Pool-based active learning in approximate linear regression

The goal of pool-based active learning is to choose the best input points to gather output values from a ‘pool’ of input samples. We develop two pool-based active learning criteria for linear regression. The first criterion allows us to obtain a closed-form solution so it is computationally very efficient. However, this solution is not necessarily optimal in the single-trial generalization error analysis. The second criterion can give a better solution, but it does not have a closed-form solution and therefore some additional search strategy is needed. To cope with this problem, we propose a practical procedure which enables us to efficiently search for a better solution around the optimal solution of the first method. Simulations with toy and benchmark datasets show that the proposed active learning method compares favorably with other active learning methods as well as the baseline passive learning scheme. Furthermore, the usefulness of the proposed active learning method is also demonstrated in wafer alignment in semiconductor exposure apparatus.     

High-pressure torsion for production of magnetoresistance in Cu–Co alloy

The process of high-pressure torsion (HPT) was applied to control the size and distribution of ferromagnetic Co particles in a Cu–Co alloy. Electron probe microanalysis, X-ray diffraction analysis, and transmission electron microscopy confirmed that the Co particles were significantly refined through fragmentation and dissolved with intense straining by HPT. Magnetoresistance appeared by ~2.5% at 77 K with an isotropic feature corresponding to giant magnetoresistance (GMR). It is demonstrated that HPT is a potential process for creating GMR in the Cu–Co alloy prepared by conventional ingot metallurgy.     

Pressure dependence of magnetoresistance for Fe/Cr multilayers

The magnetoresistance (MR) in Fe/Cr magnetic multilayers (MML) has been measured under high pressure up to 2.5 GPa. It is found that the spin-dependent scattering plays an important role in the pressure dependence of MR ratio. In the present work, for [Fe(20/spl Aring/)/Cr(10 /spl Aring/)]/sub 20/ MML with antiferromagnetic (AF) state, the pressure coefficient of saturation field H/sub s/ is (1/H/sub s/)(/spl part/H/sub s///spl part/P)=3.3/spl times/10/sup -2/ GPa/sup -1/ between 0.1 and 2.5 GPa. We found that the (1/|J|)(/spl part/|J|//spl part/P) for AF-Fe/Cr MML with polycrystalline structure is opposite in sign to that with epitaxial one. For [Fe(20 /spl Aring/)/Cr(22/spl Aring/)]/sub 20/ MML with ferromagnetic (F) state, anisotropic magnetoresistance (AMR) decreases with increasing pressure. It is suggested that the anisotropy constant decreases with weakening spin-orbit interaction at high pressure. At high field, the AMR is easily suppressed by applying pressure while the giant magnetoresistance around H/sub s/ increases slightly with increasing pressure for polycrystalline Fe/Cr MML.     

A robust SMES controller design for stabilization of inter-area oscillations based on wide area synchronized phasor measurements

This paper proposes a robust power controller design of superconducting magnetic energy storage (SMES) based on wide area synchronized phasor measurement units (PMUs) for stabilization of inter-area oscillation. The structure of active and reactive power controllers of SMES is the first-order lead/lag compensator. Assuming multiple PMUs are located in an interconnected power system, the steady state phasor data are obtained by applying the small load perturbation. Using the phasor data, the simplified oscillation model (SOM) included with SMES power controllers can be identified and applied to estimate the dominant inter-area oscillation modes. In the robust control design, unstructured system uncertainties such as various operating conditions, system parameters variation, etc., are represented by the inverse additive perturbation and included in the SOM. To enhance the system robust stability margin, the optimization of SMES control parameters is solved by genetic algorithm in the SOM. Simulation studies in the West Japan 6-machine power system confirm that the robustness of the proposed SMES is much superior to the conventional SMES against various operating conditions and fault locations.     

Back gate effects on threshold voltage sensitivity to SOI thicknessin fully-depleted SOI MOSFETs

The dependence of threshold voltage on silicon-on-insulator (SOI) thickness is studied on fully-depleted SOI MOSFETs, and, for this purpose, back-gate oxide thickness and back gate voltage are varied. When the back gate oxide is thinner than the critical thickness dependent on the back gate voltage, the threshold voltage has a minimum in cases where the SOI film thickness is decreased, because of capacitive coupling between the SOI layer and the back gate. This fact suggests that threshold voltage fluctuations due to SOI thickness variations are reduced by controlling the back gate voltage and thinning the back gate oxide     

PIV measurements of flows around the wind turbines with a flanged-diffuser shroud

The wind turbines with a flanged-diffuser shroud -so called ＂wind lens turbine＂- are developed as one of high performance wind turbines by Ohya et al. In order to investigate the flow characteristics and flow acceleration, the paper presents the flow velocity measurements of a long-type and a compact-type wind turbines with a flanged-diffuser shroud by particle image velocimetry. In the case of the long type wind turbine, the velocity vec- tors of the inner flow field of the diffuser for turbine blades rotating and no blades rotating are presented at Rey- nolds number, 0.9x105. Fur~thermore the flow fields between with and without rotating are compared. Through the PIV measurement results, one can realize that the turbine blades rotating affects as suppress the disturbance and the flow separation near the inner wall of the diffuser. The time average velocity vectors are made on the av- erage of the instantaneous velocity data. There are two large vortices in downstream region of the diffuser. One vortex behind the flange acts as suck in wind to the diffuser and raise the inlet flow velocity. Another large vortex appears in downstream. It might be act as blockage vortex of main flow. The large blockage vortex is not clear in the instantaneous velocity vectors, however it exists clearly in the time average flow field. The flow field around the wind turbine with a compact-type flanged-diffuser shroud is also investigated. The flow pattern behind the flange of the compact-type turbine is the same as the long-type one. It means that the effect of flow acceleration is caused by the unsteady vortices behind the flange. The comparison with CFD and PIV results of meridional time-average streamlines after the compact-type diffuser is also presented.     

Crosstalk between Metabolic Disorders and Immune Cells

Metabolic syndrome results from multiple risk factors that arise from insulin resistance induced by abnormal fat deposition. Chronic inflammation owing to obesity primarily results from the recruitment of pro-inflammatory M1 macrophages into the adipose tissue stroma, as the adipocytes within become hypertrophied. During obesity-induced inflammation in adipose tissue, pro-inflammatory cytokines are produced by macrophages and recruit further pro-inflammatory immune cells into the adipose tissue to boost the immune response. Here, we provide an overview of the biology of macrophages in adipose tissue and the relationship between other immune cells, such as CD4+ T cells, natural killer cells, and innate lymphoid cells, and obesity and type 2 diabetes. Finally, we discuss the link between the human pathology and immune response and metabolism and further highlight potential therapeutic targets for the treatment of metabolic disorders.     

Evaluation of scenario-based modularization for lifecycle design

Although modular design is a key technique for lifecycle design, effectiveness of modularity on environmental consciousness is not clear. This paper proposes an evaluation method of modular product from the viewpoint of resource efficiency. First, a product is modularized by applying our modularization method based on its lifecycle scenario. Second, the modular structure is evaluated by assuming that each module goes through preferable lifecycle paths (e.g., upgrading, reuse, and closed-loop recycling) designated by the scenario but unmodularized components go through unpreferable paths. This paper also illustrates a case study in order to discuss effectiveness of the proposed method.     

Improving the Hydrogen Reaction Kinetics of Complex Hydrides

Alanates, borohydrides, and amides are complex hydrides with high concentration hydrogen that have been actively investigated for materials‐based hydrogen storage on‐board polymer electrolyte membrane fuel cell (PEMFC) vehicle applications. The major challenge is to release hydrogen at fuel cell working temperature range at fast enough rate without simultaneous desorption of fuel cell poisoning impurities. We review recent progress in hydrogen reaction mechanism and schemes for complex hydride hydrogen storage.     

Application of a quick-freezing and deep-etching method to pathological diagnosis: a case of elastofibroma

A case of elastofibroma in a middle-aged Japanese woman was examined by the quick-freezing and deep-etching (QF-DE) method, as well as by immunohistochemistry and conventional electron microscopy. The slowly growing tumor developed at the right scapular region and was composed of fibrous connective tissue with unique elastic materials called elastofibroma fibers. A normal elastic fiber consists of a central core and peripheral zone, in which the latter has small aggregates of 10 nm microfibrils. By the QF-DE method, globular structures consisting of numerous fibrils (5-20 nm in width) were observed between the collagen bundles. We could confirm that they were microfibril-rich peripheral zones of elastofibroma fibers by comparing the replica membrane and conventional electron microscopy. One of the characteristics of elastofibroma fibers is that they are assumed to contain numerous microfibrils. Immunohistochemically, spindle tumor cells showed positive immunoreaction for vimentin, whereas alpha-smooth muscle actin, desmin, S-100 protein and CD34 showed negative immunoreaction. By conventional electron microscopy, the tumor cell had thin cytoplasmic processes, pinocytotic vesicles and prominent rough endoplasmic reticulum. Abundant intracytoplasmic filaments were observed in some tumor cells. Thick lamina-like structures along with their inner nuclear membrane were often observed in the tumor cell nuclei. The whole image of the tumor cell was considered to be a periosteal-derived cell, which would produce numerous microfibrils in the peripheral zone of elastofibroma fibers. This study indicated that the QF-DE method could be applied to the pathological diagnosis and analysis of pathomechanism, even for surgical specimens obtained from a patient.