37 resultados para MOVING MIRRORS
em Aston University Research Archive
Resumo:
The 'moving targets' algorithm for training recurrent networks is reviewed and applied to a task which demonstrates the ability of this algorithm to use distant contextual information. Some practical difficulties are discussed, especially with regard to the minimization process. Results on performance and computational requirements of several different 2nd-order minimization algorithms are presented for moving target problems.
Resumo:
A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The algorithm resembles back-propagation in that an error function is minimized using a gradient-based method, but the optimization is carried out in the hidden part of state space either instead of, or in addition to weight space. Computational results are presented for some simple dynamical training problems, one of which requires response to a signal 100 time steps in the past.
Resumo:
A simple method for training the dynamical behavior of a neural network is derived. It is applicable to any training problem in discrete-time networks with arbitrary feedback. The method resembles back-propagation in that it is a least-squares, gradient-based optimization method, but the optimization is carried out in the hidden part of state space instead of weight space. A straightforward adaptation of this method to feedforward networks offers an alternative to training by conventional back-propagation. Computational results are presented for simple dynamical training problems, with varied success. The failures appear to arise when the method converges to a chaotic attractor. A patch-up for this problem is proposed. The patch-up involves a technique for implementing inequality constraints which may be of interest in its own right.
Resumo:
Theory suggests that the dimensions that are incorporated in the new product screening decision will differ according to the stage of the development process. The outcome of the application of different screening dimensions would be quicker, realistic and more reliable screening decisions. This research project builds on existing new product development and screening literature by investigating new product screening in international fast moving consumer goods companies. It further builds on the existing literature by measuring decision-making relating to projects in 'real time', as managers' responses refer to projects they are currently working on. The introduction of branded consumer products allows us to evolve scales used in new product research by further developing variables relating to branding, promotion and retailer power. The project uncovers multiple dimensions of new product screening and evaluation within this branded product sector. These dimensions are found to differ in their ability to discriminate between two groups of accepted and rejected projects at each of four stages of the new product development process. This investigation provides the intelligence with which managers can determine the likelihood of project acceptance and rejection at different stages of the development process. It highlights the need for managers to apply stage-specific dimensions in the new product screening decision and advocates the redefinition of new product screening from both an academic and managerial perspective. The screening decision should not be viewed as a single, early decision in a product development process, but as a series of stage specific decisions regarding future project potential.
Resumo:
Observation of autosoliton propagation in a dispersion-managed optical transmission system controlled by in-line nonlinear fiber loop switches is reported for what is believed to be the first time. The system is based on a strong dispersion map with large amplifier spacing. Operation at transmission rates of 10 and 40 Gbits/s is demonstrated. ©2004 Optical Society of America.
Resumo:
We demonstrate multiple-peaked switching in a nonlinear-optical loop mirror and present an experimental investigation of device cascading in the soliton regime based on a sequence of two independent nonlinear-optical loop mirrors. Cascading leads to an enhanced switching response with sharper switching edges, flattened peaks, and increased interpeak extinction ratios. We observe that pulses emerging from the cascade retain the sech2 temporal profile of a soliton with minimal degradation in the spectral characteristics.
Resumo:
The authors show that by inserting nonlinear optical loop mirrors into an optical fibre transmission line, 1.5 ps solitons may be transmitted over at least 750 km, with amplifiers spaced at 15 km intervals.
Resumo:
We show that by inserting nonlinear optical loop mirrors into an optical fibre transmission line, that 1.5 ps solitons may be transmitted over at least 750 km, with amplifiers spaced at 15 km intervals.
Resumo:
The development of an all-optical communications infrastructure requires appropriate optical switching devices and supporting hardware. This thesis presents several novel fibre lasers which are useful pulse sources for high speed optical data processing and communications. They share several attributes in common: flexibility, stability and low-jitter output. They all produce short (picosecond) and are suitable as sources for soliton systems. The lasers are all-fibre systems using erbium-doped fibre for gain, and are actively-modelocked using a dual-wavelength nonlinear optical loop mirror (NOLM) as a modulator. Control over the operating wavelength and intra-cavity dispersion is obtained using a chirped in-fibre Bragg grating.Systems operating both at 76MHz and gigahertz frequencies are presented, the latter using a semiconductor laser amplifier to enhance nonlinear action in the loop mirror. A novel dual-wavelength system in which two linear cavities share a common modulator is presented with results which show that the jitter between the two wavelengths is low enough for use in switching experiments with data rates of up to 130Gbit/s.
Resumo:
Practitioners and academics are in broad agreement that, above all, organizations need to be able to learn, to innovate and to question existing ways of working. This thesis develops a model to take into account, firstly, what determines whether or not organizations endorse practices designed to facilitate learning. Secondly, the model evaluates the impact of such practices upon organizational outcomes, measured in terms of products and technological innovation. Researchers have noted that organizations that are committed to producing innovation show great resilience in dealing with adverse business conditions (e.g. Pavitt, 1991; Leonard Barton, 1998). In effect, such organizations bear many of the characteristics associated with the achievement of ‘learning organization’ status (Garvin, 1993; Pedler, Burgoyne & Boydell, 1999; Senge, 1990). Seven studies are presented to support this theoretical framework. The first empirical study explores the antecedents to effective learning. The three following studies present data to suggest that people management practices are highly significant in determining whether or not organizations are able to produce sustained innovation. The thesis goes on to explore the relationship between organizational-level job satisfaction, learning and innovation, and provides evidence to suggest that there is a strong, positive relationship between these variables. The final two chapters analyze learning and innovation within two similar manufacturing organizations. One manifests relatively low levels of innovation whilst the other is generally considered to be outstandingly innovative. I present the comparative framework for exploring the different approaches to learning manifested by the two organizations. The thesis concludes by assessing the extent to which the theoretical model presented in the second chapter is borne out by the findings of the study. Whilst this is a relatively new field of inquiry, findings reveal that organizations have a much stronger chance of producing sustained innovation where they manage people proactively where people process themselves to be satisfied at work. Few studies to date have presented empirical evidence to substantiate theoretical endorsements to engage in higher order learning, so this research makes an important contribution to existing literature in this field.
Resumo:
We investigate numerically the dependence of higher harmonics of the space-charge field on the detuning frequency between the pump waves, which form a running interference pattern. Bistability and hysteresis of harmonics are predicted for a contrast of the interference pattern m =(0.25-0.3). For contrasts m˜1 and small detuning frequencies we show the existence of a narrow resonance, connected with the nonlinear excitation of a slowly decreasing sequence of spatial harmonics. For experiments we use a BSO crystal in the optical configuration which avoids nonlinear optical distortions. The experimental data show good qualitative agreement with theory.
Resumo:
A preliminary study by Freeman et al (1996b) has suggested that when complex patterns of motion elicit impressions of 2-dimensionality, odd-item-out detection improves given targets can be differentiated on the basis of surface properties. Their results can be accounted for, it if is supposed that observers are permitted efficient access to 3-D surface descriptions but access to 2-D motion descriptions is restricted. To test the hypothesis, a standard search technique was employed, in which targets could be discussed on the basis of slant sign. In one experiment, slant impressions were induced through the summing of deformation and translation components. In a second theory were induced through the summing of shear and translation components. Neither showed any evidence of efficient access. A third experiment explored the possibility that access to these representations may have been hindered by a lack of grouping between the stimuli. Attempts to improve grouping failed to produce convincing evidence in support of life. An alternative explanation is that complex patterns of motion are simply not processed simultaneously. Psychophysical and physiological studies have, however, suggested that multiple mechanisms selective for complex motion do exist. Using a subthreshold summation technique I found evidence supporting the notion that complex motions are processed in parallel. Furthermore, in a spatial summation experiment, coherence thresholds were measured for displays containing different numbers of complex motion patches. Consistent with the idea that complex motion processing proceeds in parallel, increases in the number of motion patches were seen to decrease thresholds, both for expansion and rotation. Moreover, the rates of decrease were higher than those typically expected from probability summation, thus implying mechanisms are available, which can pool signals from spatially distinct complex motion flows.