Geometrical learning, descriptive geometry, and biomimetic pattern recognition

Studies on learning problems from geometry perspective have attracted an ever increasing attention in machine learning, leaded by achievements on information geometry. This paper proposes a different geometrical learning from the perspective of high-dimensional descriptive geometry. Geometrical properties of high-dimensional structures underlying a set of samples are learned via successive projections from the higher dimension to the lower dimension until two-dimensional Euclidean plane, under guidance of the established properties and theorems in high-dimensional descriptive geometry. Specifically, we introduce a hyper sausage like geometry shape for learning samples and provides a geometrical learning algorithm for specifying the hyper sausage shapes, which is then applied to biomimetic pattern recognition. Experimental results are presented to show that the proposed approach outperforms three types of support vector machines with either a three degree polynomial kernel or a radial basis function kernel, especially in the cases of high-dimensional samples of a finite size. (c) 2005 Elsevier B.V. All rights reserved.

Kinetic study for hopping conduction through deoxyribonucleic acid molecules

Recent experiments indicated that disorder effect in deoxyribonucleic acid (DNA) may lead to a transition of the electronic hole transport mechanism from band resonant tunneling to thermally activated hopping. In this letter, based on Mott's variable-range hopping theory, we present a kinetic study for the hole transport properties of DNA molecules. Beyond the conventional argument in large-scale systems, our numerical study for finite-size DNA molecules reveals a number of unique features for: (i) the current-voltage characteristics, (ii) the temperature and length dependence, and (iii) the transition from conducting to insulating behaviors. (c) 2005 American Institute of Physics.

A speaker-independent continuous speech recognition system using biomimetic pattern recognition

In speaker-independent speech recognition, the disadvantage of the most diffused technology (HMMs, or Hidden Markov models) is not only the need of many more training samples, but also long train time requirement. This paper describes the use of Biomimetic pattern recognition (BPR) in recognizing some mandarin continuous speech in a speaker-independent manner. A speech database was developed for the course of study. The vocabulary of the database consists of 15 Chinese dish's names, the length of each name is 4 Chinese words. Neural networks (NNs) based on Multi-weight neuron (MWN) model are used to train and recognize the speech sounds. The number of MWN was investigated to achieve the optimal performance of the NNs-based BPR. This system, which is based on BPR and can carry out real time recognition reaches a recognition rate of 98.14% for the first option and 99.81% for the first two options to the persons from different provinces of China speaking common Chinese speech. Experiments were also carried on to evaluate Continuous density hidden Markov models (CDHMM), Dynamic time warping (DTW) and BPR for speech recognition. The Experiment results show that BPR outperforms CDHMM and DTW especially in the cases of samples of a finite size.

Analysis of atomic force microscopic results of InAs islands formed by molecular beam epitaxy

Atomic force microscopy (AFM) measurements of nanometer-sized islands formed by 2 monolayers of InAs by molecular beam epitaxy have been carried out and the scan line of individual islands was extracted from raw AFM data for investigation. It is found that the base widths of nanometer-sized islands obtained by AFM are not reliable due to the finite size and shape of the contacting probe. A simple model is proposed to analyze the deviation of the measured value From the real value of the base width of InAs islands. (C) 1998 Elsevier Science B.V. All rights reserved.

Biomimetic pattern recognition for speaker-independent speech recognition

In speaker-independent speech recognition, the disadvantage of the most diffused technology ( Hidden Markov Models) is not only the need of many more training samples, but also long train time requirement. This paper describes the use of Biomimetic Pattern Recognition (BPR) in recognizing some Mandarin Speech in a speaker-independent manner. The vocabulary of the system consists of 15 Chinese dish's names. Neural networks based on Multi-Weight Neuron (MWN) model are used to train and recognize the speech sounds. Experimental results are presented to show that the system, which can carry out real time recognition of the persons from different provinces speaking common Chinese speech, outperforms HMMs especially in the cases of samples of a finite size.

AFM扫描图像重构算法的改进

针对原子力显微镜(AFM)纳米成像中存在的失真问题,研究了通过探针建模实现AFM扫描图像重构方法.目前探针盲建模算法在重构AFM图像时存在较大误差,因此提出基于探针模型预估计的AFM扫描图像重构方法.该方法采用分区探针针尖建模,并通过基于该探针模型的反卷积运算实现AFM扫描图像重构,获得比较接近真实形貌的AFM扫描图像.文中介绍了算法的具体步骤,通过仿真和实验结果证明,该方法能够有效降低AFM图像重构时引入的误差,得到的图像更能反映样品表面真实的形貌。

On the Relationship Between Generalization Error, Hypothesis Complexity, and Sample Complexity for Radial Basis Functions

In this paper, we bound the generalization error of a class of Radial Basis Function networks, for certain well defined function learning tasks, in terms of the number of parameters and number of examples. We show that the total generalization error is partly due to the insufficient representational capacity of the network (because of its finite size) and partly due to insufficient information about the target function (because of finite number of samples). We make several observations about generalization error which are valid irrespective of the approximation scheme. Our result also sheds light on ways to choose an appropriate network architecture for a particular problem.

A Quantitative Evaluation of the AVITEWRITE Model of Handwriting Learning

Much sensory-motor behavior develops through imitation, as during the learning of handwriting by children. Such complex sequential acts are broken down into distinct motor control synergies, or muscle groups, whose activities overlap in time to generate continuous, curved movements that obey an intense relation between curvature and speed. The Adaptive Vector Integration to Endpoint (AVITEWRITE) model of Grossberg and Paine (2000) proposed how such complex movements may be learned through attentive imitation. The model suggest how frontal, parietal, and motor cortical mechanisms, such as difference vector encoding, under volitional control from the basal ganglia, interact with adaptively-timed, predictive cerebellar learning during movement imitation and predictive performance. Key psycophysical and neural data about learning to make curved movements were simulated, including a decrease in writing time as learning progresses; generation of unimodal, bell-shaped velocity profiles for each movement synergy; size scaling with isochrony, and speed scaling with preservation of the letter shape and the shapes of the velocity profiles; an inverse relation between curvature and tangential velocity; and a Two-Thirds Power Law relation between angular velocity and curvature. However, the model learned from letter trajectories of only one subject, and only qualitative kinematic comparisons were made with previously published human data. The present work describes a quantitative test of AVITEWRITE through direct comparison of a corpus of human handwriting data with the model's performance when it learns by tracing human trajectories. The results show that model performance was variable across subjects, with an average correlation between the model and human data of 89+/-10%. The present data from simulations using the AVITEWRITE model highlight some of its strengths while focusing attention on areas, such as novel shape learning in children, where all models of handwriting and learning of other complex sensory-motor skills would benefit from further research.

From Parallel Sequence Representations to Calligraphic Control: A Conspiracy of Neural Circuits

Calligraphic writing presents a rich set of challenges to the human movement control system. These challenges include: initial learning, and recall from memory, of prescribed stroke sequences; critical timing of stroke onsets and durations; fine control of grip and contact forces; and letter-form invariance under voluntary size scaling, which entails fine control of stroke direction and amplitude during recruitment and derecruitment of musculoskeletal degrees of freedom. Experimental and computational studies in behavioral neuroscience have made rapid progress toward explaining the learning, planning and contTOl exercised in tasks that share features with calligraphic writing and drawing. This article summarizes computational neuroscience models and related neurobiological data that reveal critical operations spanning from parallel sequence representations to fine force control. Part one addresses stroke sequencing. It treats competitive queuing (CQ) models of sequence representation, performance, learning, and recall. Part two addresses letter size scaling and motor equivalence. It treats cursive handwriting models together with models in which sensory-motor tmnsformations are performed by circuits that learn inverse differential kinematic mappings. Part three addresses fine-grained control of timing and transient forces, by treating circuit models that learn to solve inverse dynamics problems.

Analytic expressions for the constitutive parameters of magnetoelectric metamaterials

Electromagnetic metamaterials are artificially structured media typically composed of arrays of resonant electromagnetic circuits, the dimension and spacing of which are considerably smaller than the free-space wavelengths of operation. The constitutive parameters for metamaterials, which can be obtained using full-wave simulations in conjunction with numerical retrieval algorithms, exhibit artifacts related to the finite size of the metamaterial cell relative to the wavelength. Liu showed that the complicated, frequency-dependent forms of the constitutive parameters can be described by a set of relatively simple analytical expressions. These expressions provide useful insight and can serve as the basis for more intelligent interpolation or optimization schemes. Here, we show that the same analytical expressions can be obtained using a transfer-matrix formalism applied to a one-dimensional periodic array of thin, resonant, dielectric, or magnetic sheets. The transfer-matrix formalism breaks down, however, when both electric and magnetic responses are present in the same unit cell, as it neglects the magnetoelectric coupling between unit cells. We show that an alternative analytical approach based on the same physical model must be applied for such structures. Furthermore, in addition to the intercell coupling, electric and magnetic resonators within a unit cell may also exhibit magnetoelectric coupling. For such cells, we find an analytical expression for the effective index, which displays markedly characteristic dispersion features that depend on the strength of the coupling coefficient. We illustrate the applicability of the derived expressions by comparing to full-wave simulations on magnetoelectric unit cells. We conclude that the design of metamaterials with tailored simultaneous electric and magnetic response-such as negative index materials-will generally be complicated by potentially unwanted magnetoelectric coupling. © 2010 The American Physical Society.

Dispersing Light with Surface Plasmon Polaritonic Crystals

Spectral dispersion of light on a finite-size surface plasmon polaritonic (SPP) crystal has been studied. The angular wavelength separation of one or more orders of magnitude higher than in other state-of-the-art wavelength-splitting devices available to date has been demonstrated. The two-stage process is responsible for the dispersion value, which involves conversion of the incident light into SPP Bloch modes of a nanostructure followed by the SPP Bloch waves refraction at the SPP crystal boundary. The high spectral dispersion achievable in plasmonic devices may be useful for integrated high-resolution spectroscopy in nanophotonic, optical communication and lab-on-a-chip applications.

Domains in three-dimensional ferroelectric nanostructures

We report the experimental measurement of domains in single- crystal nanocolumns of ferroelectric BaTiO3, together with a theory of domain size scaling in three- dimensional structures which explains the observations.

Sub-Wavelength Profile 2-D Leaky-Wave Antennas With Two Periodic Layers

A fast and accurate analysis and synthesis technique for high-gain sub-wavelength 2-D Fabry-Perot leaky-wave antennas (LWA) consisting of two periodic metallodielectric arrays over a ground plane is presented. Full-wave method of moments (MoM) together with reciprocity is employed for the estimation of the near fields upon plane wave illumination and the extraction of the radiation patterns of the LWA. This yields a fast and rigorous tool for the characterisation of this type of antennas. A thorough convergence study for different antenna designs is presented and the operation principles of these antennas as well as the radiation characteristics are discussed. Moreover, design guidelines to tailor the antenna profile, the dimensions of the arrays as well as the antenna directivity and bandwidth are provided. A study on the radiation efficiency for antennas with different profiles is also presented and the trade off between directivity and radiation bandwidth is discussed. Numerical examples are given throughout to demonstrate the technique. A finite size antenna model is simulated using commercial software (CST Microstripes 2009) which validates the technique.

Thermal conductivity measurement of liquids in a microfluidic device

A new microfluidic-based approach to measuring liquid thermal conductivity is developed to address the requirement in many practical applications for measurements using small (microlitre) sample size and integration into a compact device. The approach also gives the possibility of high-throughput testing. A resistance heater and temperature sensor are incorporated into a glass microfluidic chip to allow transmission and detection of a planar thermal wave crossing a thin layer of the sample. The device is designed so that heat transfer is locally one-dimensional during a short initial time period. This allows the detected temperature transient to be separated into two distinct components: a short-time, purely one-dimensional part from which sample thermal conductivity can be determined and a remaining long-time part containing the effects of three-dimensionality and of the finite size of surrounding thermal reservoirs. Identification of the one-dimensional component yields a steady temperature difference from which sample thermal conductivity can be determined. Calibration is required to give correct representation of changing heater resistance, system layer thicknesses and solid material thermal conductivities with temperature. In this preliminary study, methanol/water mixtures are measured at atmospheric pressure over the temperature range 30-50A degrees C. The results show that the device has produced a measurement accuracy of within 2.5% over the range of thermal conductivity and temperature of the tests. A relation between measurement uncertainty and the geometric and thermal properties of the system is derived and this is used to identify ways that error could be further reduced.

Interaction-induced correlations and non-Markovianity of quantum dynamics

We investigate the conditions under which the trace distance between two different states of a given open system increases in time due to the interaction with an environment, therefore signaling non-Markovianity. We find that the finite-time difference in trace distance is bounded by two sharply defined quantities that are strictly linked to the occurrence of system-environment correlations created throughout their interaction and affecting the subsequent evolution of the system. This allows us to shed light on the origin of non-Markovian behaviors in quantum dynamics. We best illustrate our findings by tackling two physically relevant examples: a non-Markovian dephasing mechanism that has been the focus of a recent experimental endeavor and the open-system dynamics experienced by a spin connected to a finite-size quantum spin chain.