arXiv subjects

Yusuke Ikeda

Publications and source records attributed to Yusuke Ikeda.

5 recordsLinked to original sources

arXiv2026

Sound Field Interpolation Using Physics-Informed Extreme Learning Machine with Pre-Training

Numerous machine learning-based sound field interpolation methods have been proposed. In particular, physics-informed neural networks (PINNs) can accurately interpolate sound fields from a small number of microphones. However, their high computational cost and long training time pose practical challenges for applications requiring real-time processing or online learning. To address this, we propose a hybrid framework that combines PINN-based pre-training with a physics-informed extreme learning machine (PIELM) tailored for acoustic fields. By replacing iterative PINN fine-tuning for each target sound field with closed-form output-layer adaptation using hidden-layer weights pre-trained by PINN, the proposed method efficiently interpolates unknown sound fields from limited observations. Simulation results under simplified one-dimensional free-field conditions demonstrate that, given a pre-trained model, the proposed method achieves interpolation accuracy comparable to that of PINN-based fine-tuning while reducing the adaptation time by more than three orders of magnitude.

eess.AS↗

arXiv2026

Hydrogenation-induced gigantic resistance decrease of palladium films deposited by high pressure magnetron sputtering

We demonstrate a pronounced decrease in the electrical resistance of highly disordered palladium (Pd) films deposited under a high working Ar pressure using a compact film coating system. The resulting resistance change ratio of up to $1/335$ is predominant among those reported previously. Film characterization suggests two primary mechanisms responsible for this significant resistance reduction: atomic force microscopy observation indicates improved electrical contacts among Pd grains, and X-ray diffraction measurement demonstrates hydrogenation-induced crystallization of Pd. These findings offer a simple scheme to enhance hydrogen sensor performance and can contribute to a more comprehensive understanding of the hydrogenation process in Pd.

cond-mat.mtrl-sci↗

arXiv2025

Spatial Interpolation of Room Impulse Responses based on Deeper Physics-Informed Neural Networks with Residual Connections

The room impulse response (RIR) characterizes sound propagation in a room from a loudspeaker to a microphone under the linear time-invariant assumption. Estimating RIRs from a limited number of measurement points is crucial for sound propagation analysis and visualization. Physics-informed neural networks (PINNs) have recently been introduced for accurate RIR estimation by embedding governing physical laws into deep learning models; however, the role of network depth has not been systematically investigated. In this study, we developed a deeper PINN architecture with residual connections and analyzed how network depth affects estimation performance. We further compared activation functions, including tanh and sinusoidal activations. Our results indicate that the residual PINN with sinusoidal activations achieves the highest accuracy for both interpolation and extrapolation of RIRs. Moreover, the proposed architecture enables stable training as the depth increases and yields notable improvements in estimating reflection components. These results provide practical guidelines for designing deep and stable PINNs for acoustic-inverse problems.

eess.AS↗

arXiv2016

Thermoelectric transport in the layered Ca$_3$Co$_{4-x}$Rh$_x$O$_9$ single crystals

We have examined an isovalent Rh substitution effect on the transport properties of the thermoelectric oxide Ca$_3$Co$_{4}$O$_9$ using single-crystalline form. With increasing Rh content $x$, both the electrical resistivity and the Seebeck coefficient change systematically up to $x=0.6$ for Ca$_3$Co$_{4-x}$Rh$_{x}$O$_9$ samples. In the Fermi-liquid regime where the resistivity behaves as $ρ=ρ_0+AT^2$ around 120 K, the $A$ value decreases with increasing Rh content, indicating that the correlation effect is weakened by Rh $4d$ electrons with extended orbitals. We find that, in contrast to such a weak correlation effect observed in the resistivity of Rh-substituted samples, the low-temperature Seebeck coefficient is increased with increasing Rh content, which is explained with a possible enhancement of a pseudogap associated with the short-range order of spin density wave. In high-temperature range above room temperature, we show that the resistivity is largely suppressed by Rh substitution while the Seebeck coefficient becomes almost temperature-independent, leading to a significant improvement of the power factor in Rh-substituted samples. This result is also discussed in terms of the differences in the orbital size and the associated spin state between Co $3d$ and Rh $4d$ electrons.

cond-mat.str-el↗

arXiv2013

Deterministic generation of Gaussian pure state in quasi-local dissipative system

This paper shows that an arbitrary Gaussian pure state can be deterministically generated in a dissipative open system that has quasi-local interactions between the subsystems and couples to surrounding environment in a local manner. A quasi-local interaction, which means that the interaction occurs among only a few subsystems, is a crucial requirement for practical engineering of a dissipative system. The key idea is that, first, an auxiliary system having local interaction with environment is prepared, and then that auxiliary system is coupled to the underlying target system via a set of two-body Hamiltonians in such a way that a desired pure state is generated. Moreover, we show that, even with a simple single-mode auxiliary system, deterministic generation of an arbitrary approximate Gaussian cluster state is possible, by devising an appropriate switching scheme. Lastly, we discuss in a specific example how much a dissipation-induced pure Gaussian state can be perturbed by decoherence and parameter error.

quant-ph↗