arXiv subjects

Hiroki Nakamura

Publications and source records attributed to Hiroki Nakamura.

At least 19 recordsLinked to original sources

arXiv2026

Spin Neural Network Potential for Magnetic Phase Transitions in Uranium Dioxide

Uranium dioxide (UO2) is a prototypical nuclear fuel material, yet predicting its thermophysical properties across a wide temperature range remains challenging. One factor contributing to this difficulty is the complex magnetic ordering at low temperatures, where spin-orbit coupling produces strong coupling between spin and lattice degrees of freedom. Direct DFT simulations of magnetic phase transitions at finite temperatures are computationally prohibitive. Here, we develop a spin neural network potential (SpinNNP) that explicitly incorporates spin degrees of freedom together with spin-orbit coupling to describe the magnetic states of UO2.Reference datasets were generated using magnetic constrained DFT+U calculations with spin-orbit coupling, covering a wide range of non-collinear spin configurations. The SpinNNP accurately reproduces DFT energies, atomic forces, spin forces, and lattice constants. Machine learning molecular dynamics simulations with spin dynamics successfully capture the antiferromagnetic-paramagnetic transition. Although the predicted magnetic ground state differs from experiment due to known limitations of the underlying DFT description, the transition temperature obtained is of the correct order of magnitude compared with experiment. These results demonstrate that machine-learning potentials can enable large-scale spin-lattice simulations of actinide oxides and provide a practical route toward predictive modeling of complex magnetic materials.

cond-mat.mtrl-sci↗

arXiv2026

EasyControlEdge: A Foundation-Model Fine-Tuning for Edge Detection

We propose EasyControlEdge, adapting an image-generation foundation model to edge detection. In real-world edge detection (e.g., floor-plan walls, satellite roads/buildings, and medical organ boundaries), crispness and data efficiency are crucial, yet producing crisp raw edge maps with limited training samples remains challenging. Although image-generation foundation models perform well on many downstream tasks, their pretrained priors for data-efficient transfer and iterative refinement for high-frequency detail preservation remain underexploited for edge detection. To enable crisp and data-efficient edge detection using these capabilities, we introduce an edge-specialized adaptation of image-generation foundation models. To better specialize the foundation model for edge detection, we incorporate an edge-oriented objective with an efficient pixel-space loss. At inference, we introduce guidance based on unconditional dynamics, enabling a single model to control the edge density through a guidance scale. Experiments on BSDS500, NYUDv2, BIPED, and CubiCasa compare against state-of-the-art methods and show consistent gains, particularly under no-post-processing crispness evaluation and with limited training data.

cs.CV↗

arXiv2025

UniEgoMotion: A Unified Model for Egocentric Motion Reconstruction, Forecasting, and Generation

Egocentric human motion generation and forecasting with scene-context is crucial for enhancing AR/VR experiences, improving human-robot interaction, advancing assistive technologies, and enabling adaptive healthcare solutions by accurately predicting and simulating movement from a first-person perspective. However, existing methods primarily focus on third-person motion synthesis with structured 3D scene contexts, limiting their effectiveness in real-world egocentric settings where limited field of view, frequent occlusions, and dynamic cameras hinder scene perception. To bridge this gap, we introduce Egocentric Motion Generation and Egocentric Motion Forecasting, two novel tasks that utilize first-person images for scene-aware motion synthesis without relying on explicit 3D scene. We propose UniEgoMotion, a unified conditional motion diffusion model with a novel head-centric motion representation tailored for egocentric devices. UniEgoMotion's simple yet effective design supports egocentric motion reconstruction, forecasting, and generation from first-person visual inputs in a unified framework. Unlike previous works that overlook scene semantics, our model effectively extracts image-based scene context to infer plausible 3D motion. To facilitate training, we introduce EE4D-Motion, a large-scale dataset derived from EgoExo4D, augmented with pseudo-ground-truth 3D motion annotations. UniEgoMotion achieves state-of-the-art performance in egocentric motion reconstruction and is the first to generate motion from a single egocentric image. Extensive evaluations demonstrate the effectiveness of our unified framework, setting a new benchmark for egocentric motion modeling and unlocking new possibilities for egocentric applications.

cs.CV↗

arXiv2024

Specific Heat Anomalies and Local Symmetry Breaking in (Anti-)Fluorite Materials: A Machine Learning Molecular Dynamics Study

Understanding the high-temperature properties of materials with (anti-)fluorite structures is crucial for their application in nuclear reactors. In this study, we employ machine learning molecular dynamics (MLMD) simulations to investigate the high-temperature thermal properties of thorium dioxide, which has a fluorite structure, and lithium oxide, which has an anti-fluorite structure. Our results show that MLMD simulations effectively reproduce the reported thermal properties of these materials. A central focus of this work is the analysis of specific heat anomalies in these materials at high temperatures, commonly referred to as Bredig, pre-melting, or $λ$-transitions. We demonstrate that a local order parameter, analogous to those used to describe liquid-liquid transitions in supercooled water and liquid silica, can effectively characterize these specific heat anomalies. The local order parameter identifies two distinct types of defective structures: lattice defect-like and liquid-like local structures. Above the transition temperature, liquid-like local structures predominate, and the sub-lattice character of mobile atoms disappears.

cond-mat.mtrl-sci↗

arXiv2024

Representation Synthesis by Probabilistic Many-Valued Logic Operation in Self-Supervised Learning

In this paper, we propose a new self-supervised learning (SSL) method for representations that enable logic operations. Representation learning has been applied to various tasks, such as image generation and retrieval. The logical controllability of representations is important for these tasks. Although some methods have been shown to enable the intuitive control of representations using natural languages as the inputs, representation control via logic operations between representations has not been demonstrated. Some SSL methods using representation synthesis (e.g., elementwise mean and maximum operations) have been proposed, but the operations performed in these methods do not incorporate logic operations. In this work, we propose a logic-operable self-supervised representation learning method by replacing the existing representation synthesis with the OR operation on the probabilistic extension of many-valued logic. The representations comprise a set of feature-possession degrees, which are truth values indicating the presence or absence of each feature in the image, and realize the logic operations (e.g., OR and AND). Our method can generate a representation that has the features of both representations or only those features common to both representations. In addition, the expression of the ambiguous presence of a feature is realized by indicating the feature-possession degree by the probability distribution of truth values of the many-valued logic. We showed that our method performs competitively in single and multi-label classification tasks compared with prior SSL methods using synthetic representations. Moreover, experiments on image retrieval using MNIST and PascalVOC showed that the representations of our method can be operated by OR and AND operations.

cs.CV↗

arXiv2023

Representation Uncertainty in Self-Supervised Learning as Variational Inference

In this study, a novel self-supervised learning (SSL) method is proposed, which considers SSL in terms of variational inference to learn not only representation but also representation uncertainties. SSL is a method of learning representations without labels by maximizing the similarity between image representations of different augmented views of an image. Meanwhile, variational autoencoder (VAE) is an unsupervised representation learning method that trains a probabilistic generative model with variational inference. Both VAE and SSL can learn representations without labels, but their relationship has not been investigated in the past. Herein, the theoretical relationship between SSL and variational inference has been clarified. Furthermore, a novel method, namely variational inference SimSiam (VI-SimSiam), has been proposed. VI-SimSiam can predict the representation uncertainty by interpreting SimSiam with variational inference and defining the latent space distribution. The present experiments qualitatively show that VI- SimSiam could learn uncertainty by comparing input images and predicted uncertainties. Additionally, we described a relationship between estimated uncertainty and classification accuracy.

cs.CV↗

arXiv2020

Three-Dimensional Euler Fluid Code for Fusion Fuel Ignition and Burning

The document describes a numerical algorithm to simulate plasmas and fluids in the 3 dimensional space by the Euler method, in which the spatial meshes are fixed to the space. The plasmas and fluids move through the spacial Euler mesh boundary. The Euler method can represent a large deformation of the plasmas and fluids. On the other hand, when the plasmas or fluids are compressed to a high density, the spatial resolution should be ensured to describe the density change precisely. The present 3D Euler code is developed to simulate a nuclear fusion fuel ignition and burning. Therefore, the 3D Euler code includes the DT fuel reactions, the alpha particle diffusion, the alpha particle deposition to heat the DT fuel and the DT fuel depletion by the DT reactions, as well as the thermal energy diffusion based on the three-temperature compressible fluid model.

physics.plasm-ph↗

arXiv2016

Multi-band Eilenberger theory of superconductivity: Systematic low-energy projection

We propose the general multi-band quasiclassical Eilenberger theory of superconductivity to describe quasiparticle excitations in inhomogeneous systems. With the use of low-energy projection matrix, the $M$-band quasiclassical Eilenberger equations are systematically obtained from $N$-band Gor'kov equations. Here $M$ is the internal degrees of freedom in the bands crossing the Fermi energy and $N$ is the degree of freedom in a model. Our framework naturally includes inter-band off-diagonal elements of Green's functions, which have usually been neglected in previous multi-band quasiclassical frameworks. The resultant multi-band Eilenberger and Andreev equations are similar to the single-band ones, except for multi-band effects. The multi-band effects can exhibit the non-locality and the anisotropy in the mapped systems. Our framework can be applied to an arbitrary Hamiltonian (e.g. a tight-binding Hamiltonian derived by the first-principle calculation). As examples, we use our framework in various kinds of systems, such as noncentrosymmetric superconductor CePt$_{3}$Si, three-orbital model for Sr$_{2}$RuO$_{4}$, heavy fermion CeCoIn$_{5}$/YbCoIn$_{5}$ superlattice, a topological superconductor with the strong spin-orbit coupling Cu$_{x}$Bi$_{2}$Se$_{3}$, and a surface system on a topological insulator.

cond-mat.supr-con↗

arXiv2015

First-principle study of antimony doping effects on the iron-based superconductor CaFe(Sb$_{x}$As$_{1-x}$)$_{2}$

We study antimony doping effects in the iron-based superconductor CaFe(Sb$_{x}$As$_{1-x}$)$_{2}$ by using the first-principle calculation. The calculations reveal that the substitution of the doped antimony atom into As of the chainlike As layers is more stable than that in FeAs layers. This prediction can be checked by experiments. Our results suggest that doping homologous elements into the chainlike As layers existing only in novel 112 system is responsible for rising up the critical temperature. We discuss antimony doping effects on the electronic structure. It is found that the calculated band structures with and without the antimony doping are similar to each other within our framework.

cond-mat.supr-con↗

arXiv2015

Surface states around a vortex in topological superconductors: Intersection of a surface and a vortex

We numerically show that the zero-energy Majorana surface states are suppressed around a vortex in the three-dimensional topological superconductors such as Cu$_{x}$Bi$_{2}$Se$_{3}$ and Sn$_{1-x}$In$_{x}$Te. On the other hand, the zero-energy Majorana bound states along the vortex line are robust against cut by the surface. The suppression of the surface bound states is similar to that with a magnetic impurity on the surface of the topological insulator. The suppression of the surface-bound states around a vortex can be observed as the unconventional energy dependence of imaging of the scanning tunneling microscopy/spectroscopy.

cond-mat.supr-con↗

arXiv2014

Spin-polarized Majorana Bound States inside a Vortex Core in Topological Superconductors

We reveal that Majorana bound states inside the vortex core in an odd-parity topological superconductivity classified as "pseudo-scalar" type in the gap function are distinctly spin-polarized by solving the massive Dirac Bogoliubov-de Gennes (BdG) equation considering the spin-orbit coupling. This result is universal for "Dirac superconductivity" whose rotational degree of freedom is characterized by the total angular momentum J = S + L and in marked contrast to the spin-degeneracy of the core bound states as the consequence of the conventional BdG equation. The spin-polarized vortex core can be easily detected by spin-sensitive probes such as the neutron scattering and other measurements well above the first critical magnetic field H_{c1}.

cond-mat.supr-con↗

arXiv2014

Quasiclassical Treatment and Odd-parity/Triplet Correspondence in Topological Superconductors

We construct a quasiclassical framework for topological superconductors with the strong spin-orbit coupling such as CuxBi2Se3. In the manner of the quasiclassical treatment, decomposing the slowly varying component from the total quasi-particle wave function, the original massive Dirac Bogoliubov-de Gennes (BdG) Hamiltonian derived from the tight-binding model represented by 8 x 8 matrix is reduced to 4 x 4 one. The resultant equations are equivalent to Andreev-type equations of singlet or triplet superconductors, in which the apparent spin-orbit coupling vanishes. Using this formalism, we find a fact that the odd-parity superconductivity in topological superconductors turns to the spin-triplet one. % without the spin-orbit coupling through the quasiclassical treatment. Moreover, in terms of the quasiclassical treatment, we show that the topologically-protected zero-energy states in topological superconductors has the correspondence to the Andreev bound states established in a long history of studies for the unconventional superconductors. This clearly indicates that low-energy non-trivial superconducting properties in the topological superconductors can be analyzed by the established theoretical descriptions on the spin-triplet superconductors.

cond-mat.supr-con↗

arXiv2013

Phonons and Spin Excitations in Fe-Based Superconductor Ca10Pt4As8(Fe1-xPtxAs)10 (x ~ 0.2)

By means of neutron inelastic scattering, magnetic excitations and phonons were measured for a single crystal of slightly overdoped superconductor Ca10Pt4As8(Fe1-xPtxAs)10 (x ~ 0.2) with the transition temperature Tc of ~33 K. Below Tc, magnetic excitation spectra \{chi}"(Q, ω) measured at Q = QM [magnetic Γ points] are gapped, and in the relatively higher ω region, the \{chi}"(QM, ω)-increase was observed with decreasing T, where the maximum of the increase was found at ~18 meV at 3 K (<< Tc). These characteristics are favorable to orbital-fluctuation-mediated superconductivity with the so-called S++symmetry of the order parameter. The energy dependence of δQ-temperature(T) curve of the magnetic excitations seems to have anomalous behavior in rather wide T region above Tc, δQ being the width of the Q scan profile. In the phonon measurements, we observed softening of the in-plane TA mode, which corresponds to the elastic constant C66. This softening seems to start at rather high temperature T, as T is lowered. Additionally, anomalous increase in spectral weights of the TO-phonons at around QM in the region 35 < ω < 40 meV was found even above Tc, as T is lowered from ambient T . Because the spectral weights in this ω region mainly correspond to the in-plane motions of Fe atoms and because orbital fluctuations are expected to be strong at around QM, the result may present clues to investigate a possible coupling between the fluctuations of the orbitals and lattice system.

cond-mat.supr-con↗

arXiv2013

Inhomogeneity Effects in Topological Superconductors

We have constructed a quasiclassical framework on superconductors with strong spin-orbit couplings, applicable to CuxBi2Se3[Y. Nagai, H. Nakamura, and M. Machida: arXiv:1305.3025]. The notable point is that in this framework the Bogoliubov-de Gennes Hamiltonians with suggested odd-parity pairing states turn to quasiclassical ones with usual spin-triplet Cooper pairs. Using this quasiclassical theory, we can investigate inhomogeneity effects such as the phenomena with vortices and surfaces in this superconductors and shed light on the pairing state of topological superconductors. In this paper, we apply the quasiclassical framework to the surface bound states with the Dirac-cone energy dispersion originated from the topological invariant in the parent compound Bi2Se3 in order to investigate the robustness of these bound states under the superconducting order parameter. The odd-parity gap functions can not open on the Dirac-cone-dispersion band in the Cu-doped Bi$_{2}$Se$_{3}$ superconductor. We show that the massless Dirac quasiparticles originated from the normal-state topological invariant and the Majorana quasiparticles coexist with each other on the surface in the odd-parity topological superconductivity. Inhomogeneity effects can be easily investigated with the use of our quasiclassical framework in topological superconductors.

cond-mat.supr-con↗

arXiv2012

Rotational Isotropy Breaking as Proof for Spin-polarized Cooper Pairs in the Topological Superconductor CuxBi2Se3

In a promising candidate of topological superconductors, CuxBi2Se3, we propose a way to exclusively determine the pairing symmetry. The proposal suggests that the angle dependence of the thermal conductivity in the basal ab-plane shows a distinct strong anisotropy only when the pairing symmetry is an odd-parity spin-polarized triplet below the superconducting transition temperature (Tc). Such striking isotropy breaking below Tc is explicitly involved in Dirac formalism for superconductors, in which the spin-orbit coupling is essential. We classify possible gap functions based on the Dirac formalism and clarify an origin of the isotropy breaking.

cond-mat.supr-con↗

arXiv2011

Superconducting gap function in the organic superconductor (TMTSF)2ClO4 with anion ordering; First-principles calculations and quasiclassical analysis for angle-resolved heat capacity

We calculate angle-dependent heat capacity in a low magnetic field range on the basis of Kramer-Pesch approximation together with an electronic structure obtained by first-principles calculations to determine a superconducting gap function of (TMTSF)2ClO4 through its comparisons with experiments. The present comparative studies reveal that a nodal d-wave gap function consistently explains the experimental results for (TMTSF)2ClO4. Especially, it is emphasized that the observed unusual axis-asymmetry of the angle-dependence eliminates the possibility of s-wave and node-less d-wave functions. It is also found that the directional ordering of ClO4 anions does not have any significant effects on the Fermi surface structure contrary to the previous modelings since the two Fermi surfaces obtained by the band calculations almost cross within the present full accuracy in first-principles calculations.

cond-mat.supr-con↗

arXiv2011

First-principles Studies for the Hydrogen Doping Effects on Iron-based Superconductors

We study hydrogen doping effects in an iron-based superconductor LaFeAsO_(1-y) by using the first-principles calculation and explore the reason why the superconducting transition temperature is remarkably enhanced by the hydrogen doping. The present calculations reveal that a hydrogen cation stably locating close to an iron atom attracts a negatively-charged FeAs layer and results in structural distortion favorable for further high temperature transition. In fact, the lattice constant a averaged over the employed supercell shrinks and then the averaged As-Fe-As angle approaches 109.74 degrees with increasing the hydrogen doping amount. Moreover, the calculations clarify electron doping effects of the solute hydrogen and resultant Fermi-level shift. These insights are useful for design of high transition-temperature iron-based superconductors.

cond-mat.supr-con↗

arXiv2010

Ambegaokar-Baratoff relations of Josephson critical current in heterojunctions with multi-gap superconductors

An extension of the Ambegaokar-Baratoff relation to a superconductor-insulator-superconductor (SIS) Josephson junction with multiple tunneling channels is derived. Appling the resultant relation to a SIS Josephson junction formed by an iron-based (five-band) and a single-band Bardeen-Cooper-Schrieffer (BCS) type superconductors, a theoretical bound of the Josephson critical current ($I_{\rm c}$) multiplied by the resistance of the junction ($R_{\rm n}$) is given. We reveal that such a bound is useful for identifying the pairing symmetry of iron-pnictide superconductors. One finds that if a measured value of $I_{\rm c}R_{\rm n}$ is smaller than the bound then the symmetry is $\pm s$-wave, and otherwise $s$-wave without any sign changes. In addition, we stress that temperature dependence of $I_{\rm c}R_{\rm n}$ is sensitive to the difference of the gap functions from the BCS type gap formula in the above heterojunction.

cond-mat.supr-con↗