神经网络在少数投影图像重建中的应用

本文提出了将两种神经网络模型进行套用的图像重建算法。算法将图像重建问题转化为神经网络优化计算,利用Hopfield神经网络（HNN）将各种优化准则转化为模型的能量函数,并应用Adaline网络调整各准则在能量函数中所占的比重,从而求取图像重建的最优解。将其应用于少数投影图像重建,体现出重建精度高、收敛快等特点,取得了令人满意的效果。     

激光实时全息法用于双尖牙受力分析的研究

本文介绍了激光实时全息法用于双尖牙受力分析的研究。模型试验表明：在双尖牙牙尖部位及牙颈部位有明显的应力集中,牙尖部位受力较大,牙颈部位次之,折裂大多从两牙尖的正中裂开,折裂方式与临床观察相吻合。本文对此进行了相应的分析,并对牙骨质剥落等问题提出防治的建议。     

Non‐Lithographic Fabrication of All‐2D α‐MoTe2 Dual Gate Transistors

As one of the emerging new transition‐metal dichalcogenides materials, molybdenum ditelluride (α‐MoTe₂) is attracting much attention due to its optical and electrical properties. This study fabricates all‐2D MoTe₂‐based field effect transistors (FETs) on glass, using thin hexagonal boron nitride and thin graphene in consideration of good dielectric/channel interface and source/drain contacts, respectively. Distinguished from previous works, in this study, all 2D FETs with α‐MoTe₂ nanoflakes are dual‐gated for driving higher current. Moreover, for the present 2D dual gate FET fabrications on glass, all thermal annealing and lithography processes are intentionally exempted for fully non‐lithographic method using only van der Waal's forces. The dual‐gate MoTe₂ FET displays quite a high hole and electron mobility over ≈20 cm² V^?1 s^?1 along with ON/OFF ratio of ≈10⁵ in maximum as an ambipolar FET and also demonstrates high drain current of a few tens‐to‐hundred μA at a low operation voltage. It appears promising enough to drive organic light emitting diode pixels and NOR logic functions on glass.     

Scalable Model of On-Wafer Interconnects for High-Speed CMOS ICs

This paper describes the development of an equivalent circuit model of on-wafer interconnects for high-speed CMOS integrated circuits. By strategically cascading two-pi blocks together, the lumped model can characterize the distributed effects. Besides, the elaborately proposed model characterizes the frequency-variant characteristics with frequency-independent components. Thus, the model can be easily plugged into commercial computer-aided design tools. By adopting a newly invented optimization algorithm, namely, particle swarm optimization (PSO), the model parameters are extracted and formulated as empirical expressions. Therein, with each set of the geometrical parameters, the interconnect behaviors can be accurately predicted. The accuracy of the model is validated by comparisons with the on-wafer measurements up to 30 GHz. Moreover, the scalability of the proposed model is also discussed     

Disclosing the Role of Defect-Engineered Metal–Organic Frameworks in Mixed Matrix Membranes for Efficient CO2 Separation: A Joint Experimental-Computational Exploration

Incorporation of defects in metal–organic frameworks (MOFs) offers new opportunities for manipulating their microporosity and functionalities. The so-called “defect engineering” has great potential to tailor the mass transport properties in MOF/polymer mixed matrix membranes (MMMs) for challenging separation applications, for example, CO₂ capture. This study first investigates the impact of MOF defects on the membrane properties of the resultant MOF/polymer MMMs for CO₂ separation. Highly porous defect-engineered UiO-66 nanoparticles are successfully synthesized and incorporated into a CO₂-philic crosslinked poly(ethylene glycol) diacrylate (PEGDA) matrix. A thorough joint experimental/simulation characterization reveals that defect-engineered UiO-66/PEGDA MMMs exhibit nearly identical filler–matrix interfacial properties regardless of the defect concentrations of their parental UiO-66 filler. In addition, non-equilibrium molecular dynamics simulations in tandem with gas transport studies disclose that the defects in MOFs provide the MMMs with ultrafast transport pathways mainly governed by diffusivity selectivity. Ultimately, MMMs containing the most defective UiO-66 show the most enhanced CO₂/N₂ separation performance—CO₂ permeability = 470 Barrer (four times higher than pure PEGDA) and maintains CO₂/N₂ selectivity = 41—which overcomes the trade-off limitation in pure polymers. The results emphasize that defect engineering in MOFs would mark a new milestone for the future development of optimized MMMs.     

Thermoelectric and Magnetic Properties of Pt-Substituted $${BaFe_{4-{x}}Pt_{{x}}Sb_{12}}$$ Compounds

\({BaFe_{4-{x}}Pt_{{x}}Sb_{12}}\) (x = 0, 0.1, 0.2) compounds were prepared by melting and annealing, followed by a spark plasma sintering method. Low-temperature thermoelectric and magnetic properties were investigated based on Seebeck coefficient, electrical and thermal conductivity and magnetization measurements. The structural properties of \({BaFe_{4-{x}}Pt_{{x}}Sb_{12}}\) (x = 0, 0.1, 0.2) compounds were ascertained by powder x-ray diffraction analysis, confirming that all samples have a main phase of a skutterudite structure with the space group Im\({\mathrm {\bar{3}}}\). The lattice parameters obtained, 9.202(5), 9.199(5) and 9.202(1) Å for x = 0, 0.1 and 0.2, respectively, were found consistent with literature. The Seebeck coefficient sign shows that holes are dominant carriers in all compounds. The local maximum Seebeck coefficient was observed around 50 K which may be a trace of paramagnon-drag effect of charge carriers. Thermal conductivity and electrical resistivity measurements were carried out between 4.2 and 300 K. Temperature dependence of electrical resistivity reflects that all samples show semi-metallic behavior in our temperature range of 4.2–300 K. Samples for x = 0.1 and x = 0.2 show Kondo-like behavior. In magnetization measurement, we observe that there are two successive magnetic transitions in Pt-substituted compounds; however, there is only one (transition from a paramagnetic state to long-range magnetic ordering) in Pt-free compounds. In Pt-substituted compounds, the first transition appears at \( T _{ {\rm c}}\) = 48 K. In addition, the second transition is observed at \( T _{ {\rm irr}}\) = 30 K where an intermediate state is observed before the magnetic ordering transforms to an irreversible ferromagnetic state. We concluded that Pt substitution on the Fe side effectual on the thermoelectric and magnetic properties of \({BaFe_{4-{x}}Pt_{{x}}Sb_{12}}\) (x = 0, 0.1, 0.2) compounds.     

Performance analysis of bandwidth allocation strategy with state-dependent Bernoulli access and preemptive priority in wideband integrated networks

In this paper, we study a bandwidth allocation strategy with state-dependent Bernoulli access (SDBA) and preemptive priority (PP) to serve wideband (WB) and narrowband (NB) traffics in wideband integrated networks. In this strategy, total bandwidth is divided into three regions, bandwidth for WB traffic, bandwidth for NB traffic and shared bandwidth. Under this environment, the minimum usable bandwidths for queueable NB traffic are controlled adaptively according to the system states. And the bandwidths used by the NB traffic are probabilistically preempted by a blockable WB traffic, when the number of NB messages waiting in the buffer is varying. The strategy is analyzed using the Neuts' matrix analytic approach. It shows better performance than other schemes which do not use SDBA and PP. In addition, the strategy with multiple buffer thresholds controls bandwidth allocation more dynamically and shows more performance build-up than the strategy with single buffer threshold.     

Channel assignment with QoS guarantees for a multiclass multicode CDMA system

In a wireless system that supports multimedia services, each traffic requires different quality of service (QoS) at both communication on radio links and connection admission. We initially derive the uplink capacity satisfying the QoS constraint on radio links in a multiclass multicode code-division multiple-access (CDMA) system. Based on the derived capacity, the number of channel elements, which is one of the system resources, is determined. Then, we define the QoS parameters associated with connection processes. To guarantee the defined QoS at the connection level, under given channel elements, we propose a channel-assignment scheme with dynamic priority adjustment (DPA). The proposed scheme gives multipriority to different traffic classes. Real-time classes can preempt non-real-time classes with restricted preemptive priority. Such restriction is regulated by preemption-free code channels and a buffer threshold for non-real-time classes. Among real-time classes, different priorities are assigned to each traffic class by code reservation parameters. These multipriority parameters are dynamically adjusted in order to guarantee different QoS requirements. We analyze the DPA scheme by the matrix-geometric method, and evaluate the performance of each traffic class. The results show that the proposed scheme flexibly guarantees QoS depending on traffic loading condition and achieves high channel utilization.     

Modeling of power supply noise in large chips using the circuit-based finite-difference time-domain method

In this paper, a multilayered on-chip power distribution network consisting of two million passive elements has been modeled using the finite-difference time-domain (FDTD) method. In this method, a branch capacitor has been used. The use of the branch capacitor is important for simulating multilayered power grids. In addition, a method for including the CMOS inverter characteristics into the FDTD simulation has been presented. As an example of the application of this method, an H-tree clock network was simulated to compute the power supply noise distribution across an entire chip. Various scenarios with varying decoupling capacitances, load capacitances, number of clock buffers, and rise times have been analyzed to demonstrate the importance of circuit nonlinearity on power supply noise. Also, a method has been presented for analyzing package and board planes. Based on the methods presented, the interaction between chip and package has been discussed for capturing the resonant behavior that is otherwise absent when each section of the system is analyzed separately.     

Potential Anomaly Separation and Archeological Site Localization Using Genetically Trained Multi‐level Cellular Neural Networks

In this paper, a supervised algorithm for the evaluation of geophysical sites using a multi‐level cellular neural network (ML‐CNN) is introduced, developed, and applied to real data. ML‐CNN is a stochastic image processing technique based on template optimization using neighborhood relationships of the pixels. The separation/enhancement and border detection performance of the proposed method is evaluated by various interesting real applications. A genetic algorithm is used in the optimization of CNN templates. The first application is concerned with the separation of potential field data of the Dumluca chromite region, which is one of the rich reserves of Turkey; in this context, the classical approach to the gravity anomaly separation method is one of the main problems in geophysics. The other application is the border detection of archeological ruins of the Hittite Empire in Turkey. The Hittite civilization sites located at the Sivas‐Altinyayla region of Turkey are among the most important archeological sites in history, one reason among others being that written documentation was first produced by this civilization.