On sparsity and overcompleteness in image models

Computational models of visual cortex, and in particular those based on sparse coding, have enjoyed much recent attention. Despite this currency, the question of how sparse or how over-complete a sparse representation should be, has gone without principled answer. Here, we use Bayesian model-selection methods to address these questions for a sparse-coding model based on a Student-t prior. Having validated our methods on toy data, we find that natural images are indeed best modelled by extremely sparse distributions; although for the Student-t prior, the associated optimal basis size is only modestly over-complete.

Probabilistic Amplitude Demodulation

Auditory scene analysis is extremely challenging. One approach, perhaps that adopted by the brain, is to shape useful representations of sounds on prior knowledge about their statistical structure. For example, sounds with harmonic sections are common and so time-frequency representations are efficient. Most current representations concentrate on the shorter components. Here, we propose representations for structures on longer time-scales, like the phonemes and sentences of speech. We decompose a sound into a product of processes, each with its own characteristic time-scale. This demodulation cascade relates to classical amplitude demodulation, but traditional algorithms fail to realise the representation fully. A new approach, probabilistic amplitude demodulation, is shown to out-perform the established methods, and to easily extend to representation of a full demodulation cascade.

Two problems with variational expectation maximisation for time-series models

Variational methods are a key component of the approximate inference and learning toolbox. These methods fill an important middle ground, retaining distributional information about uncertainty in latent variables, unlike maximum a posteriori methods (MAP), and yet generally requiring less computational time than Monte Carlo Markov Chain methods. In particular the variational Expectation Maximisation (vEM) and variational Bayes algorithms, both involving variational optimisation of a free-energy, are widely used in time-series modelling. Here, we investigate the success of vEM in simple probabilistic time-series models. First we consider the inference step of vEM, and show that a consequence of the well-known compactness property of variational inference is a failure to propagate uncertainty in time, thus limiting the usefulness of the retained distributional information. In particular, the uncertainty may appear to be smallest precisely when the approximation is poorest. Second, we consider parameter learning and analytically reveal systematic biases in the parameters found by vEM. Surprisingly, simpler variational approximations (such a mean-field) can lead to less bias than more complicated structured approximations.

Vowel normalisation: Time-domain processing of the internal dynamics of speech

Human listeners can identify vowels regardless of speaker size, although the sound waves for an adult and a child speaking the ’same’ vowel would differ enormously. The differences are mainly due to the differences in vocal tract length (VTL) and glottal pulse rate (GPR) which are both related to body size. Automatic speech recognition machines are notoriously bad at understanding children if they have been trained on the speech of an adult. In this paper, we propose that the auditory system adapts its analysis of speech sounds, dynamically and automatically to the GPR and VTL of the speaker on a syllable-to-syllable basis. We illustrate how this rapid adaptation might be performed with the aid of a computational version of the auditory image model, and we propose that an auditory preprocessor of this form would improve the robustness of speech recognisers.

Making the case for inclusive design.

This paper describes the University of Cambridge, Engineering Design Centre's (EDC) case for inclusive design, based on 10 years of research, promotion and knowledge transfer. In summary, inclusive design applies an understanding of customer diversity to inform decisions throughout the development process, in order to better satisfy the needs of more people. Products that are more inclusive can reach a wider market, improve customer satisfaction and drive business success. The rapidly ageing population increases the importance of this approach. The case presented here has helped to convince BT, Nestlé and others to adopt an inclusive approach.

Age, technology prior experience and ease of use: Who's doing what?

Designers often assume that their users will have some digital technological prior experience. We examined these levels of prior experience by surveying frequency and ease of technology use with a range of technology products. 362 people participated as part of a UK nationwide larger survey of people's capabilities and characteristics to inform product design. We found that frequency and self-reported ease of use are indeed correlated for all of the products. Furthermore, both frequency and ease of use declined significantly with age for most of the products. In fact, 29% of the over 65s had never or rarely used any of the products, except for digital TV. We conclude that interfaces need to be designed carefully to avoid implicit assumptions about users' previous technology use.

Rebuilding the limit order book: Sequential Bayesian inference on hidden states

The limit order book of an exchange represents an information store of market participants' future aims and for many traders the information held in this store is of interest. However, information loss occurs between orders being entered into the exchange and limit order book data being sent out. We present an online algorithm which carries out Bayesian inference to replace information lost at the level of the exchange server and apply our proof of concept algorithm to real historical data from some of the world's most liquid futures contracts as traded on CME GLOBEX, EUREX and NYSE Liffe exchanges. © 2013 © 2013 Taylor & Francis.

The effect of carbon nanotube orientation on erosive wear resistance of CNT-epoxy based composites

Aligned carbon nanotube (CNT) polymer composites are envisioned as the next-generation composite materials for a wide range of applications. In this work, we investigate the erosive wear behavior of epoxy matrix composites reinforced with both randomly dispersed and aligned carbon nanotube (CNT) arrays. The aligned CNT composites are prepared in two different configurations, where the sidewalls and ends of nanotubes are exposed to the composite surface. Results have shown that the composite with vertically aligned CNT-arrays exhibits superior erosive wear resistance compared to any of the other types of composites, and the erosion rate reaches a similar performance level to that of carbon steel at 20° impingement angle. The erosive wear mechanism of this type of composite, at various impingement angles, is studied by Scanning Electron Microscopy (SEM). We report that the erosive wear performance shows strong dependence on the alignment geometries of CNTs within the epoxy matrix under identical nanotube loading fractions. Correlations between the eroded surface roughness and the erosion rates of the CNT composites are studied by surface profilometry. This work demonstrates methods to fabricate CNT based polymer composites with high loading fractions of the filler, alignment control of nanotubes and optimized erosive wear properties. © 2014 Elsevier Ltd. All rights reserved.

First demonstration of a full C-band tunable WDM-PON system with novel high-temperature DS-DBR lasers

We demonstrate automatic operation of a cooler-less tunable-laser based WDM-PON system. Using a pilot-tone based overhead channel and centralized wavelength locking scheme, 1 Gb/s and 10 Gb/s data transmission is demonstrated in a multi-user set-up. © 2013 Optical Society of America.