Factors influencing robustness and effectiveness of conditional random fields in active learning frameworks

Active learning approaches reduce the annotation cost required by traditional supervised approaches to reach the same effectiveness by actively selecting informative instances during the learning phase. However, effectiveness and robustness of the learnt models are influenced by a number of factors. In this paper we investigate the factors that affect the effectiveness, more specifically in terms of stability and robustness, of active learning models built using conditional random fields (CRFs) for information extraction applications. Stability, defined as a small variation of performance when small variation of the training data or a small variation of the parameters occur, is a major issue for machine learning models, but even more so in the active learning framework which aims to minimise the amount of training data required. The factors we investigate are a) the choice of incremental vs. standard active learning, b) the feature set used as a representation of the text (i.e., morphological features, syntactic features, or semantic features) and c) Gaussian prior variance as one of the important CRFs parameters. Our empirical findings show that incremental learning and the Gaussian prior variance lead to more stable and robust models across iterations. Our study also demonstrates that orthographical, morphological and contextual features as a group of basic features play an important role in learning effective models across all iterations.

On the benefits of equicorrelation for portfolio allocation

The importance of modelling correlation has long been recognised in the field of portfolio management, with largedimensional multivariate problems increasingly becoming the focus of research. This paper provides a straightforward and commonsense approach toward investigating a number of models used to generate forecasts of the correlation matrix for large-dimensional problems.We find evidence in favour of assuming equicorrelation across various portfolio sizes, particularly during times of crisis. During periods of market calm, however, the suitability of the constant conditional correlation model cannot be discounted, especially for large portfolios. A portfolio allocation problem is used to compare forecasting methods. The global minimum variance portfolio and Model Confidence Set are used to compare methods, while portfolio weight stability and relative economic value are also considered.

Stability Of Large Damped Flexible Spacecraft With Stored Angular Momentum

Vibrational stability of large flexible structurally damped spacecraft carrying internal angular momentum and undergoing large rigid body rotations is analysed modeling the systems as elastic continua. Initially, analytical solutions to the motion of rigid gyrostats under torque-free conditions are developed. The solutions to the gyrostats modeled as axisymmetric and triaxial spacecraft carrying three and two constant speed momentum wheels, respectively, with spin axes aligned with body principal axes are shown to be complicated. These represent extensions of solutions for simpler cases existing in the literature. Using these solutions and modal analysis, the vibrational equations are reduced to linear ordinary differential equations. Equations with periodically varying coefficients are analysed applying Floquet theory. Study of a few typical beam- and plate-like spacecraft configurations indicate that the introduction of a single reaction wheel into an axisymmetric satellite does not alter the stability criterion. However, introduction of constant speed rotors deteriorates vibrational stability. Effects of structural damping and vehicle inertia ratio are also studied.

Stability of Fe3+ -VO defect pairs in polycrystalline Pb(Ti,Zr)O3

The intensity of the EPR signal with g = 5.985 arising from a ferric ion â oxygen vacancy defect pair (Fe3+ â VO) in PbTiO3, varies with the extent of PbO nonstoichiometry at constant Fe3+ content due to an increased oxygen vacancy concentration. In PZT solid solutions, the signal intensity decreases with an increase in Zr. A lower intensity is also noticed for Fe3+ â VO signals in PbZrO3. This behaviour is explained on the basis of PbO nonstoichiometry arising from independent Pb- and O-vacancies as well as the randomly distributed crystallographic shear (CS) plane defects. The contribution to PbO nonstoichiometry from CS planes is larger in high zirconium compositions of PZT.

Use of Multiblade Coordinates for Helicopter Flap-Lag Stability with Dynamic Inflow

Rotor flap-lag stability in forward flight is studied with and without dynamic inflow feedback via a multiblade coordinate transformation (MCT). The algebra of MCT is found to be so involved that it requires checking the final equations by independent means. Accordingly, an assessment of three derivation methods is given. Numerical results are presented for three- and four-bladed rotors up to an advance ratio of 0.5. While the constant-coefficient approximation under trimmed conditions is satisfactory for low-frequency modes, it is not satisfactory for high-frequency modes or for untrimmed conditions. The advantages of multiblade coordinates are pronounced when the blades are coupled by dynamic inflow.

Influence of spatial variability of permeability property on steady state seepage flow and slope stability analysis

In recent years, spatial variability modeling of soil parameters using random field theory has gained distinct importance in geotechnical analysis. In the present Study, commercially available finite difference numerical code FLAC 5.0 is used for modeling the permeability parameter as spatially correlated log-normally distributed random variable and its influence on the steady state seepage flow and on the slope stability analysis are studied. Considering the case of a 5.0 m high cohesive-frictional soil slope of 30 degrees, a range of coefficients of variation (CoV%) from 60 to 90% in the permeability Values, and taking different values of correlation distance in the range of 0.5-15 m, parametric studies, using Monte Carlo simulations, are performed to study the following three aspects, i.e., (i) effect ostochastic soil permeability on the statistics of seepage flow in comparison to the analytic (Dupuit's) solution available for the uniformly constant permeability property; (ii) strain and deformation pattern, and (iii) stability of the given slope assessed in terms of factor of safety (FS). The results obtained in this study are useful to understand the role of permeability variations in slope stability analysis under different slope conditions and material properties. (C) 2009 Elsevier B.V. All rights reserved.

Stability of a cylindrical plasma in the presence of a monotonic field-II

The stability of an incompressible inviscid, perfectly conducting cylindrical plasma against azimuthal disturbances in the presence of a monotonic decreasing magnetic field having a constant pitch is discussed by using energy principle. The results obtained by this principle are compared for m = 1 mode (which is a dangerous mode in which there is a lateral shift of the entire column) with that obtained by normal mode analysis. It is found that m = 1 mode is always unstable. Further, an axial line current, external axial field and the surface tension tend to stabilise m ≠ modes.

Stability of systems with power-law non-linearities

It is shown that a sufficient condition for the asymptotic stability-in-the-large of an autonomous system containing a linear part with transfer function G(jω) and a non-linearity belonging to a class of power-law non-linearities with slope restriction [0, K] in cascade in a negative feedback loop is ReZ(jω)[G(jω) + 1 K] ≥ 0 for all ω where the multiplier is given by, Z(jω) = 1 + αjω + Y(jω) - Y(-jω) with a real, y(t) = 0 for t < 0 and ∫ 0 ∞ |y(t)|dt < 1 2c2, c2 being a constant associated with the class of non-linearity. Any allowable multiplier can be converted to the above form and this form leads to lesser restrictions on the parameters in many cases. Criteria for the case of odd monotonic non-linearities and of linear gains are obtained as limiting cases of the criterion developed. A striking feature of the present result is that in the linear case it reduces to the necessary and sufficient conditions corresponding to the Nyquist criterion. An inequality of the type |R(T) - R(- T)| ≤ 2c2R(0) where R(T) is the input-output cross-correlation function of the non-linearity, is used in deriving the results.

L2-Stability Of A Class Of Nonlinear-Systems

Sufficient conditions are given for the L2-stability of a class of feedback systems consisting of a linear operator G and a nonlinear gain function, either odd monotone or restricted by a power-law, in cascade, in a negative feedback loop. The criterion takes the form of a frequency-domain inequality, Re[1 + Z(jω)] G(jω) δ > 0 ω ε (−∞, +∞), where Z(jω) is given by, Z(jω) = β[Y1(jω) + Y2(jω)] + (1 − β)[Y3(jω) − Y3(−jω)], with 0 β 1 and the functions y1(·), y2(·) and y3(·) satisfying the time-domain inequalities, ∝−∞+∞¦y1(t) + y2(t)¦ dt 1 − ε, y1(·) = 0, t < 0, y2(·) = 0, t > 0 and ε > 0, and , c2 being a constant depending on the order of the power-law restricting the nonlinear function. The criterion is derived using Zames' passive operator theory and is shown to be more general than the existing criteria

On the L2 stability of linear multidimensional time varying systems

This paper analyzes the L2 stability of solutions of systems with time-varying coefficients of the form [A + C(t)]x′ = [B + D(t)]x + u, where A, B, C, D are matrices. Following proof of a lemma, the main result is derived, according to which the system is L2 stable if the eigenvalues of the coefficient matrices are related in a simple way. A corollary of the theorem dealing with small periodic perturbations of constant coefficient systems is then proved. The paper concludes with two illustrative examples, both of which deal with the attitude dynamics of a rigid, axisymmetric, spinning satellite in an eccentric orbit, subject to gravity gradient torques.

Bounded-input/bounded-output stability of linear multidimensional time- varying systems

The present study of the stability of systems governed by a linear multidimensional time-varying equation, which are encountered in spacecraft dynamics, economics, demographics, and biological systems, gives attention the lemma dealing with L(inf) stability of an integral equation that results from the differential equation of the system under consideration. Using the proof of this lemma, the main result on L(inf) stability is derived according; a corollary of the theorem deals with constant coefficient systems perturbed by small periodic terms. (O.C.)

Structure and thermal stability relationships in ring-substituted arylammonium nitrates

The thermal stability of ring-substituted arylammonium nitrates has been investigated using thermal methods of analysis. The decomposition temperature of meta- and para-substituted derivatives is found to be linearly related to the Hammett substituent constant σ. The activation energy for decomposition determined by isothermal gravimetry increases with the increasing basicity of the corresponding amine. The results suggest that the primary step in the decomposition process of these salts is proton abstraction by the anion from the arylammonium ion.

Phase stability and dielectric properties of ceramics in the Pb(Zn1/3Nb2/3)O3-Pb(Ni1/3Nb2/3)O3-PbTio3 relaxor ferroelectric system

The perovskite structure in Pb(Zn1/3Nb2/3)O3 can be stabilized by the addition of Pb(Ni1/3Nb2/3)O3 and PbTiO3.Pb(Ni1/3Nb2/3)O3 assists in lowering the sintering temperature and shifting the Curie temperature of ceramics while PbTiO3 helps to optimize the dielectric properties. The phase stability and dielectric properties of several compositions in the Pb(Zn1/3Nb2/3)O3-Pb(Ni1/3Nb2/3)O3-PbTiO3 ternary relaxor ferroelectric system were investigated for possible capacitor applications. The effect of calcining and sintering temperature on the stability of perovskite phase in PZN rich compositions was studied extensively as a function of composition. The boundary line separating perovskite and mixed phases was determined for compositions near PZN. Several compositions can be sintered below 1050°C. The dielectric properties of compositions near the mixed phase boundary showed strong dependence on the percentage of pyrochlore phase. Compositions with a dielectric constant of 12.500 at room temperature have been identified which meet Z5T and Y5U specifications for dielectric constant and tan δ.

Microstructural stability and superplasticity in an electrodeposited nanocrystalline Ni-P alloy

Stabilization of nanocrystalline grain sizes by second phase particles can facilitate superplasticity at high strain rates and/or low temperatures. A metastable single phase nano-Ni-P alloy prepared by electrodeposition, with a grain size of similar to 6 nm, transforms to a nanoduplex structure at T> 673 K, with similar to 4 vol.% Ni3P particles at triple junctions and within Ni grains. The nanoduplex microstructure is reasonably stable up to 777 K, and the growth of Ni grains occurs in a coupled manner with the growth of Ni3P particles such that the ratio of the two mean sizes (Z) is essentially constant. High temperature tests for a grain size of 290 nm reveal superplastic behavior with an optimum elongation to failure of 810% at a strain rate of 7 x 10(-4) s(-1) and a relatively low temperature of 777 K. Superplastic deformation enhances both grain growth and the ratio Z, implying that grain boundary sliding (GBS) significantly influences the microstructural dynamics. Analysis of the deformation processes suggests that superplasticity is associated with GBS controlled by the overcoming of intragranular particles by dislocations, so that deformation is independent of the grain size. The nano-Ni-P alloy exhibits lower ductility than nano-Ni due to concurrent cavitation caused by higher stresses. (C) 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Linear stability, transient growth and the role of viscosity stratification in compressible Couette flow

Linear stability and the nonmodal transient energy growth in compressible plane Couette flow are investigated for two prototype mean flows: (a) the uniform shear flow with constant viscosity, and (b) the nonuniform shear flow with stratified viscosity. Both mean flows are linearly unstable for a range of supersonic Mach numbers (M). For a given M, the critical Reynolds number (Re) is significantly smaller for the uniform shear flow than its nonuniform shear counterpart; for a given Re, the dominant instability (over all streamwise wave numbers, α) of each mean flow belongs to different modes for a range of supersonic M. An analysis of perturbation energy reveals that the instability is primarily caused by an excess transfer of energy from mean flow to perturbations. It is shown that the energy transfer from mean flow occurs close to the moving top wall for “mode I” instability, whereas it occurs in the bulk of the flow domain for “mode II.” For the nonmodal transient growth analysis, it is shown that the maximum temporal amplification of perturbation energy, Gmax, and the corresponding time scale are significantly larger for the uniform shear case compared to those for its nonuniform counterpart. For α=0, the linear stability operator can be partitioned into L∼L̅ +Re2 Lp, and the Re-dependent operator Lp is shown to have a negligibly small contribution to perturbation energy which is responsible for the validity of the well-known quadratic-scaling law in uniform shear flow: G(t∕Re)∼Re2. In contrast, the dominance of Lp is responsible for the invalidity of this scaling law in nonuniform shear flow. An inviscid reduced model, based on Ellingsen-Palm-type solution, has been shown to capture all salient features of transient energy growth of full viscous problem. For both modal and nonmodal instability, it is shown that the viscosity stratification of the underlying mean flow would lead to a delayed transition in compressible Couette flow.