Analytic model for turn off in the silicon-on-insulator LIGBT

Lateral insulated gate bipolar transistors (LIGBTs) in silicon-on-insulator (SOI) show a unique turn off characteristic when compared to junction-isolated RESURF LIGBTs or vertical IGBTs. The turn off characteristic shows an extended `terrace' where, after the initial fast transient characteristic of IGBTs due to the loss of the electron current, the current stays almost at the same value for an extended period of time, before suddenly dropping to zero. In this paper, we show that this terrace arises because there is a value of LIGBT current during switch off where the rate of expansion of the depletion region with respect to the anode current is infinite. Once this level of anode current is approached, the depletion region starts to expand very rapidly, and is only stopped when it reaches the n-type buffer layer surrounding the anode. Once this happens, the current rapidly drops to zero. A quasi-static analytic model is derived to explain this behaviour. The analytically modelled turn off characteristic agrees well with that found by numerical simulation.

Copula-based Kernel Dependency Measures

The paper presents a new copula based method for measuring dependence between random variables. Our approach extends the Maximum Mean Discrepancy to the copula of the joint distribution. We prove that this approach has several advantageous properties. Similarly to Shannon mutual information, the proposed dependence measure is invariant to any strictly increasing transformation of the marginal variables. This is important in many applications, for example in feature selection. The estimator is consistent, robust to outliers, and uses rank statistics only. We derive upper bounds on the convergence rate and propose independence tests too. We illustrate the theoretical contributions through a series of experiments in feature selection and low-dimensional embedding of distributions.

Efficient truncation scheme for modeling deepwater mooring lines

This paper is aimed at enabling the confident use of existing model test facilities for ultra deepwater application without having to compromise on the widely accepted range of scales currently used by the floating production industry. Passive line truncation has traditionally been the preferred method of creating an equivalent numerical model at reduced depth; however, these techniques tend to suffer in capturing accurately line dynamic response and so reproducing peak tensions. In an attempt to improve credibility of model test data the proposed truncation procedure sets up the truncated model, based on line dynamic response rather than quasi-static system stiffness. The upper sections of each line are modeled in detail, capturing the wave action zone and all coupling effects with the vessel. These terminate to an approximate analytical model that aims to simulate the remainder of the line. Stages 1 & 2 are used to derive a water depth truncation ratio. Here vibration decay of transverse elastic waves is assessed and it is found that below a certain length criterion, the transverse vibrational characteristics for each line are inertia driven, hence with respect to these motions the truncated model can assume a linear damper whose coefficient depends on the local line properties and vibration frequency. Stage 3 endeavors to match the individual line stiffness between the full depth and truncated models. In deepwater it is likely that taut polyester moorings will be used which are predominantly straight and have high axial stiffness that provides the principal restoring force to static and low frequency vessel motions. Consequently, it means that the natural frequencies of axial vibrations are above the typical wave frequency range allowing for a quasi-static solution. In cases of exceptionally large wave frequency vessel motions, localized curvature at the chain seabed segment and tangential skin drag on the polyester rope can increase dynamic peak tensions considerably. The focus of this paper is to develop an efficient scheme based on analytic formulation, for replicating these forces at the truncation. The paper will close with an example case study of a single mooring under extreme conditions that replicates exactly the static and dynamic characteristics of the full depth line. Copyright © 2012 by the International Society of Offshore and Polar Engineers (ISOPE).

Structure Discovery in Nonparametric Regression through Compositional Kernel Search

Despite its importance, choosing the structural form of the kernel in nonparametric regression remains a black art. We define a space of kernel structures which are built compositionally by adding and multiplying a small number of base kernels. We present a method for searching over this space of structures which mirrors the scientific discovery process. The learned structures can often decompose functions into interpretable components and enable long-range extrapolation on time-series datasets. Our structure search method outperforms many widely used kernels and kernel combination methods on a variety of prediction tasks.

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.

Determinantal Clustering Processes - A Nonparametric Bayesian Approach to Kernel Based Semi-Supervised Clustering

Semi-supervised clustering is the task of clustering data points into clusters where only a fraction of the points are labelled. The true number of clusters in the data is often unknown and most models require this parameter as an input. Dirichlet process mixture models are appealing as they can infer the number of clusters from the data. However, these models do not deal with high dimensional data well and can encounter difficulties in inference. We present a novel nonparameteric Bayesian kernel based method to cluster data points without the need to prespecify the number of clusters or to model complicated densities from which data points are assumed to be generated from. The key insight is to use determinants of submatrices of a kernel matrix as a measure of how close together a set of points are. We explore some theoretical properties of the model and derive a natural Gibbs based algorithm with MCMC hyperparameter learning. The model is implemented on a variety of synthetic and real world data sets.

Unified analytic model for current-voltage behavior in amorphous oxide semiconductor TFTs

We present a simple and semi-physical analytical description of the current-voltage characteristics of amorphous oxide semiconductor thin-film transistors in the above-threshold and sub-threshold regions. Both regions are described by single unified expression that employs the same set of model parameter values directly extracted from measured terminal characteristics. The model accurately reproduces measured characteristics of amorphous semiconductor thin film transistors in general, yielding a scatter of < 4%. © 1980-2012 IEEE.

The Random Forest Kernel and other kernels for big data from random partitions

We present Random Partition Kernels, a new class of kernels derived by demonstrating a natural connection between random partitions of objects and kernels between those objects. We show how the construction can be used to create kernels from methods that would not normally be viewed as random partitions, such as Random Forest. To demonstrate the potential of this method, we propose two new kernels, the Random Forest Kernel and the Fast Cluster Kernel, and show that these kernels consistently outperform standard kernels on problems involving real-world datasets. Finally, we show how the form of these kernels lend themselves to a natural approximation that is appropriate for certain big data problems, allowing $O(N)$ inference in methods such as Gaussian Processes, Support Vector Machines and Kernel PCA.

Student-t processes as alternatives to Gaussian processes

We investigate the Student-t process as an alternative to the Gaussian process as a non-parametric prior over functions. We derive closed form expressions for the marginal likelihood and predictive distribution of a Student-t process, by integrating away an inverse Wishart process prior over the co-variance kernel of a Gaussian process model. We show surprising equivalences between different hierarchical Gaussian process models leading to Student-t processes, and derive a new sampling scheme for the inverse Wishart process, which helps elucidate these equivalences. Overall, we show that a Student-t process can retain the attractive properties of a Gaussian process - a nonparamet-ric representation, analytic marginal and predictive distributions, and easy model selection through covariance kernels - but has enhanced flexibility, and predictive covariances that, unlike a Gaussian process, explicitly depend on the values of training observations. We verify empirically that a Student-t process is especially useful in situations where there are changes in covariance structure, or in applications such as Bayesian optimization, where accurate predictive covariances are critical for good performance. These advantages come at no additional computational cost over Gaussian processes.