Neural population coding is optimized by discrete tuning curves

The sigmoidal tuning curve that maximizes the mutual information for a Poisson neuron, or population of Poisson neurons, is obtained. The optimal tuning curve is found to have a discrete structure that results in a quantization of the input signal. The number of quantization levels undergoes a hierarchy of phase transitions as the length of the coding window is varied. We postulate, using the mammalian auditory system as an example, that the presence of a subpopulation structure within a neural population is consistent with an optimal neural code.

Broadly tunable high-power random fibre laser

As shown recently, a long telecommunication fibre may be treated as a natural one-dimensional random system, where lasing is possible due to a combination of random distributed feedback via Rayleigh scattering by natural refractive index inhomogeneities and distributed amplification through the Raman effect. Here we present a new type of a random fibre laser with a narrow (∼1 nm) spectrum tunable over a broad wavelength range (1535-1570 nm) with a uniquely flat (∼0.1 dB) and high (>2 W) output power and prominent (>40 %) differential efficiency, which outperforms traditional fibre lasers of the same category, e.g. a conventional Raman laser with a linear cavity formed in the same fibre by adding point reflectors. Analytical model is proposed that explains quantitatively the higher efficiency and the flatter tuning curve of the random fiber laser compared to conventional one. The other important features of the random fibre laser like "modeless" spectrum of specific shape and corresponding intensity fluctuations as well as the techniques of controlling its output characteristics are discussed. Outstanding characteristics defined by new underlying physics and the simplicity of the scheme implemented in standard telecom fibre make the demonstrated tunable random fibre laser a very attractive light source both for fundamental science and practical applications such as optical communication, sensing and secure transmission. © 2012 Copyright Society of Photo-Optical Instrumentation Engineers (SPIE).

High-power widely tunable Raman fiber laser

Tunable Raman fiber lasers have attracted great interest owing to their high efficiency and reliability important for applications, such as optical fiber communications and sensing, spectroscopy, and instrument testing. Their tuning range is defined by the Raman gain bandwidth amounting to about 40 nm in telecom spectral range (∼1550 nm) for conventional silica single mode fibers (SMF). To increase the range, highly nonlinear fibers which broaden pump spectrum may be incorporated in the cavity of Raman fiber lasers, see e.g. [1]. Another approach is to involve Rayleigh scattering forming random distributed feedback in a relatively long fiber resulting in prominent flattening of the tuning curve [2]. In this paper we report on combination of these two techniques in tunable Raman fiber lasers thus providing great improvement of their output characteristics. © 2013 IEEE.

Upper and lower bounds on the learning curve for Gaussian processes

In this paper we introduce and illustrate non-trivial upper and lower bounds on the learning curves for one-dimensional Gaussian Processes. The analysis is carried out emphasising the effects induced on the bounds by the smoothness of the random process described by the Modified Bessel and the Squared Exponential covariance functions. We present an explanation of the early, linearly-decreasing behavior of the learning curves and the bounds as well as a study of the asymptotic behavior of the curves. The effects of the noise level and the lengthscale on the tightness of the bounds are also discussed.

A two-factor model of the U.K. yield curve

I model the forward premium in the U.K. gilt-edged market over the period 1982–96 using a two-factor general equilibrium model of the term structure of interest rates. The model permits the decomposition of the forward premium into separate components representing interest rate expectations, the risk premia associated with each of the underlying factors, and terms capturing the direct impact of the variances of the factors on the shape of the forward curve.

Improved supercontinuum generation by dispersion tuning and dual wavelength pumping

Supercontinuum generation in ultra-long Raman fibre laser cavities is compared for a range of fibre dispersions in the anomalous and normal regimes. For normal dispersion improved performance and efficiency is achieved using dual wavelength pumping.

Factors determining the growth curve of the foliose lichen Parmelia conspersa

Growth curves of the foliose lichen Parmelia conspersa (Ehrh. Ex Ach.)Ach. Were obtained by plotting radial growth (RGR, mm yr-1) of the fastest measured lobe, the slowest measured lobe, a randomly selected lobe, and by averaging a sample of lobes from each thallus against thallus diameter. Growth curves derived from the fastest-growing lobe and by averaging lobes were asymptotic and could be fitted by the growth model of Aplin and Hill. Mean lobe width increased with thallus size, reaching a maximum at approx. 4.5 cm thallus diameter. In four out of six thalli, radial growth of lobes over four months was positively correlated with initial lobe width or area. The RGR of isolated lobes was unaffected until the base of the lobe was removed to within 1-2 mm of the tip. The concentration (micrograms mg-1 biomass) of ribitol, arabitol and mannitol was greater in the marginal lobes of large than in small thalli. The results suggested that the growth curve of P. conspersa is determined by processes that occur within individual marginal lobes and can be explained by the Aplin and Hill model. Changes in lobe width and in the productive capacity of individual lobes with thallus size are likely to be more important factors than the degree of translocation within the lobe in determining the growth curve.

Growth curve of the lichen Rhizocarpon geographicum

Data on the growth curve of the lichen Rhizocarpon geographicum were obtained by measuring the radial growth rates (mm per 1.5 years) of 39 thalli from 2 to 65 mm in diameter growing in the same environment. An Aplin and Hill plot (r2 – r1 against ln r2 – ln r1) of the data and regression analyses suggested an initial phase of growth (up to a diameter of about 7 mm) in which the relative growth rate increased rapidly. This was followed by a phase in which the relative growth rate fell but the radial growth rate continued to rise (7 to 20 mm in diameter). Radial growth was then relatively constant until about 45 mm diameter and then declined. The Aplin and Hill model did not fit the data as a whole but may apply for a transient period in thalli between about 7 and 16 mm in diameter. The curve shows some similarities to that suggested by lichenometric studies but differs in showing a less steep decline in growth rate after the ‘great’ period.

Statnote 19: non-linear regression: fitting an exponential curve to data

Non-linear relationships are common in microbiological research and often necessitate the use of the statistical techniques of non-linear regression or curve fitting. In some circumstances, the investigator may wish to fit an exponential model to the data, i.e., to test the hypothesis that a quantity Y either increases or decays exponentially with increasing X. This type of model is straight forward to fit as taking logarithms of the Y variable linearises the relationship which can then be treated by the methods of linear regression.

Statnote 20: non-linear regression: fitting a general polynomial curve

In some circumstances, there may be no scientific model of the relationship between X and Y that can be specified in advance and indeed the objective of the investigation may be to provide a ‘curve of best fit’ for predictive purposes. In such an example, the fitting of successive polynomials may be the best approach. There are various strategies to decide on the polynomial of best fit depending on the objectives of the investigation.

Neural network enhanced self tuning adaptive control application for non-linear control of dynamic systems

The main theme of research of this project concerns the study of neutral networks to control uncertain and non-linear control systems. This involves the control of continuous time, discrete time, hybrid and stochastic systems with input, state or output constraints by ensuring good performances. A great part of this project is devoted to the opening of frontiers between several mathematical and engineering approaches in order to tackle complex but very common non-linear control problems. The objectives are: 1. Design and develop procedures for neutral network enhanced self-tuning adaptive non-linear control systems; 2. To design, as a general procedure, neural network generalised minimum variance self-tuning controller for non-linear dynamic plants (Integration of neural network mapping with generalised minimum variance self-tuning controller strategies); 3. To develop a software package to evaluate control system performances using Matlab, Simulink and Neural Network toolbox. An adaptive control algorithm utilising a recurrent network as a model of a partial unknown non-linear plant with unmeasurable state is proposed. Appropriately, it appears that structured recurrent neural networks can provide conveniently parameterised dynamic models for many non-linear systems for use in adaptive control. Properties of static neural networks, which enabled successful design of stable adaptive control in the state feedback case, are also identified. A survey of the existing results is presented which puts them in a systematic framework showing their relation to classical self-tuning adaptive control application of neural control to a SISO/MIMO control. Simulation results demonstrate that the self-tuning design methods may be practically applicable to a reasonably large class of unknown linear and non-linear dynamic control systems.

Spatial tuning studies of the pattern evoked electroretinogram

The locus of origin of the pattern evoked electroretinogram, (PERG), has been the subject of considerable discussion. A novel approach was adopted in this study to further elaborate the nature of the PERG evoked by pattern onset/offset presentation. The PERG was found to be linearly related to stimulus contrast and in particular was linearly related to the temporal contrast of the retinal image, when elicited by patterns of low spatial frequency. At high spatial frequencies the retinal image contrast is significantly reduced because of optical degradation. This is described by the eye's modulation transfer function (MTF). The retinal contrast of square wave grating and chequerboard patterns of increasing spatial frequency were found by filtering their Fourier transforms by the MTF. The filtered pattern harmonics were then resynthesised to constitute a profile of retinal image illuminance from which the temporal and spatial contrast of the image could be calculated. If the PERG is a pure illuminance response it should be spatially insensitive and dependent upon the temporal contrast of stimulation. The calculated loss of temporal contrast for finer patterns was expressed as a space-averaged temporal contrast attentuation factor. This factor, applied to PERGs evoked by low spatial frequency patterns, was used to predict the retinal illuminance response elicited by a finer pattern. The predicted response was subtracted from the recorded signal and residual waveform was proposed to represent specific activity. An additional correction for the attenuation of spatial contrast was applied to the extracted pattern specific response. Pattern specific responses computed for different spatial frequency patterns in this way are the predicted result of iso-contrast pattern stimulation. The pattern specific responses demonstrate a striking bandpass spatial selectivity which peaks at higher spatial frequencies in the more central retina. The variation of spatial sensitivity with eccentricity corresponds closely with estimated ganglion receptive field centre separation and psychophysical data. The variation of retinal structure with eccentricity, in the form of the volumes of the nuclear layers, was compared with the amplitudes of the computed retinal illuminance and pattern specific responses. The retinal illuminance response corresponds more closely to the outer and inner nuclear layers whilst the pattern specific response appears more closely related to the ganglion cell layer. In general the negative response transients correspond to the more proximal retinal layers. This thesis therefore supports the proposed contribution of proximal retinal cell activity to the PERG and describes techniques which may be further elaborated for more detailed studies of retinal receptive field dimensions.

Tuning of the charge in octahedral ferric complexes based on pyridoxal-N substituted thiosemicarbazone ligands

Four novel mononuclear coordination compounds namely: [Fe(Hthpy)2](SO4)1/2·3.5H2O 1, [Fe(Hthpy)2]NO3·3H2O 2, [Fe(H2mthpy)2](CH3C6H4SO3)3·CH3CH2OH 3 and [Fe(Hethpy)(ethpy)]·8H2O 4, (H2thpy = pyridoxalthiosemicarbazone, H2mthpy = pyridoxal-4-methylthiosemicarbazone, H2ethpy = pyridoxal-4-ethylthiosemicarbazone), were synthesized in the absence or presence of organic base, Et3N and NH3. Compounds 1 and 2 are monocationic, and were prepared using the singly deprotonated form of pyridoxalthiosemicarbazone. Both compounds crystallise in the monoclinic system, C2/c and P21/c space group for 1 and 2, respectively. Complex 3 is tricationic, it is formed with neutral bis(ligand) complex and possesses an interesting 3D channel architecture, the unit cell is triclinic, P1 space group. For complex 4, the pH value plays an important role during its synthesis; 4 is neutral and crystallises with two inequivalent forms of the ligand: the singly and the doubly deprotonated chelate of H2ethpy, the unit cell is monoclinic, C2/c space group. Notably, in 1 and 4, there is an attractive infinite three dimensional hydrogen bonding network in the crystal lattice. Magnetic measurements of 1 and 4 revealed that a rather steep spin transition from the low spin to high spin Fe(III) states occurs above 300 K in the first heating step. This transition is accompanied by the elimination of solvate molecules and thus, stabilizes the high spin form due to the breaking of hydrogen bonding networks; compared to 2 and 3, which keep their low spin state up to 400 K.

Fibre laser using a microchannel based loss tuning element for refractive index sensing

We have proposed and demonstrated a fibre laser system using a microchannel as a cavity loss tuning element for surrounding medium refractive index (SRI) sensing. A ~6µm width microchannel was created by femtosecond (fs) laser inscription assisted chemical etching in the cavity fibre, which offers a direct access to the external liquids. When the SRI changes, the microchannel behaves as a loss tuning element, hence modulating the laser cavity loss and output power. The results indicate that the presented laser sensing system has a linear response to the SRI with a sensitivity in the order of 10-5. Using higher pump power and more sensitive photodetector, the SRI sensitivity could be further enhanced.