An investigation of the effects of correlation and autocorrelation on classifier fusion and optimal classifier ensembles

The purpose of this research was to study various fusion strategies where the levels of correlation between features and auto-correlation within features could be controlled. The fusion strategies were chosen to reflect decision-level fusion (ISOC and ROC), feature level fusion, via a single Generalized Regression Neural Network (GRNN) employing all available features, and an intermediate level of fusion that employed the outputs of individual classifiers, in this case posterior probability estimates, before they are subjected to thresholds and mapped into decisions. This latter scheme involved fusing the posterior probability estimates by employing them as features in a probabilistic neural network. Correlation was injected into the data set both within a feature set (auto-correlation) and across feature sets, and sample size was varied for a two class problem. The fusion methods were then extended to three classifiers, and a method is demonstrated that selects the optimal classifier ensemble.     

Crystallography of duplex AlN–Nb(C,N) precipitates in 0.2% C steel

Crystallographic aspects related to the formation of duplex AlN–Nb(C,N) precipitates in 0.2% C steel have been evaluated by transmission electron microscopy. Nb(C,N) precipitation occurs with a compromise orientation relationship on favorably oriented AlN particles in the microstructure to minimize the barrier to Nb(C,N) nucleation and growth.     

Further improvements to the Bauer-Millward attack on the Hill cipher

The Bauer-Millward attack on the Hill cipher seeks to recover the individual rows of the decryption matrix instead of the entire matrix. The authors present some improvements to the Bauer-Millward attack that reduce the computational workload by a factor of φ(L), where φ is the Euler totient function, and L is the length of the alphabet. They also present a method of efficiently combining the top rows to recover the decryption matrix. Performance statistics are provided for blocks up to size 5, and example analyses are given for blocks up to size 8. Previous authors have stopped at size 4.     

The effects of grain-refining precipitates on the development of toughness in 4340 steel

Grain-refining precipitates have been identified as microstructural features that adversely affect the toughness of grain-refined, high-strength steels. For the case of an air-melt, aluminum-killed 4340 steel processed in a conventional manner, the presence of coarse A1N and TiN precipitates is shown to have a deleterious effect on both room-temperature impact toughness and fracture toughness. Coarse A1N precipitates, which evolve during the primary and secondary breakdown of the steel, are largely responsible for the development of quasicleavage fracture at ambient temperatures and the formation of fine-scale microvoids at upper-shelf temperatures. The refinement of grain-refining precipitates provides substantial increases in both lower-shelf and upper-shelf toughness for conditions associated with a critical damage mechanism of fracture. Improvements in upper-shelf toughness specifically result from a change in the species of grain-refining precipitate that nucleates secondary microvoids during the latter stages of fracture, viz. a refinement-induced transition from A1N to smaller TiN precipitates in the microstructure. This refinement-based mechanism of toughening is operative over a broad range of strength, although improvements in the transverse toughness of 4340 steel diminish with increases in sulfur content above 0.01 wt pct.