Ainu Internal Reconstruction: on the Origin and Typological Context of the Affiliative Form

This paper discusses the history and prehistory of the so-called affiliative form in Ainu, also known as the “possessive” or “concrete” form of nominals. In earlier research, this form has been understood as belonging to the sphere of suffixal morphology, complicated by the impact of vowel harmony and/or nominal classes. This paper shows, however, that the marking of the affiliative form actually involves a trace of the original stem-final vowel otherwise lost in the language, followed by a recently grammaticalized separate particle. This means that Ainu must have undergone a general process of final vowel loss, which has substantially changed its morpheme structure and ultimately caused the morphophonological alternations connected with the affiliative form. This conclusion potentially opens up the way towards a more comprehensive internal reconstruction of Ainu.

Cost-effective Indoor Localization for Autonomous Robots using Kinect and WiFi Sensors

Indoor localization has been considered to be the most fundamental problem when it comes to providing a robot with autonomous capabilities. Although many algorithms and sensors have been proposed, none have proven to work perfectly under all situations. Also, in order to improve the localization quality, some approaches use expensive devices either mounted on the robots or attached to the environment that don't naturally belong to human environments. This paper presents a novel approach that combines the benefits of two localization techniques, WiFi and Kinect, into a single algorithm using low-cost sensors. It uses separate Particle Filters (PFs). The WiFi PF gives the global location of the robot using signals of Access Point devices from different parts of the environment while it bounds particles of the Kinect PF, which determines the robot's pose locally. Our algorithm also tackles the Initialization/Kidnapped Robot Problem by detecting divergence on WiFi signals, which starts a localization recovering process. Furthermore, new methods for WiFi mapping and localization are introduced.

Faster RooFitting: Automated parallel calculation of collaborative statistical models

RooFit [1, 2] is the main statistical modeling and fitting package used to extract physical parameters from reduced particle collision data, e.g. the Higgs boson experiments at the LHC [3, 4]. RooFit aims to separate particle physics model building and fitting (the users’ goals) from their technical implementation and optimization in the back-end. In this paper, we outline our efforts to further optimize this back-end by automatically running parts of user models in parallel on multi-core machines. A major challenge is that RooFit allows users to define many different types of models, with different types of computational bottlenecks. Our automatic parallelization framework must then be flexible, while still reducing run time by at least an order of magnitude, preferably more. We have performed extensive benchmarks and identified at least three bottlenecks that will benefit from parallelization. We designed a parallelization framework that allows us to parallelize likelihood minimization with high performance by splitting over partial derivatives in the minimizer. The basis of the framework is a task queue approach. Preliminary results show speed-ups of factor 2 to 20, depending on the exact model and parallelization strategy.

Asteroid Visited by Mission Spews Rocks into Space

Mission scientists observed 11 separate particle ejection events in a 1-month period. They are still trying to figure out what could be causing the particle plumes.

Using apparatus with vortex layer of ferromagnetic particles for production of unburnt synthanite

Here is offered a method of production of unburnt synthanite from natural minerals using apparatus with vortex layer of ferromagnetic particles. There was fulfilled comparative quality analyses of anhydrite binding substances, produced by anhydrite rock milling together with complex solidifying activator in traditional ball mills, and substances produced by our method. It was stated that achieved rising of the hydration activity of synthanite by collective milling of the raw material components was a result of ferromagnetic particles hindered impact under magnetic field influence. The reason for such activity is not only increase in specific surface area of the milling material, but changes in anhydrite structure at the crystalline grid level, preserving a lot of defects on the separate particle surface.

A HIGH-PERFORMANCE METHOD FOR SIMULATING SURFACE RAINFALL-RUNOFF DYNAMICS USING PARTICLE SYSTEM

The simulation of rainfall-runoff process is essential for disaster emergency and sustainable development. One common disadvantage of the existing conceptual hydrological models is that they are highly dependent upon specific spatial-temporal contexts. Meanwhile, due to the inter-dependence of adjacent flow paths, it is still difficult for the RS or GIS supported distributed hydrological models to achieve high-performance application in real world applications. As an attempt to improve the performance efficiencies of those models, this study presents a high-performance rainfall-runoff simulating framework based on the flow path network and a separate particle system. The vector-based flow path lines are topologically linked to constrain the movements of independent rain drop particles. A separate particle system, representing surface runoff, is involved to model the precipitation process and simulate surface flow dynamics. The trajectory of each particle is constrained by the flow path network and can be tracked by concurrent processors in a parallel cluster system. The result of speedup experiment shows that the proposed framework can significantly improve the simulating performance just by adding independent processors. By separating the catchment elements and the accumulated water, this study provides an extensible solution for improving the existing distributed hydrological models. Further, a parallel modeling and simulating platform needs to be developed and validate to be applied in monitoring real world hydrologic processes.

A HIGH-PERFORMANCE METHOD FOR SIMULATING SURFACE RAINFALL-RUNOFF DYNAMICS USING PARTICLE SYSTEM

The simulation of rainfall-runoff process is essential for disaster emergency and sustainable development. One common disadvantage of the existing conceptual hydrological models is that they are highly dependent upon specific spatial-temporal contexts. Meanwhile, due to the inter-dependence of adjacent flow paths, it is still difficult for the RS or GIS supported distributed hydrological models to achieve high-performance application in real world applications. As an attempt to improve the performance efficiencies of those models, this study presents a high-performance rainfall-runoff simulating framework based on the flow path network and a separate particle system. The vector-based flow path lines are topologically linked to constrain the movements of independent rain drop particles. A separate particle system, representing surface runoff, is involved to model the precipitation process and simulate surface flow dynamics. The trajectory of each particle is constrained by the flow path network and can be tracked by concurrent processors in a parallel cluster system. The result of speedup experiment shows that the proposed framework can significantly improve the simulating performance just by adding independent processors. By separating the catchment elements and the accumulated water, this study provides an extensible solution for improving the existing distributed hydrological models. Further, a parallel modeling and simulating platform needs to be developed and validate to be applied in monitoring real world hydrologic processes.

Theory of segmented particle filters

We study the asymptotic behavior of a new particle filter approach for the estimation of hidden Markov models. In particular, we develop an algorithm where the latent-state sequence is segmented into multiple shorter portions, with an estimation technique based upon a separate particle filter in each portion. The partitioning facilitates the use of parallel processing, which reduces the wall-clock computational time. Based upon this approach, we introduce new estimators of the latent states and likelihood which have similar or better variance properties compared to estimators derived from standard particle filters. We show that the likelihood function estimator is unbiased, and show asymptotic normality of the underlying estimators.

Coupled Lattice Boltzmann and Meshless Simulation of Natural Convection in the Presence of Volumetric Radiation

In this work, the coupled lattice Boltzmann and direct collocation meshless (LB–DCM) method is introduced to solve the natural convection in the presence of volumetric radiation in irregular geometries. LB–DCM is a hybrid approach based on a common multiscale Boltzmann-type model. Separate particle distribution functions with multirelaxation time (MRT) and lattice Bhatnagar–Gross–Krook (LBGK) models are used to calculate the flow field and the thermal field, respectively. The radiation transfer equation is computed using the meshless method with moving least-squares (MLS) approximation. The LB–DCM code is first validated by the case of coupled convection–radiation flows in a square cavity. Comparisons show that this combined method is accurate and efficient. Then, the coupled convective and radiative heat transfer in two complex geometries are simulated at various parameters, such as eccentricity, Rayleigh number, and convection–radiation parameter. Numerical results show that the LB–DCM combination is a potential technique for the multifield coupling models, especially with the curved boundary.