arXiv subjects

Hiroshi Morita

Publications and source records attributed to Hiroshi Morita.

6 recordsLinked to original sources

arXiv2025

Real-Time Moving Flock Detection in Pedestrian Trajectories Using Sequential Deep Learning Models

Understanding collective pedestrian movement is crucial for applications in crowd management, autonomous navigation, and human-robot interaction. This paper investigates the use of sequential deep learning models, including Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, and Transformers, for real-time flock detection in multi-pedestrian trajectories. Our proposed approach consists of a two-stage process: first, a pre-trained binary classification model is used for pairwise trajectory classification, and second, the learned representations are applied to identify multi-agent flocks dynamically. We validate our method using real-world group movement datasets, demonstrating its robustness across varying sequence lengths and diverse movement patterns. Experimental results indicate that our model consistently detects pedestrian flocks with high accuracy and stability, even in dynamic and noisy environments. Furthermore, we extend our approach to identify other forms of collective motion, such as convoys and swarms, paving the way for more comprehensive multi-agent behavior analysis.

cs.LG↗

arXiv2023

Least Squares Estimation of a Quasiconvex Regression Function

We develop a new approach for the estimation of a multivariate function based on the economic axioms of quasiconvexity (and monotonicity). On the computational side, we prove the existence of the quasiconvex constrained least squares estimator (LSE) and provide a characterization of the function space to compute the LSE via a mixed integer quadratic programme. On the theoretical side, we provide finite sample risk bounds for the LSE via a sharp oracle inequality. Our results allow for errors to depend on the covariates and to have only two finite moments. We illustrate the superior performance of the LSE against some competing estimators via simulation. Finally, we use the LSE to estimate the production function for the Japanese plywood industry and the cost function for hospitals across the US.

stat.ME↗

arXiv2021

Molecular-shape- and size-independent power-law dependence of percolation thresholds on radius of gyration in ideal molecular systems

Three-dimensional single-component ideal gas systems composed of model homogeneous rigid molecules in various molecular shapes and sizes are simulated by a molecular Monte Carlo simulation technique. We reveal that percolation thresholds of such single-component systems result in, when the molecular volume is fixed, power-law decreasing functions of the radius of gyration (gyradius) of the molecules. The systems with the same parameter set of the molecular volume and radius of gyration, but in different molecular shapes, show the identical value of the percolation threshold. Moreover, we also reveal that a dimensionless scale-free parameter, which is the ratio between the radius of gyration and real cube root of the molecular volume, uniquely determines the percolation threshold.

cond-mat.soft↗

arXiv2019

Conducting transition analysis of thin films composed of long flexible macromolecules: Percolation study

Simulating percolation and critical phenomena of labelled species inside films composed of single-component linear homogeneous macromolecules using molecular Monte Carlo method in 3 dimensions, we study dependence of these conducting transition and critical phenomena upon both thermal movement, i.e. spontaneous mobility, and extra-molecular topological constraints of the molecules. Systems containing topological constraints and/or composed of immobile particles, e.g. lattice models and chemical gelation, were studied in conventional works on percolation. Coordinates of the randomly distributed particles in the conventional lattice models are limited to discrete lattice points. Moreover, each particle is spatially fixed at the distributed position, which results in a temporally unchanged network structure. Although each polymer in the chemical gels can spontaneously move in the continuous space, the network structure is fixed when cross-linking reaction ends. By contrast to these conventional systems, all the molecules in the present system freely move and spontaneously diffuse in the continuous space. The network structure of the present molecules continues changing dynamically. The percolation and critical phenomena of such dynamic network structures are examined here. We reveal that these phenomena also occur in the present system, and that both the universality class and percolation threshold are independent from the extra-molecular topological constraints.

cond-mat.soft↗

arXiv2019

An axiomatic nonparametric production function estimator: Modeling production in Japan's cardboard industry

We develop a new approach to estimate a production function based on the economic axioms of the Regular Ultra Passum law and convex non-homothetic input isoquants. Central to the development of our estimator is stating the axioms as shape constraints and using shape constrained nonparametric regression methods. We implement this approach using data from the Japanese corrugated cardboard industry from 1997-2007. Using this new approach, we find most productive scale size is a function of the capital-to-labor ratio and the largest firms operate close to the largest most productive scale size associated with a high capital-to-labor ratio. We measure the productivity growth across the panel periods based on the residuals from our axiomatic model. We also decompose productivity into scale, input mix, and unexplained effects to clarify the sources the productivity differences and provide managers guidance to make firms more productive.

stat.AP↗

arXiv2012

Shear induced breakup of droplets in a colloidal dispersion

We present numerical results for the breakup of a pair of colloidal particles enveloped by a droplet under shear flow. The smoothed profile method is used to accurately account for the hydrodynamic interactions between particles due to the host fluid. We observe that the critical capillary number, $Ca_{\rm B}$, at which droplets breakup depends on a velocity ratio, $E$, defined as the ratio of the boundary shift velocity (that restores the droplet shape to a sphere) to the diffusive flux velocity in units of the particle radius $a$. For $E < 10$, $Ca_{B}$ is independent of $E$, as is consistent with the regime studied by Taylor. When $E > 10$, $Ca_{B}$ behaves as $Ca_{\rm B} = 2E^{-1}$, which confirms Karam and Bellinger's hypothesis. As a consequence, droplet break up will occur when the time scale of droplet deformation $\dotγ^{-1}$ is smaller than the diffusive time scale $t_{D} \equiv a^{2}/Lτ$ in units of $a$, where $L$ is the diffusion constant and $τ$ is the 2nd order coefficient of the Ginzburg-Landau type free energy of the binary mixture. We emphasize that the breakup of droplet dispersed particles is not only governed by a balance of forces. We find that velocity competition is one of the important contributing factor.

cond-mat.soft↗