arXiv Science⌕ Search

arXiv subjects

Masato Ito

Publications and source records attributed to Masato Ito.

At least 19 recordsLinked to original sources
arXiv2026

Directional Symmetry Breaking of Spherical Active Colloids by Magnetoviscous Coupling

Harnessing active matter calls for strategies that break the directional symmetry of self-propelled motion without altering the propulsion mechanism itself. Here, we show that magnetically inert spherical active colloids can be steered through the anisotropic viscous response of a ferrofluid under a uniform magnetic field. Self-propelled Janus colloids exhibit robust cross-field motion transverse to the magnetic field, although the applied magnetic field directly controls neither the particles nor their propulsion speed. Quantitative measurements reveal an emergent reorientation torque that grows with both propulsion speed and magnetic field strength. A squirmer model in a magnetoviscous medium captures these observations and shows that the torque arises from the coupling between swimmer-generated flow and anisotropic rotational viscosity. Our findings establish a hydrodynamic basis for converting viscous dissipation into directional symmetry breaking through anisotropic rheology, providing a route to field-controlled material transport by active matter.

cond-mat.soft↗
arXiv2026

AE-LLM: Adaptive Efficiency Optimization for Large Language Models

Large Language Models (LLMs) have achieved remarkable success across diverse applications, yet their deployment remains challenging due to substantial computational costs, memory requirements, and energy consumption. Recent empirical studies have demonstrated that no single efficiency technique is universally optimal; instead, the effectiveness of methods such as efficient attention mechanisms, mixture-of-experts (MoE), parameter-efficient fine-tuning, and quantization varies significantly depending on task characteristics, resource constraints, and model scales. Building upon these insights, we propose AE-LLM, a unified framework that automatically selects and combines optimal efficiency techniques tailored to specific deployment scenarios. Our approach introduces a multi-objective optimization framework that jointly considers accuracy, latency, memory footprint, and energy consumption, while accounting for hardware constraints and task requirements. We develop an efficient search algorithm that explores the combinatorial space of efficiency techniques across architecture, fine-tuning, and inference stages, identifying Pareto-optimal configurations. Extensive experiments across 15 models (0.5B-70B parameters) and 10 diverse tasks demonstrate that AE-LLM achieves an average of $2.8\times$ improvement in efficiency metrics while maintaining competitive accuracy (within 1.2\% of baseline), compared to static efficiency configurations. Furthermore, our framework generalizes effectively to vision-language models, achieving similar efficiency gains. Our contributions provide practitioners with an automated tool for navigating the complex trade-off landscape of LLM efficiency optimization.

cs.LG↗
arXiv2026

Contextual Discrepancy-Aware Contrastive Learning for Robust Medical Time Series Diagnosis in Small-Sample Scenarios

Medical time series data, such as EEG and ECG, are vital for diagnosing neurological and cardiovascular diseases. However, their precise interpretation faces significant challenges due to high annotation costs, leading to data scarcity, and the limitations of traditional contrastive learning in capturing complex temporal patterns. To address these issues, we propose CoDAC (Contextual Discrepancy-Aware Contrastive learning), a novel framework that enhances diagnostic accuracy and generalization, particularly in small-sample settings. CoDAC leverages external healthy data and introduces a Contextual Discrepancy Estimator (CDE), built upon a Transformer-based Autoencoder, to precisely quantify abnormal signals through context-aware anomaly scores. These scores dynamically inform a Dynamic Multi-views Contrastive Framework (DMCF), which adaptively weights different temporal views to focus contrastive learning on diagnostically relevant, discrepant regions. Our encoder combines dilated convolutions with multi-head attention for robust feature extraction. Comprehensive experiments on Alzheimer's Disease EEG, Parkinson's Disease EEG, and Myocardial Infarction ECG datasets demonstrate CoDAC's superior performance across all metrics, consistently outperforming state-of-the-art baselines, especially under low label availability. Ablation studies further validate the critical contributions of CDE and DMCF. CoDAC offers a robust and interpretable solution for medical time series diagnosis, effectively mitigating data scarcity challenges.

cs.LG↗
arXiv2025

XDR-LVLM: An Explainable Vision-Language Large Model for Diabetic Retinopathy Diagnosis

Diabetic Retinopathy (DR) is a major cause of global blindness, necessitating early and accurate diagnosis. While deep learning models have shown promise in DR detection, their black-box nature often hinders clinical adoption due to a lack of transparency and interpretability. To address this, we propose XDR-LVLM (eXplainable Diabetic Retinopathy Diagnosis with LVLM), a novel framework that leverages Vision-Language Large Models (LVLMs) for high-precision DR diagnosis coupled with natural language-based explanations. XDR-LVLM integrates a specialized Medical Vision Encoder, an LVLM Core, and employs Multi-task Prompt Engineering and Multi-stage Fine-tuning to deeply understand pathological features within fundus images and generate comprehensive diagnostic reports. These reports explicitly include DR severity grading, identification of key pathological concepts (e.g., hemorrhages, exudates, microaneurysms), and detailed explanations linking observed features to the diagnosis. Extensive experiments on the Diabetic Retinopathy (DDR) dataset demonstrate that XDR-LVLM achieves state-of-the-art performance, with a Balanced Accuracy of 84.55% and an F1 Score of 79.92% for disease diagnosis, and superior results for concept detection (77.95% BACC, 66.88% F1). Furthermore, human evaluations confirm the high fluency, accuracy, and clinical utility of the generated explanations, showcasing XDR-LVLM's ability to bridge the gap between automated diagnosis and clinical needs by providing robust and interpretable insights.

cs.CV↗
arXiv2018

Lambert W function and hanging chain revisited

In classical physics, calculating the slack of a hanging chain is a problem that has attracted interest. This study aims to solve this problem through experiment and theory. When the length and distance of both the ends of a hanging chain are given, the length of slack can be expressed by a Lambert W function or an irrational function within a certain distance. Herein, a simple observation of the slack is presented. The result obtained is one of the applications of the Lambert W function in the field of physics. Though the shape of a hanging chain is well-known to be a catenary, the calculation of slack possesses richer aspects of mathematical physics.

physics.class-ph↗
arXiv2009

In-situ observation of a soap film catenoid - a simple educational physics experiment

The solution to the Euler-Lagrange equation is an extremal functional.To understand that the functional is stationary at local extrema (maxima or minima), we propose a physics experiment that involves using soap film to form a catenoid. A catenoid is a surface that is formed between two coaxial circular rings and is classified mathematically as a minimal surface.Using soap film, we create catenoids between two rings and characterize the catenoid in-situ while varying distance between rings. The shape of the soap film is very interesting and can be explained using dynamic mechanics. By observing catenoid, physics students can observe local extrema phenomena. We stress that in-situ observation of soap film catenoids is an appropriate physics experiment that combines theory and experimentation.

physics.class-ph↗
arXiv2009

Anisotropic Evolution Driven by Kinetic Term

We present a simple model where anisotropic evolution is driven by kinetic term in extra dimensions. By introducing a canonical or a ghost kinetic term, the possibility of anisotropy is studied.

hep-th↗
arXiv2007

Cosmological solutions for model with a $1/H^{2}$ term

We drive the cosmological solutions of five-dimensional model with $1/H^{2}$ term $(H^{2}\equiv H_{MNPQ}H^{MNPQ})$, where $H_{MNPQ}$ is 4-form field strength. The behaviors of the scale factors and the scalar potential in effective theory are examined.As a consequence, we show that the universe changes from decelerated expansion to accelerated expansion in Einstein frame of the four-dimensional theory.

hep-th↗
arXiv2005

Holographic dark energy model with non-minimal coupling

We find that holographic dark energy model with non-minimally coupled scalar field gives rise to an accelerating universe by choosing Hubble scale as IR cutoff. We show viable range of a non-minimal coupling parameter in the framework of this model.

hep-th↗
arXiv2004

Scalar potential from de Sitter brane in 5D and effective cosmological constant

We derive the scalar potential in zero mode effective action arising from a de Sitter brane embedded in five dimensions with bulk cosmological constant $Λ$. The scalar potential for a scalar field canonically normalized is given by the sum of exponential potentials. In the case of $Λ=0$ and $Λ>0$, we point out that the scalar potential has an unstable local maximum at the origin and exponentially vanishes for large positive scalar field. In the case of $Λ<0$, the scalar potential has an unstable local maximum at the origin and a stable local minimum, it is shown that the positive cosmological constant in brane is reduced by negative potential energy of scalar at minimum.

hep-th↗
arXiv2003

Localized gravity on de Sitter brane in five dimensions

We consider a single brane embedded in five dimensions with vanishing and positive bulk cosmological constant. In this setup, the existence of $dS_{4}$ brane is allowed. We explore the gravitational fluctuations on the brane, and we point out that the usual four-dimensional gravity can be reproduced by a normalizable zero mode and that the continuous massive modes are separated by a mass gap from zero mode. We derive the relation among the cosmological constant in the brane, and the fundamental scale of five dimensions and Planck scale. Finally we show that not zero but tiny observed cosmological constant cannot be realized in this setup.

hep-th↗
arXiv2003

On the solutions to accelerating cosmologies

Motivated by recent accelerating cosmological model, we derive the solutions to vacuum Einstein equation in $(d+1)$-dimensional Minkowski space with $n$-dimensional hyperbolic manifold. The conditions of accelerating expansion are given in such a set up.

hep-th↗
arXiv2003

Casimir Energies due to Matter Fields in $T^{2}$ and $T^{2}/Z_{2}$ Compactifications

We calculate the Casimir energies due to matter fields with various boundary conditions along two compact directions in $T^{2}$ compactification. We discuss whether the Casimir energies generate attractive or repulsive forces. On the theories with extra dimensions, the Casimir energy plays a crucial role in the mechanism for stabilizing the size of extra dimensions. Finally we argue a procedure of the application to $Z_{2}$ orbifold.

hep-ph↗
arXiv2003

Correction terms to Newton law due to induced gravity in AdS background

We calculate small correction terms to gravitational potential on Randall-Sundrum brane with an induced Einstein term. The behaviors of the correction terms depend on the magnitudes of $AdS$ radius $k^{-1}$ and a characteristic length scale $ł$ of model. We represent the gravitational potential for arbitrary $k$ and $ł$ at all distances.

hep-th↗
arXiv2002

Linearized gravity in flat braneworlds with anisotropic brane tension

We study the four-dimensional gravitational fluctuation on anisotropic brane tension embedded in braneworlds with vanishing bulk cosmological constant. In this setup, warp factors have two types (A and B) and we point out that the two types correspond to positive and negative tension brane, respectively. We show that volcano potential in the model of type A has singularity and the usual Newton's law is reproduced by the existence of normalizable zero mode. While, in the case of type B, the effective Planck scale is infinite so that there is no normalizable zero mode.

hep-th↗
arXiv2002

Precocious Gauge Symmetry Breaking in $SU(6) \times SU(2)_R$ Model

In the $SU(6) \times SU(2)_R$ string-inspired model, we evolve the couplings and the masses down from the string scale $M_S$ using the renormalization group equations and minimize the effective potential. This model has the flavor symmetry including the binary dihedral group $\tilde{D}_4$. We show that the scalar mass squared of the gauge non-singlet matter field possibly goes negative slightly below the string scale. As a consequence, the precocious radiative breaking of the gauge symmetry down to the standard model gauge group can occur. In the present model, the large Yukawa coupling which plays an important role in the symmetry breaking is identical with the colored Higgs coupling related to the longevity of the proton.

hep-ph↗
arXiv2002

Cosmological constant and curved 5D geometry

We study the value of cosmological constant in de Sitter brane embedded in five dimensions with positive, vanishing and negative bulk cosmological constant. In the case of negative bulk cosmological constant, we show that not zero but tiny four-dimensional cosmological constant can be realized by tiny deviation from bulk curvature of the Randall-Sundrum model.

hep-ph↗