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.

A reversible infinite HMM using normalised random measures

Copyright 2014 by the author(s). We present a nonparametric prior over reversible Markov chains. We use completely random measures, specifically gamma processes, to construct a countably infinite graph with weighted edges. By enforcing symmetry to make the edges undirected we define a prior over random walks on graphs that results in a reversible Markov chain. The resulting prior over infinite transition matrices is closely related to the hierarchical Dirichlet process but enforces reversibility. A reinforcement scheme has recently been proposed with similar properties, but the de Finetti measure is not well characterised. We take the alternative approach of explicitly constructing the mixing measure, which allows more straightforward and efficient inference at the cost of no longer having a closed form predictive distribution. We use our process to construct a reversible infinite HMM which we apply to two real datasets, one from epigenomics and one ion channel recording.

MoT - Mixture of trees probabilistic graphical model for video segmentation

We present a novel mixture of trees (MoT) graphical model for video segmentation. Each component in this mixture represents a tree structured temporal linkage between super-pixels from the first to the last frame of a video sequence. Our time-series model explicitly captures the uncertainty in temporal linkage between adjacent frames which improves segmentation accuracy. We provide a variational inference scheme for this model to estimate super-pixel labels and their confidences in nearly realtime. The efficacy of our approach is demonstrated via quantitative comparisons on the challenging SegTrack joint segmentation and tracking dataset [23].

A derivation of the asymptotic random-coding prefactor

This paper studies the subexponential prefactor to the random-coding bound for a given rate. Using a refinement of Gallager's bounding techniques, an alternative proof of a recent result by Altuǧ and Wagner is given, and the result is extended to the setting of mismatched decoding. © 2013 IEEE.

A fully three-dimensional atomistic quantum mechanical study on random dopant-induced effects in 25-nm MOSFETs

A fully 3-D atomistic quantum mechanical simulation is presented to study the random dopant-induced effects in nanometer metal-oxide-semiconductor field-effect transistors. The empirical pseudopotential is used to represent the single particle Hamiltonian, and the linear combination of bulk band method is used to solve the million atom Schrodinger equation. The gate threshold fluctuation and lowering due to the discrete dopant configurations are studied. It is found that quantum mechanical effects increase the threshold fluctuation while decreasing the threshold lowering. The increase of threshold fluctuation is in agreement with the researchers' early study based on an approximated density gradient approach. However, the decrease in threshold lowering is in contrast with the density gradient calculations.

An ultra-low power CMOS random number generator

This paper proposes an ultra-low power CMOS random number generator (RING), which is based on an oscillator-sampling architecture. The noisy oscillator consists of a dual-drain MOS transistor, a noise generator and a voltage control oscillator. The dual-drain MOS transistor can bring extra-noise to the drain current or the output voltage so that the jitter of the oscillator is much larger than the normal oscillator. The frequency division ratio of the high-frequency sampling oscillator and the noisy oscillator is small. The RNG has been fabricated in a 0.35 mu m CMOS process. It can produce good quality bit streams without any post-processing. The bit rate of this RNG could be as high as 100 kbps. It has a typical ultra-low power dissipation of 0.91 mu W. This novel circuit is a promising unit for low power system and communication applications. (c) 2007 Elsevier Ltd. All rights reserved.

Random telegraph noise in the photoluminescence of individual GaxIn1-xAs quantum dots in GaAs

We have investigated random telegraph noise in the photoluminescence from InGaAs quantum dots in GaAs. Dots switching among two and three levels have been measured. The experiments show that the switching InGaAs dots behave very similarly to switching InP dots in GaInP. but differently from the more commonly investigated colloidal dots. The switching is attributed to defects, and we show that the switching can be used as a monitor of the defect.

Fully lithography independent fabrication of nanogap electrodes for lateral phase-change random access memory application

Submitted by 阎军 (yanj@red.semi.ac.cn) on 2010-06-07T01:33:41Z No. of bitstreams: 1 ApplPhysLett_96_213505.pdf: 1153920 bytes, checksum: 69931d8deb797813dd478b5dd0e292c0 (MD5)