Noisy recurrent neural networks: the continuous-time case

The classical stochastic analog of the deterministic linear system in engineering is the linear system driven by white noise. As the promise of artificial neural networks in modeling nonlinear systems continues to grow, the need for a stochastic analog with quantitative foundations for analysis and synthesis will increase. This paper presents recent work in this direction, examining recurrent neural nets (RNN driven by white noise. We examine the effect of noise on the typical continuous-time RNN model. First, we perform qualitative analysis establishing uniform boundedness of moments of the neuron states over time. To enable practical application, however, it is necessary to relate these properties to useful measures that can be estimated. We thus subsequently derive bias and variance measures for the noisy RNN with respect to the corresponding deterministic RNN. This has significant practical implications, since net design is nonminimal in the sense that several nets can solve the same problem. The results allow the user to evaluate given RNN for noise performance. The designer can use these results to constrain the design space so that the design satisfies performance specifications whenever possible. An example is provided using the measures derived in this paper to predetermine the best among several RNN designs for a given problem.     

Design and characterization of asymetrical super-lattice Si/4H-SiC pin photo diode array: a potential opto-sensor for future applications in bio-medical domain

Microsystem Technologies - Photo-sensors are integral part of different bio-medical diagnostic equipment. Each type of bio-molecules possess unique spectral fingerprint in visible wavelength region...     

Favorable influence of ssDNA-functionalized SWCNT on the navigation pattern of C. elegans

Microsystem Technologies - Single-Walled Carbon Nanotubes (SWCNTs) are widely used as potential carriers in drug delivery systems. The objective of this work was to observe the effects of pristine,...     

Design of an Ultra-Compact and Highly-Sensitive Temperature Sensor Using Photonic Crystal Based Single Micro-Ring Resonator and Cascaded Micro-Ring Resonator

Silicon - In the present report, a photonic crystal based micro-ring resonator (MRR) structure is proposed which is very compact in size and has very fast response and is employed for temperature...     

Recovery and Recycling Silica Flux in Submerged Arc Welding – Acceptable Properties and Economical Correlation

Silicon - This research aims at optimum utilization of waste recycled submerged arc welding slag along with virgin flux in place of pure virgin flux without compromising the quality of welding and...     

Evaluation of wear of turning carbide inserts using neural networks

Recent trends, being towards mostly unmanned automated machining systems and consistent system operations, need reliable on-line monitoring processes. A proper on-line cutting tool condition monitoring system is essential for deciding when to change the tool. Many methods have been attempted in this connection.Recently, artificial neural networks have been tried for this purpose because of its inherent simplicity and reasonably quick data-processing capability. The present work uses the back propagation algorithm for training the neural network of 5-3-1 structure. The technique shows close matching of estimation of average flank wear and directly measured wear value. Thus the system developed demonstrates the possibility of successful tool wear monitoring on-line.     

Study of polyaniline polyacrylamide composites by positron annihilation technique

Polyaniline doped with nonoxidizing Bronsted acids is recognized as a conducting material, as its electrical conductivity changes with percentage of doping. In the present work, different percentages of doped polyaniline were blended with polyacrylamide and their electrical conductivities as well as the positron annihilation lifetimes were measured. Analysis of data yielded three lifetime components. It was observed that the value of the short lifetime component remained constant for doping concentration, whereas that of the intermediate component τ₂ decreased. The relative intensity pertaining to τ₂, however, increased with the increase in doped PANI concentration. © 2003 Wiley Periodicals, Inc. J Appl Polym Sci 89: 930–933, 2003     

Al2O3─SiC Composites from Aluminosilicate Precursors

Al₂O₃ and SiC composite materials have been produced from mixtures of aluminosilicates (both natural minerals and synthetic) and carbon as precursor materials. These composites are produced by heating a mixture of kaolinite (or synthetic aluminosilicates) and carbon in stoichiometric proportion above 1550°C, so that only Al₂O₃ and SiC remain as the major phases. A similar process has also been used for synthesizing other composite powders having mixtures of Al₂O₃, SiC, TiC, and ZrO₂ in different proportions (all compounds together or selective mixtures of some of them), as desired. The microstructure of hot-pressed dense compacts, produced from these powders, revealed that the SiC phase is distributed very homogeneously, even occasionally within Al₂O₃ grains on a nanosize scale. The homogeneous distribution of SiC particles within the system produced high fracture toughness of the hot-pressed material (K_IC∼ 7.0 MPa · m^1/2) and having Vicker's hardness values greater than 2000 kgf/mm².