the will to understand the underlying principles stripped away from specifics and irrelevant detail. Despite different terminology, they can help us to create a common language and point us to theories which have been gathering dust for centuries. It is often those pioneers that provide inspiration and novel ideas without even realising it. William Fitzgerald, or Bill as we all knew him, was such a person. A unique scientist whose knowledge spanned the fields of statistical physics, statistics, probability theory, signal processing and information theory. He was a statistical physicist by training and contributed a number of publications to the field while at Birmingham, Zurich and Grenoble. Working in industry, Bill's interests shifted to signal processing theory and applications. This continued during his time at Cambridge, where one of his main contributions was to draw the Signal Processing community's attention to the field of Bayesian Statistics. Together with a small number of other pioneers, Bill outlined how statistical signal processing problems should include prior information as well as illustrating the technical means of calculating non-tractable estimates in the Bayesian framework. The numerical method were based on sampling, an idea he borrowed from statistical physics. Many of Bill's students have become experts in Bayesian techniques and Bill's influence has extended to industrial companies, start-ups he was part of and workshops he organised: all were inspired by his Bayesian approach to Signal Processing. Bill provided vision, inspiration and the persistence to explore Bayesian methods in many fields, including beamforming, image segmentation, computational biophysics, online data analytics and many more. His book, co-authored with Joseph Ruanaidh, has influenced researchers and students all around the world. We, five of his ex-students and colleagues, whose careers have been heavily influenced by Bill, believed that a fitting way to bid farewell to Bill was to have a special issue in his honour. This issue has contributors whose work has been influenced by Bill personally, as well as those influenced by his ideas. We hope that it provides a rich spectrum of contemporary research on Bayesian methods: some are tutorial in nature and some discuss current research. We hope it will have a lasting impact and will go some way to introducing new Bayesian ideas.
URL: http://www.sciencedirect.com/science/article/pii/S1051200415002675?np=y
DOI: 10.1016/j.dsp.2015.09.001
Subject Bayesian signal processing
Bayes theory
G.3 PROBABILITY AND STATISTICS. Probabilistic algorithms (including Monte Carlo)
62F15 Bayesian inference
62C10 Bayesian problems; characterization of Bayes procedures

Icona documento 1) Download Document PDF

Icona documento Open access Icona documento Restricted Icona documento Private


Per ulteriori informazioni, contattare: Librarian http://puma.isti.cnr.it

Valid HTML 4.0 Transitional