A robust Bayesian two-sample test for detecting intervals of differential gene expression in microarray time series.

Understanding the regulatory mechanisms that are responsible for an organism's response to environmental change is an important issue in molecular biology. A first and important step towards this goal is to detect genes whose expression levels are affected by altered external conditions. A range of methods to test for differential gene expression, both in static as well as in time-course experiments, have been proposed. While these tests answer the question whether a gene is differentially expressed, they do not explicitly address the question when a gene is differentially expressed, although this information may provide insights into the course and causal structure of regulatory programs. In this article, we propose a two-sample test for identifying intervals of differential gene expression in microarray time series. Our approach is based on Gaussian process regression, can deal with arbitrary numbers of replicates, and is robust with respect to outliers. We apply our algorithm to study the response of Arabidopsis thaliana genes to an infection by a fungal pathogen using a microarray time series dataset covering 30,336 gene probes at 24 observed time points. In classification experiments, our test compares favorably with existing methods and provides additional insights into time-dependent differential expression.

Learning shape priors for single view reconstruction

In this paper, we aim to reconstruct free-from 3D models from a single view by learning the prior knowledge of a specific class of objects. Instead of heuristically proposing specific regularities and defining parametric models as previous research, our shape prior is learned directly from existing 3D models under a framework based on the Gaussian Process Latent Variable Model (GPLVM). The major contributions of the paper include: 1) a probabilistic framework for prior-based reconstruction we propose, which requires no heuristic of the object, and can be easily generalized to handle various categories of 3D objects, and 2) an attempt at automatic reconstruction of more complex 3D shapes, like human bodies, from 2D silhouettes only. Qualitative and quantitative experimental results on both synthetic and real data demonstrate the efficacy of our new approach. ©2009 IEEE.

MicroRNAs and other small RNAs enriched in the Arabidopsis RNA-dependent RNA polymerase-2 mutant.

The Arabidopsis genome contains a highly complex and abundant population of small RNAs, and many of the endogenous siRNAs are dependent on RNA-Dependent RNA Polymerase 2 (RDR2) for their biogenesis. By analyzing an rdr2 loss-of-function mutant using two different parallel sequencing technologies, MPSS and 454, we characterized the complement of miRNAs expressed in Arabidopsis inflorescence to considerable depth. Nearly all known miRNAs were enriched in this mutant and we identified 13 new miRNAs, all of which were relatively low abundance and constitute new families. Trans-acting siRNAs (ta-siRNAs) were even more highly enriched. Computational and gel blot analyses suggested that the minimal number of miRNAs in Arabidopsis is approximately 155. The size profile of small RNAs in rdr2 reflected enrichment of 21-nt miRNAs and other classes of siRNAs like ta-siRNAs, and a significant reduction in 24-nt heterochromatic siRNAs. Other classes of small RNAs were found to be RDR2-independent, particularly those derived from long inverted repeats and a subset of tandem repeats. The small RNA populations in other Arabidopsis small RNA biogenesis mutants were also examined; a dcl2/3/4 triple mutant showed a similar pattern to rdr2, whereas dcl1-7 and rdr6 showed reductions in miRNAs and ta-siRNAs consistent with their activities in the biogenesis of these types of small RNAs. Deep sequencing of mutants provides a genetic approach for the dissection and characterization of diverse small RNA populations and the identification of low abundance miRNAs.

Bayesian Active Learning for Classification and Preference Learning

Information theoretic active learning has been widely studied for probabilistic models. For simple regression an optimal myopic policy is easily tractable. However, for other tasks and with more complex models, such as classification with nonparametric models, the optimal solution is harder to compute. Current approaches make approximations to achieve tractability. We propose an approach that expresses information gain in terms of predictive entropies, and apply this method to the Gaussian Process Classifier (GPC). Our approach makes minimal approximations to the full information theoretic objective. Our experimental performance compares favourably to many popular active learning algorithms, and has equal or lower computational complexity. We compare well to decision theoretic approaches also, which are privy to more information and require much more computational time. Secondly, by developing further a reformulation of binary preference learning to a classification problem, we extend our algorithm to Gaussian Process preference learning.

Dynamic analysis of the brushless doubly-fed induction generator during symmetrical three-phase voltage dips

The Brushless Doubly-Fed Induction Generator (BDFIG) shows commercial promise as replacement for doublyfed slip-ring generators for wind power applications by offering reduced capital and operational costs due to its brushless operation. In order to facilitate its commercial deployment, the capabilities of the BDFIG system to comply with grid code requirements have to be assessed. This paper, for the first time, studies the performance of the BDFIG under grid fault ride-through and presents the dynamic behaviour of the machine during three-phase symmetrical voltage dips. Both full and partial voltage dips are studied using a vector model. Simulation and experimental results are provided for a 180 frame BDFIG.

A novel vector control approach for Single Phase Brushless Doubly Fed Machine

This paper proposed a novel control scheme for operating the Single Phase Brushless Doubly-Fed Machine (SPB) based on Stator-Flux-Oriented control algorithm. The SPB is a new type of Brushless Doubly-Fed Machine (BDFM) which shows a potential in applications which require adjustable speed such as Wind Power generation and speed adjustable Drive. The SPB can be applied to single-phase power system and the lower cost of the SPB makes the SPB suitable for low-rated power conversion applications. This paper develops the control scheme of the SPB with explicit mathematical analysis and block diagram of the controller. Experimental verification is also given. © 2011 IEEE.

On-line policy optimisation of spoken dialogue systems via live interaction with human subjects

Statistical dialogue models have required a large number of dialogues to optimise the dialogue policy, relying on the use of a simulated user. This results in a mismatch between training and live conditions, and significant development costs for the simulator thereby mitigating many of the claimed benefits of such models. Recent work on Gaussian process reinforcement learning, has shown that learning can be substantially accelerated. This paper reports on an experiment to learn a policy for a real-world task directly from human interaction using rewards provided by users. It shows that a usable policy can be learnt in just a few hundred dialogues without needing a user simulator and, using a learning strategy that reduces the risk of taking bad actions. The paper also investigates adaptation behaviour when the system continues learning for several thousand dialogues and highlights the need for robustness to noisy rewards. © 2011 IEEE.

Symmetrical low voltage ride-through of the Brushless Doubly-Fed Induction Generator

The Brushless Doubly-Fed Induction Generator (BDFIG) shows commercial promise for wind power generation due to its lower cost and higher reliability compared to the Doubly-Fed Induction Generator (DFIG). For the purposes of commercialisation, the BDFIG must meet grid codes at all times. Nowadays, all new wind generators have to ride through certain grid faults, and the Low-Voltage Ride Through (LVRT) capability has become one of the most important points on which to assess the performance a generator. This paper, for the first time, proposes a control scheme to enable the the BDFIG to ride through symmetrical voltage dips. Simulation results and experimental results on a prototype BDFIG show that the proposed scheme gives the capability to ride through low voltage faults. © 2011 IEEE.

Symmetrical low voltage ride-through of a 250 kW Brushless DFIG

The Brushless Doubly-Fed Induction Generator (Brushless DFIG) shows commercial promise for wind power generation due to its lower cost and higher reliability when compared with the conventional Doubly-Fed Induction Generator (DFIG). In the most recent grid codes, wind generators are required to be able to ride through a low voltage fault and meet the reactive current demand from the grid. Hence, a Low-Voltage Ride-Through (LVRT) capability is important for wind generators which are integrated into the grid. In this paper the authors propose a control strategy enabling the Brushless DFIG to successfully ride through a symmetrical voltage dip. The control strategy has been implemented on a 250 kW Brushless DFIG and the experimental results indicate that LVRT is possible without a crowbar.

Random function priors for exchangeable arrays with applications to graphs and relational data

A fundamental problem in the analysis of structured relational data like graphs, networks, databases, and matrices is to extract a summary of the common structure underlying relations between individual entities. Relational data are typically encoded in the form of arrays; invariance to the ordering of rows and columns corresponds to exchangeable arrays. Results in probability theory due to Aldous, Hoover and Kallenberg show that exchangeable arrays can be represented in terms of a random measurable function which constitutes the natural model parameter in a Bayesian model. We obtain a flexible yet simple Bayesian nonparametric model by placing a Gaussian process prior on the parameter function. Efficient inference utilises elliptical slice sampling combined with a random sparse approximation to the Gaussian process. We demonstrate applications of the model to network data and clarify its relation to models in the literature, several of which emerge as special cases.

Kernel conditional quantile estimation via reduction revisited

Quantile regression refers to the process of estimating the quantiles of a conditional distribution and has many important applications within econometrics and data mining, among other domains. In this paper, we show how to estimate these conditional quantile functions within a Bayes risk minimization framework using a Gaussian process prior. The resulting non-parametric probabilistic model is easy to implement and allows non-crossing quantile functions to be enforced. Moreover, it can directly be used in combination with tools and extensions of standard Gaussian Processes such as principled hyperparameter estimation, sparsification, and quantile regression with input-dependent noise rates. No existing approach enjoys all of these desirable properties. Experiments on benchmark datasets show that our method is competitive with state-of-the-art approaches. © 2009 IEEE.

Lateral spreading-induced abutment rotation in the 2011 Christchurch earthquake: Observations and analysis

A series of strong earthquakes near Christchurch, New Zealand, occurred between September 2010 and December 2011, causing widespread liquefaction throughout the city's suburbs. Lateral spreading developed along the city's Avon River, damaging many of the bridges east of the city centre. The short-to medium-span bridges exhibited a similar pattern of deformation, involving back-rotation of their abutments and compression of their decks. By explicitly considering the rotational equilibrium of the abutments about their point of contact with the rigid bridge decks, it is shown that relatively small kinematic demands from the laterally spreading backfill soil are needed to initiate pile yielding, and that this mode of deformation should be taken into account in the design of the abutments and abutment piles.

Avoiding pathologies in very deep networks

Choosing appropriate architectures and regularization strategies of deep networks is crucial to good predictive performance. To shed light on this problem, we analyze the analogous problem of constructing useful priors on compositions of functions. Specifically, we study the deep Gaussian process, a type of infinitely-wide, deep neural network. We show that in standard architectures, the representational capacity of the network tends to capture fewer degrees of freedom as the number of layers increases, retaining only a single degree of freedom in the limit. We propose an alternate network architecture which does not suffer from this pathology. We also examine deep covariance functions, obtained by composing infinitely many feature transforms. Lastly, we characterize the class of models obtained by performing dropout on Gaussian processes.