A neural network model of the olfactory system of mice: computer simulation of the attention behavior of mice for some components in an odor

Recently, it was observed that mice could identify an odor by paying attention to only a few of its components. Further, it has been reported that each individual is attracted to different components of an odor. This behavior is referred to as “attention”; however, its mechanism has yet to be completely elucidated. In this paper, we first propose a novel artificial neural network model based on the biological structure of an olfactory system. Then a series of computer simulations of odorant discrimination are performed to evaluate the attention ability of the proposed model. Finally, we changed the connective weights between the neurons to simulate individual differences. The simulation results indicate that the inhibitory connections from the piriform cortex to the olfactory bulb may contribute to the individual differences observed in the behavioral experiment. This work was presented in part at the 12th International Symposium on Artificial Life and Robotics, Oita, Japan, January 25–27, 2007     

A design study on MOX-fueled small fast reactors for standardization of a small fast nuclear reactor system

A way of development to standardize a small fast nuclear reactor system, which is considered one of the suitable concepts at next generation for satisfying such needs as generality, small dependence on natural resources, safety and non-proliferation, is proposed. This process consists of three steps : the first is to demonstrate the basic system within a short period based on current techniques, the second is to achieve greatly higher economy, and the final is to standardize the commercial system that can economically compete with or overcome current light water reactors. A technical investigation is conducted on the performance of a mixed-oxide (MOX)-fueled small fast reactor with a reflector-driven reactivity control system to satisfy the needs at the first step, considering plenty of accomplishments on the MOX fuel and its advantage for limiting the duration of development to the level required at the stage. The results obtained from a series of neutronic and thermal-hydraulic calculations show the feasibility of a small fast reactor that produces the electric power of about 50MW, achieves about two-year consecutive operation with high safety performance and is greatly flexible for updating the system. A mixed-nitride-fueled core is found to be promising past the first stage.     

The verification tests of the melting conditions for homogenization of metallic LLW at the JAEA

To clarify the operating conditions for the homogenization of metallic low-level radioactive wastes (LLW), melting tests were conducted in the metal melting system of the Advanced Volume Reduction Facilities at the Japan Atomic Energy Agency (JAEA). Simulated metallic LLW with a non-radioactive tracer were melted in the induction furnace of the metal melting system. Regardless of the chemical composition or amounts of the simulated wastes, the tracer distribution in the solidified products was nearly homogeneous when the simulated wastes were completely melted in the range from 1550 °C to 1638 °C.     

Fabrication and mechanical properties of α-Al2O3/β-Al2O3/Al/Si composites by liquid displacement reaction

Al₂O₃/metal composites were fabricated by heating three kinds of commercial mullite refractories in contact with Al, and their mechanical properties were investigated. Aluminum reacted with the mullite and SiO₂-glass constituting the mullite refractories and changed them into -Al₂O₃ and Si. Simultaneously, -Al₂O₃ was formed by the reactions among -Al₂O₃ and sodium and potassium oxides in the glassy phase. Also, Al penetrated into the -Al₂O₃/-Al₂O₃/Si composite by partly dissolving Si. Finally, the mullite refractories were changed into -Al₂O₃/-Al₂O₃/Al/Si composites. The phase contents, microstructures and mechanical properties of the resulting composites varied with the composition of the refractories. The content of -Al₂O₃ in the composite was lowest at the lowest Na₂O and K₂O contents in the refractories. Silicon in the composite had its highest content at the highest SiO₂ content. The composite fabricated from SiO₂/Al₂O₃ (in mol) (SAR)=1.85 consisted of 2–5 m Al₂O₃ grains embedded in metal, but that from SAR=1.05 showed a complicated microstructure with small and large grains. The bending strength of the composites fabricated from the refractories of SAR=1.85, 1.24, and 1.05 were 327, 405 and 421 MPa, respectively. Also, the corresponding fracture toughness values were 5.2, 6.1, and 5.5 MPam , respectively.     

Numerical computation for Rayleigh-Benard convection of water in a magnetic field

The derivation process for the model equation is shown for the natural convection of water (diamagnetic) under both gravity and magnetizing force fields and numerically solved for the Rayleigh-Benard convection in a shallow cylinder heated from below and cooled from above. The cylindrical enclosure was located at two levels in the bore of a super-conducting magnet, where the radial component of the magnetizing force is minimal and its axial component prevails. The cylindrical enclosure was assumed to be located coaxially with the bore of the magnet, and a two-dimensional model equation was presumed. Sample computations were carried out without or with a gravity force for various strengths of Rayleigh number and magnetic induction. When the enclosure was placed above the coil center, where the magnetizing force is opposed to the gravitational force, the average Nusselt number decreased with increasing strength of the magnetic field. When the enclosure was placed below the coil center, where the magnetizing force is parallel to gravity, the average Nusselt number increased above unity even at Ra=1000 and 1500. All of the data agreed favorably with the classical experimental data of Silveston when plotted against the magnetic Rayleigh number proposed by Braithwaite et al.     

A study on the preparation of supported metal oxide catalysts using JRC-reference catalysts. I. Preparation of a molybdena–alumina catalyst. Part 3. Drying process

In the present part of the group study on the preparation of 13 wt% MoO₃/Al₂O₃, the effects of drying processes were investigated on the physicochemical and catalytic properties. Two series of catalysts were prepared by a conventional impregnation technique and by an equilibrium adsorption method using a common extrudate support. XPS and EPMA results demonstrated that the distribution of Mo oxide species in extrudates was strongly affected by drying processes. A rapid drying, in particular at a reduced pressure, was found to induce a strong segregation of Mo oxides on the outer surface of the extrudates, forming a sharp egg shell type distribution of Mo. On the other hand, drying under static conditions produced a moderate egg shell type distribution, suggesting that a slow drying rate is favorable for a homogeneous distribution of Mo. The equilibrium adsorption technique was found to provide considerably flat Mo profiles inside the extrudates except for the utmost surfaces where Mo concentrations increased steeply.     

Modified Rule Ensemble Method for Binary Data and Its Applications

Ensemble learning, which combines multiple base learners to improve statistical prediction accuracy, is frequently used in statistical science and data mining. However, because of their “black box” nature, ensemble learning models are difficult to interpret. A recently proposed rule ensemble method known as RuleFit presents the base learner as a production rule and also generates a measure that influences the response variable. The RuleFit method for binary response applies a squared-error ramp loss function, and base learners are weighted by shrinkage regression using the lasso method. Thus, RuleFit is not constructed by a logistic regression model. Moreover, highly correlated pairs of base learners may be excessively pruned by the lasso method. In this study, we solved the excess pruning problem by constructing RuleFit within a logistic regression framework, weighting the base learners by elastic net. The effectiveness ofour proposed RuleFit model is illustrated through a real data set. In small-scale simulations, this method demonstrated higher predictive performance than the original RuleFit model.