arXiv subjects

Keita Saito

Publications and source records attributed to Keita Saito.

6 recordsLinked to original sources

arXiv2026

Development of an early warning method incorporating pre-supernova neutrino light curves

Massive stars ($M>8\mathrm{M_\odot}$) emit neutrinos known as pre-supernova (pre-SN) neutrinos through thermal and nuclear interactions for cooling the stellar core during the final stage of stellar evolution. Real-time monitoring of their pre-SN neutrino interaction rate offers a crucial opportunity to issue an early warning to a core-collapse supernova. Some neutrino detectors, including KamLAND and Super-Kamiokande already operate pre-SN alarm systems based on a statistically significant excess of the observed event rate over the expected background. To improve alarm sensitivity, an alarm method which incorporates the time evolution of the observed pre-SN neutrino event rate was proposed in A. Sheshukov et al. (2021). We evaluate the performance of the light-curve likelihood approach under realistic KamLAND and SK operating conditions, including realistic background rates, global false-alarm-rate calibration, and the combined alarm. The results demonstrate a significant improvement in the alarm time and distance compared to the conventional rate-only method, while maintaining the same false alarm rate.

astro-ph.HE↗

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

Direct measurement of osmotic pressure and interparticle interactions in colloidal dispersions

Colloidal dispersions are widely found in systems ranging from natural environments to industrial materials. Their macroscopic properties such as viscosity and light scattering depend on their dispersibility, which is characterized by interparticle interactions. Osmotic pressure is induced in a solution with a concentration gradient, in which dispersity is one of the major factors governing the behavior of solutes. Thus, examining the relationship between the interparticle interactions and osmotic pressure may reveal colloidal dispersive properties. Although measuring the osmotic pressure is useful to understand dispersion systems, osmotic pressure is usually extremely low, and only limited experimental methods are available. In this study, we demonstrate that both osmotic pressure and interparticle interactions can be measured within the same experimental system, an optical tweezer system. The directly measured pressure is consistent with both the Brownian dynamics simulation and theoretical results based on the effective hard-sphere model, both of which were calculated using the interparticle interactions directly measured in the experiment. This agreement demonstrates the applicability of the proposed technique for investigating dispersive properties based on particle-level interactions. The proposed technique enables bottom-up design of colloidal materials through particle-level modifications.

cond-mat.soft↗

arXiv2025

Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment

Large language models (LLMs) are increasingly deployed in real-world systems, making it critical to understand their vulnerabilities. While data poisoning attacks during RLHF/DPO alignment have been studied empirically, their theoretical foundations remain unclear. We investigate the minimum-cost poisoning attack required to steer an LLM's policy toward an attacker's target by flipping preference labels during RLHF/DPO, without altering the compared outputs. We formulate this as a convex optimization problem with linear constraints, deriving lower and upper bounds on the minimum attack cost. As a byproduct of this theoretical analysis, we show that any existing label-flipping attack can be post-processed via our proposed method to reduce the number of label flips required while preserving the intended poisoning effect. Empirical results demonstrate that this cost-minimization post-processing can significantly reduce poisoning costs over baselines, particularly when the reward model's feature dimension is small relative to the dataset size. These findings highlight fundamental vulnerabilities in RLHF/DPO pipelines and provide tools to evaluate their robustness against low-cost poisoning attacks.

cs.LG↗

arXiv2024

Combined Pre-Supernova Alert System with Kamland and Super-Kamiokande

Preceding a core-collapse supernova, various processes produce an increasing amount of neutrinos of all flavors characterized by mounting energies from the interior of massive stars. Among them, the electron antineutrinos are potentially detectable by terrestrial neutrino experiments such as KamLAND and Super-Kamiokande via inverse beta decay interactions. Once these pre-supernova neutrinos are observed, an early warning of the upcoming core-collapse supernova can be provided. In light of this, KamLAND and Super-Kamiokande, both located in the Kamioka mine in Japan, have been monitoring pre-supernova neutrinos since 2015 and 2021, respectively. Recently, we performed a joint study between KamLAND and Super-Kamiokande on pre-supernova neutrino detection. A pre-supernova alert system combining the KamLAND detector and the Super-Kamiokande detector was developed and put into operation, which can provide a supernova alert to the astrophysics community. Fully leveraging the complementary properties of these two detectors, the combined alert is expected to resolve a pre-supernova neutrino signal from a 15 M$_{\odot}$ star within 510 pc of the Earth, at a significance level corresponding to a false alarm rate of no more than 1 per century. For a Betelgeuse-like model with optimistic parameters, it can provide early warnings up to 12 hours in advance.

hep-ex↗

arXiv2023

Verbosity Bias in Preference Labeling by Large Language Models

In recent years, Large Language Models (LLMs) have witnessed a remarkable surge in prevalence, altering the landscape of natural language processing and machine learning. One key factor in improving the performance of LLMs is alignment with humans achieved with Reinforcement Learning from Human Feedback (RLHF), as for many LLMs such as GPT-4, Bard, etc. In addition, recent studies are investigating the replacement of human feedback with feedback from other LLMs named Reinforcement Learning from AI Feedback (RLAIF). We examine the biases that come along with evaluating LLMs with other LLMs and take a closer look into verbosity bias -- a bias where LLMs sometimes prefer more verbose answers even if they have similar qualities. We see that in our problem setting, GPT-4 prefers longer answers more than humans. We also propose a metric to measure this bias.

cs.CL↗