Kernelized sorting

Object matching is a fundamental operation in data analysis. It typically requires the definition of a similarity measure between the classes of objects to be matched. Instead, we develop an approach which is able to perform matching by requiring a similarity measure only within each of the classes. This is achieved by maximizing the dependency between matched pairs of observations by means of the Hilbert Schmidt Independence Criterion. This problem can be cast as one of maximizing a quadratic assignment problem with special structure and we present a simple algorithm for finding a locally optimal solution.

Distribution matching for transduction

Many transductive inference algorithms assume that distributions over training and test estimates should be related, e.g. by providing a large margin of separation on both sets. We use this idea to design a transduction algorithm which can be used without modification for classification, regression, and structured estimation. At its heart we exploit the fact that for a good learner the distributions over the outputs on training and test sets should match. This is a classical two-sample problem which can be solved efficiently in its most general form by using distance measures in Hilbert Space. It turns out that a number of existing heuristics can be viewed as special cases of our approach.

Contraction and observer design on cones

We consider the problem of positive observer design for positive systems defined on solid cones in Banach spaces. The design is based on the Hilbert metric and convergence properties are analyzed in the light of the Birkhoff theorem. Two main applications are discussed: positive observers for systems defined in the positive orthant, and positive observers on the cone of positive semi-definite matrices with a view on quantum systems. © 2011 IEEE.

Consensus in non-commutative spaces

Convergence analysis of consensus algorithms is revisited in the light of the Hilbert distance. The Lyapunov function used in the early analysis by Tsitsiklis is shown to be the Hilbert distance to consensus in log coordinates. Birkhoff theorem, which proves contraction of the Hilbert metric for any positive homogeneous monotone map, provides an early yet general convergence result for consensus algorithms. Because Birkhoff theorem holds in arbitrary cones, we extend consensus algorithms to the cone of positive definite matrices. The proposed generalization finds applications in the convergence analysis of quantum stochastic maps, which are a generalization of stochastic maps to non-commutative probability spaces. ©2010 IEEE.

Efficient decoding with generative score-spaces using the expectation semiring

State-of-the-art speech recognisers are usually based on hidden Markov models (HMMs). They model a hidden symbol sequence with a Markov process, with the observations independent given that sequence. These assumptions yield efficient algorithms, but limit the power of the model. An alternative model that allows a wide range of features, including word- and phone-level features, is a log-linear model. To handle, for example, word-level variable-length features, the original feature vectors must be segmented into words. Thus, decoding must find the optimal combination of segmentation of the utterance into words and word sequence. Features must therefore be extracted for each possible segment of audio. For many types of features, this becomes slow. In this paper, long-span features are derived from the likelihoods of word HMMs. Derivatives of the log-likelihoods, which break the Markov assumption, are appended. Previously, decoding with this model took cubic time in the length of the sequence, and longer for higher-order derivatives. This paper shows how to decode in quadratic time. © 2013 IEEE.

Sequence variations in the mitochondrial DNA control region and their implications for the phylogeny of the Cypriniformes

The mitochondrial DNA control region of six cobitids and two catostomids was sequenced and compared with sequences of other cypriniforms to study their sequence variations. The extended termination associated sequence (ETAS) domain, central domain, and conserved sequence block (CSB) domain were partitioned and the ETAS sequence, CSB-D, CSB-E, ECSB-F, CSB1, CSB2, and CSB3 were identified. It is suggested that the "hairpin" TACAT-ATGTA is the key sequence of ETAS and GACATA is the symbol of CSB1. Phylogenetic analysis based on the CSB domain showed that all cyprinids evolved as one monophyletic group, while the non-cyprinid Cypriniformes could be another monophyly that is in accordance with the hypothesis proposed by Siebert. Further analysis of the phylogeny of the Cobitoidei was also conducted and it is tentatively suggested that their relationships are Catostomidae + (Gyrinocheilidae + (Botiinae + (Homalopteridae + (Cobitinae + Nemacheilinae)))).

Digits speech recognition based on geometrical learning

We investigate the use of independent component analysis (ICA) for speech feature extraction in digits speech recognition systems.We observe that this may be true for a recognition tasks based on geometrical learning with little training data. In contrast to image processing, phase information is not essential for digits speech recognition. We therefore propose a new scheme that shows how the phase sensitivity can be removed by using an analytical description of the ICA-adapted basis functions via the Hilbert transform. Furthermore, since the basis functions are not shift invariant, we extend the method to include a frequency-based ICA stage that removes redundant time shift information. The digits speech recognition results show promising accuracy, Experiments show method based on ICA and geometrical learning outperforms HMM in different number of train samples.

基于模板匹配和SVM的草图符号自适应识别方法

草图符号的自适应学习中，不同用户的训练样本数量可能不同。保持在不同样本数量下良好的学习效果成为需要解决的一个重要问题．提出一种自适应的草图符号识别方法，该方法采用与训练样本个数相关的分类器组合策略将模板匹配方法和SVM统计分类方法进行了高效组合．它通过利用支持小样本学习的模板匹配方法和支持大量样本学习的SVM方法，并同时利用草图符号中的在线信息和离线信息，实现了不同样本个数下自适应的符号学习和识别．基于该方法，文中设计并实现了支持自适应识别的草图符号组件．最后，利用扩展的PIBGToolkit开发出原型系统IdeaNote．评估表明，该方法可以在24类草图符号分别使用1到20个训练样本时具有较高的识别正确率和较好的时间性能．

建筑平面图理解中对墙体符号的识别方法

一些由图纸矢量化或者CAD软件绘制的建筑平面图只包含基本的几何图形元素如线段、弧线等,需要通过符号识别得到关于建筑结构的信息,例如墙体拓扑结构、门窗的位置、型号等。该文提出了一种对建筑平面图的墙体符号识别算法,结合建筑平面图的领域知识,首先利用几何运算从平面图中提取出构成墙体候选符号的子图,再使用逻辑方法推理确定墙体元素符号并得到完整的墙体拓扑结构信息。

建筑平面图纸理解中关于符号识别的新方法

实现计算机图纸理解的关键在于对图纸中符号的识别,该文提出了一种建筑平面CAD图纸中符号识别的新方法,该方法首先由建筑符号中的几何图元构造出一种矩形网格结构,并用图元与网格之间的约束替换几何图元之间的约束来建立符号的几何特征描述。基于这种描述提出了一个统一的识别方法,最后给出了应用实例,取得了满意的识别效果。

Optimized logarithmic phase masks used to generate defocus invariant modulation transfer function for wavefront coding system

In a previous Letter [Opt. Lett. 33, 1171 (2008)], we proposed an improved logarithmic phase mask by making modifications to the original one designed by Sherif. However, further studies in another paper [Appl. Opt. 49, 229 (2010)] show that even when the Sherif mask and the improved one are optimized, their corresponding defocused modulation transfer functions (MTFs) are still not stable with respect to focus errors. So, by further modifying their phase profiles, we design another two logarithmic phase masks that exhibit more stable defocused MTF. However, with the defocus-induced phase effect considered, we find that the performance of the two masks proposed in this Letter is better than the Sherif mask, but worse than our previously proposed phase mask, according to the Hilbert space angle. (C) 2010 Optical Society of America

Crossing the phantom divide in brane cosmology with curvature corrections and brane-bulk energy transfer

We consider the Randall-Sundrum brane-world model with bulk-brane energy transfer where the Einstein-Hilbert action is modified by curvature correction terms: a four-dimensional scalar curvature from induced gravity on the brane, and a five-dimensional Gauss-Bonnet curvature term. It is remarkable that these curvature terms will not change the dynamics of the brane universe at low energy. Parameterizing the energy transfer and taking the dark radiation term into account, we find that the phantom divide of the equation of state of effective dark energy could be crossed, without the need of any new dark energy components. Fitting the two most reliable and robust SNIa datasets, the 182 Gold dataset and the Supernova Legacy Survey (SNLS), our model indeed has a small tendency of phantom divide crossing for the Gold dataset, but not for the SNLS dataset. Furthermore, combining the recent detection of the SDSS baryon acoustic oscillations peak (BAO) with lower matter density parameter prior, we find that the SNLS dataset also mildly favors phantom divide crossing.

Non-goal scene analysis for soccer video

The broadcast soccer video is usually recorded by one main camera, which is constantly gazing somewhere of playfield where a highlight event is happening. So the camera parameters and their variety have close relationship with semantic information of soccer video, and much interest has been caught in camera calibration for soccer video. The previous calibration methods either deal with goal scene, or have strict calibration conditions and high complexity. So, it does not properly handle the non-goal scene such as midfield or center-forward scene. In this paper, based on a new soccer field model, a field symbol extraction algorithm is proposed to extract the calibration information. Then a two-stage calibration approach is developed which can calibrate camera not only for goal scene but also for non-goal scene. The preliminary experimental results demonstrate its robustness and accuracy. (c) 2010 Elsevier B.V. All rights reserved.

High-dimensional assembly depending on polyoxoanion templates, metal ion coordination geometries, and a flexible bis(imidazole) ligand

By introducing the flexible 1,1'-(1,4-butanediyl)bis(imidazole) (bbi) ligand into the polyoxovanadate system, five novel polyoxoanion-templated architectures based on [As8V14O42](4-) and [V16O38Cl](6-) building blocks were obtained: [M(bbi)(2)](2)[As8V14O42(H2O)] [M = Co (1), Ni (2), and Zn (3)], [Cu(bbi)](4)[As8V14O42(H2O)] (4), and [Cu(bbi)](6)[V16O38Cl] (5). Compounds 1-3 are isostructural, and they exhibit a binodal (4,6)-connected 2D structure with Schlafli symbol (3(4)center dot 4(2))(3(4)center dot 4(4)center dot 5(4)center dot 6(3))(2), in which the polyoxoanion induces a closed four-membered circuit of M-4(bbi)(4). Compound 4 exhibits an interesting 3D framework constructed from tetradentate [As8V14O42](4-) cluster anions and cationic ladderlike double chains. There exists a bigger M-8(bbi)(6)O-2 circuit in 4. The 3D extended structure of 5 is composed of heptadentate [V16O38Cl](6-) anions and flexural cationic chains; the latter consists of six Cu(bbi) segments arranged alternately. It presents the largest 24-membered circuit of M-24(bbi)(24) so far observed made of bbi molecules and transition-metal cations. Investigation of their structural relations shows the important template role of the polyoxoanions and the synergetic interactions among the polyoxoanions, transition-metal ions, and flexible ligand in the assembly process.

多尺度分解与相位反褶积提高海上地震资料分辨率和信噪比

Offshore seismic exploration is full of high investment and risk. And there are many problems, such as multiple. The technology of high resolution and high S/N ratio on marine seismic data processing is becoming an important project. In this paper, the technology of multi-scale decomposition on both prestack and poststack seismic data based on wavelet and Hilbert-Huang transform and the theory of phase deconvolution is proposed by analysis of marine seismic exploration, investigation and study of literatures, and integration of current mainstream and emerging technology. Related algorithms are studied. The Pyramid algorithm of decomposition and reconstruction had been given by the Mallat algorithm of discrete wavelet transform In this paper, it is introduced into seismic data processing, the validity is shown by test with field data. The main idea of Hilbert-Huang transform is the empirical mode decomposition with which any complicated data set can be decomposed into a finite and often small number of intrinsic mode functions that admit well-behaved Hilbert transform. After the decomposition, a analytical signal is constructed by Hilbert transform, from which the instantaneous frequency and amplitude can be obtained. And then, Hilbert spectrum. This decomposition method is adaptive and highly efficient. Since the decomposition is based on the local characteristics of the time scale of data, it is applicable to nonlinear and non-stationary processes. The phenomenons of fitting overshoot and undershoot and end swings are analyzed in Hilbert-Huang transform. And these phenomenons are eliminated by effective method which is studied in the paper. The technology of multi-scale decomposition on both prestack and poststack seismic data can realize the amplitude preserved processing, enhance the seismic data resolution greatly, and overcome the problem that different frequency components can not restore amplitude properly uniformly in the conventional method. The method of phase deconvolution, which has overcome the minimum phase limitation in traditional deconvolution, approached the base fact well that the seismic wavelet is phase mixed in practical application. And a more reliable result will be given by this method. In the applied research, the high resolution relative amplitude preserved processing result has been obtained by careful analysis and research with the application of the methods mentioned above in seismic data processing in four different target areas of China Sea. Finally, a set of processing flow and method system was formed in the paper, which has been carried on in the application in the actual production process and has made the good progress and the huge economic benefit.