Bayesian and L1 Approaches to Sparse Unsupervised Learning

The use of L1 regularisation for sparse learning has generated immense research interest, with successful application in such diverse areas as signal acquisition, image coding, genomics and collaborative filtering. While existing work highlights the many advantages of L1 methods, in this paper we find that L1 regularisation often dramatically underperforms in terms of predictive performance when compared with other methods for inferring sparsity. We focus on unsupervised latent variable models, and develop L1 minimising factor models, Bayesian variants of "L1", and Bayesian models with a stronger L0-like sparsity induced through spike-and-slab distributions. These spike-and-slab Bayesian factor models encourage sparsity while accounting for uncertainty in a principled manner and avoiding unnecessary shrinkage of non-zero values. We demonstrate on a number of data sets that in practice spike-and-slab Bayesian methods outperform L1 minimisation, even on a computational budget. We thus highlight the need to re-assess the wide use of L1 methods in sparsity-reliant applications, particularly when we care about generalising to previously unseen data, and provide an alternative that, over many varying conditions, provides improved generalisation performance.

Exploiting variable-width features in large vocabulary speech recognition

The use of variable-width features (prosodics, broad structural information etc.) in large vocabulary speech recognition systems is discussed. Although the value of this sort of information has been recognized in the past, previous approaches have not been widely used in speech systems because either they have not been robust enough for realistic, large vocabulary tasks or they have been limited to certain recognizer architectures. A framework for the use of variable-width features is presented which employs the N-Best algorithm with the features being applied in a post-processing phase. The framework is flexible and widely applicable, giving greater scope for exploitation of the features than previous approaches. Large vocabulary speech recognition experiments using TIMIT show that the application of variable-width features has potential benefits.

Gaussian mixture modeling with Gaussian process latent variable models

Density modeling is notoriously difficult for high dimensional data. One approach to the problem is to search for a lower dimensional manifold which captures the main characteristics of the data. Recently, the Gaussian Process Latent Variable Model (GPLVM) has successfully been used to find low dimensional manifolds in a variety of complex data. The GPLVM consists of a set of points in a low dimensional latent space, and a stochastic map to the observed space. We show how it can be interpreted as a density model in the observed space. However, the GPLVM is not trained as a density model and therefore yields bad density estimates. We propose a new training strategy and obtain improved generalisation performance and better density estimates in comparative evaluations on several benchmark data sets. © 2010 Springer-Verlag.

An evaluation of volumetric interest points

This paper presents the first performance evaluation of interest points on scalar volumetric data. Such data encodes 3D shape, a fundamental property of objects. The use of another such property, texture (i.e. 2D surface colouration), or appearance, for object detection, recognition and registration has been well studied; 3D shape less so. However, the increasing prevalence of depth sensors and the diminishing returns to be had from appearance alone have seen a surge in shape-based methods. In this work we investigate the performance of several detectors of interest points in volumetric data, in terms of repeatability, number and nature of interest points. Such methods form the first step in many shape-based applications. Our detailed comparison, with both quantitative and qualitative measures on synthetic and real 3D data, both point-based and volumetric, aids readers in selecting a method suitable for their application. © 2011 IEEE.

Variable repetition frequency femtosecond-pulse surface emitting semiconductor laser

We report a femtosecond-pulse vertical-external-cavity surface-emitting laser with a continuous repetition frequency tuning range of 8 near 1 GHz. A constant average output power of 56 ± 1 mW and near-transform-limited pulse duration of 450 ± 20 fs were observed across the entire tuning range. © 2011 American Institute of Physics.

QM/MM simulation of liquid water with an adaptive quantum region.

The simulation of complex chemical systems often requires a multi-level description, in which a region of special interest is treated using a computationally expensive quantum mechanical (QM) model while its environment is described by a faster, simpler molecular mechanical (MM) model. Furthermore, studying dynamic effects in solvated systems or bio-molecules requires a variable definition of the two regions, so that atoms or molecules can be dynamically re-assigned between the QM and MM descriptions during the course of the simulation. Such reassignments pose a problem for traditional QM/MM schemes by exacerbating the errors that stem from switching the model at the boundary. Here we show that stable, long adaptive simulations can be carried out using density functional theory with the BLYP exchange-correlation functional for the QM model and a flexible TIP3P force field for the MM model without requiring adjustments of either. Using a primary benchmark system of pure water, we investigate the convergence of the liquid structure with the size of the QM region, and demonstrate that by using a sufficiently large QM region (with radius 6 Å) it is possible to obtain radial and angular distributions that, in the QM region, match the results of fully quantum mechanical calculations with periodic boundary conditions, and, after a smooth transition, also agree with fully MM calculations in the MM region. The key ingredient is the accurate evaluation of forces in the QM subsystem which we achieve by including an extended buffer region in the QM calculations. We also show that our buffered-force QM/MM scheme is transferable by simulating the solvated Cl(-) ion.

Robust variable horizon MPC with move blocking

This paper introduces a new formulation of variable horizon model predictive control (VH-MPC) that utilises move blocking for reducing computational complexity. Various results pertaining to move blocking are derived, following which, a generalised blocked VH-MPC controller is formulated for linear discrete-time systems. Robustness to bounded disturbances is ensured through the use of tightened constraints. The resulting time-varying control scheme is shown to guarantee robust recursive feasibility and finite-time completion. An example is then presented for a particular choice of blocking regime, as would be applicable to vehicle manœuvring problems. Simulations demonstrate the efficacy of the formulation. © 2012 Elsevier B.V. All rights reserved.

The development of a hot rolling process for variable cross-section I-beams

To meet targeted reductions in CO 2 emissions by 2050, demand for metal must be cut, for example through the use of lightweight technologies. However, the efficient production of weight optimized components often requires new, more flexible forming processes. In this paper, a novel hot rolling process is presented for forming I-beams with variable cross-section, which are lighter than prismatic alternatives. First, the new process concept is presented and described. A detailed computational and experimental analysis is then conducted into the capabilities of the process. Results show that the process is capable of producing defect free I-beams with variations in web depth of 30-50%. A full analysis of the process then indicates the likely failure modes, and identifies a safe operating window. Finally, the implications of these results for producing lightweight beams are discussed. © 2012 Elsevier B.V. All rights reserved.

Robust pairwise matching of interest points with complex wavelets

We present a matching framework to find robust correspondences between image features by considering the spatial information between them. To achieve this, we define spatial constraints on the relative orientation and change in scale between pairs of features. A pairwise similarity score, which measures the similarity of features based on these spatial constraints, is considered. The pairwise similarity scores for all pairs of candidate correspondences are then accumulated in a 2-D similarity space. Robust correspondences can be found by searching for clusters in the similarity space, since actual correspondences are expected to form clusters that satisfy similar spatial constraints in this space. As it is difficult to achieve reliable and consistent estimates of scale and orientation, an additional contribution is that these parameters do not need to be determined at the interest point detection stage, which differs from conventional methods. Polar matching of dual-tree complex wavelet transform features is used, since it fits naturally into the framework with the defined spatial constraints. Our tests show that the proposed framework is capable of producing robust correspondences with higher correspondence ratios and reasonable computational efficiency, compared to other well-known algorithms. © 1992-2012 IEEE.