Molecular Dynamics Simulation of Persistent Slip Bands Formation in Nickel-base Superalloys

Persistent slip band （PSB） is an important and typical microstructure generated during fatigue crack initiation. Intensive work has been done to investigate the mechanisms of the formation of persistent slip bands since the 1950s when Wadsworth[1] observed the fatigue fracture in copper. Simulations have indicated that PSBs formation during fatigue crack initiation is related to the dislocation driving force and interaction. In this paper, a molecular dynamics （MD） simulation associated with embedded atom model （EAM） is applied to the PSBs formation in nickel-base superalloys with different microstructure and temperature under tensile- tensile loadings. Five MD models with different microstructure （pure 5/ phase and γ/γ＇ phase）, grain orientation （[1 0 0][0 1 0][0 0 1] and [1 1 1][1 0 1][1 2 1]） and simulation temperature （300 K, 600 K, 900 K） were built up in these simulations. Our results indicated that within the γ phase by massive dislocations, pile-up and propagation which can penetrate the grain. Also, it is found that the temperature will affect the material fatigue performance and blur PSBs appearance. The simulation results are in strong agreement with published experimental test result. This simulation is based on the work[2]. The highlights of the article include： 1） investigation of the PSB formation via molecular dynamics simulation with three different parameters, 2） conduct of a new deformation and velocity combination controlled simulation for the PSB formation, 3） high-performance computing of PSB formation, and 4） systematic analysis of the PSB formation at the atomic scale in which the dislocation plays a critical role.

Defence Against Adversarial Attacks Using Clustering Algorithm

Deep learning model is vulnerable to adversarial examples in the task of image classification. In this paper, a cluster-based method for defending against adversarial examples is proposed. Each adversarial example before attacking a classifier is reconstructed by a clustering algorithm according to the pixel values. The MNIST database of handwritten digits was used to assess the defence performance of the method under the fast gradient sign method (FGSM) and the DeepFool algorithm. The defence model proposed is simple and the trained classifier does not need to be retrained.

Twist-teleportation-based local discrimination of maximally entangled states

In this work,we study the local distinguishability of maximally entangled states(MESs).In particular,we are concerned with whether any fixed number of MESs can be locally distinguishable for sufficiently large dimensions.Fan and Tian et al.have already obtained two satisfactory results for the generalized Bell states(GBSs)and the qudit lattice states when applied to prime or prime power dimensions.We construct a general twist-teleportation scheme for any orthonormal basis with MESs that is inspired by the method used in[Phys.Rev.A 70,022304(2004)].Using this teleportation scheme,we obtain a sufficient and necessary condition for one-way distinguishable sets of MESs,which include the GBSs and the qudit lattice states as special cases.Moreover,we present a generalized version of the results in[Phys.Rev.A 92,042320(2015)]for the arbitrary dimensional case.

Optimization RFID-enabled Retail Store Management with Complex Event Processing

Radio frequency identification(RFID) enabled retail store management needs workflow optimization to facilitate real-time decision making. In this paper, complex event processing(CEP) based RFID-enabled retail store management is studied, particularly focusing on automated shelf replenishment decisions. We define different types of event queries to describe retailer store workflow action over the RFID data streams on multiple tagging levels(e.g., item level and container level). Non-deterministic finite automata(NFA)based evaluation models are used to detect event patterns. To manage pattern match results in the process of event detection, optimization algorithm is applied in the event model to share event detection results. A simulated RFID-enabled retail store is used to verify the effectiveness of the method, experiment results show that the algorithm is effective and could optimize retail store management workflow.