arXiv subjects

Hidetoshi Kawase

Publications and source records attributed to Hidetoshi Kawase.

6 recordsLinked to original sources

arXiv2026

Cardinality-Stratified Interaction Decomposition for Interpretable Pairwise and Higher-Order Structure in Transactional Basket Data

Transactional basket data can reveal associations among items, but observed co-occurrence conflates item-specific relations with basket-size structure and unmodeled higher-order dependence. We introduce Cardinality-Stratified Interaction Decomposition (CSID), an interpretable framework that decomposes log-odds contrasts stratified by the number of remaining items into item-set-specific and cardinality-common components, without fitting a global joint distribution. CSID uses an information-weighted, gauge-constrained ridge projection to estimate pair and triple components and to diagnose higher-order contributions to pairwise structure. CSID is designed primarily for interpretable decomposition of association structure rather than for full-distribution prediction. In a simulation with zero pair effects, increasingly strong small-basket cardinality potentials drive ordinary Ising couplings spuriously negative, whereas CSID pair estimates remain centered near zero. Detection power rises with the magnitude of planted triple effects, and local deprojection reduces pair-coefficient RMSE from 0.244 to 0.073. Across three grocery datasets, high-information triple components are reproducible over time. In the matched cross-period partial-transfer evaluation, transferred CSID triple components show closer agreement with later-period stratified contrasts than the nodewise-symmetrized cardinality-aware higher-order pseudolikelihood comparator, with gains in weighted Lin's concordance correlation of 0.038--0.122. These results support CSID as an exploratory and interpretable decomposition framework for pairwise and higher-order association structure in transactional data.

cs.LG↗

arXiv2026

Beyond a symmetry-restricted VEV ansatz: transverse stability of an $SU(5)$ special-subgroup vacuum

Extended Higgs sectors can realize multistage breaking chains in grand unified theories (GUTs), and the resulting intermediate gauge phases can be relevant to cosmological defects. Restricting vacuum expectation values (VEVs) to configurations that preserve a chosen subgroup is an efficient way to identify candidate stationary points, while their stability against fluctuations outside this subspace is a separate question. We study this question in a renormalizable $SU(5)$ GUT Higgs sector with an adjoint $24_H$ and a complex symmetric $15_H$, motivated by a special-subgroup realization of Langacker-Pi monopole erasure. On the standard adjoint background, the $15_H$ VEVs that preserve the intermediate gauge group form a two-block family with independent color- and weak-block amplitudes. Portal interactions can make these amplitudes unequal without changing the unbroken gauge group. Two independent mixed quartic invariants become identical when evaluated on this two-block field space, although they contribute differently to color-weak fluctuations. We derive the complete spectrum of these modes and identify two bounded-from-below parameter choices that induce the same potential on the two-block field space but have different curvatures in transverse directions. The common stationary configuration is a local minimum in one case and a saddle with six negative physical modes in the other. Direct Hessian calculations in the full 54-dimensional real scalar field space reproduce the analytic spectrum and verify both the minimum-saddle distinction and locally stable benchmark vacua.

hep-ph↗

arXiv2026

Collective Kernel EFT for Pre-activation ResNets

In finite-width deep neural networks, the empirical kernel $G$ evolves stochastically across layers. We develop a collective kernel effective field theory (EFT) for pre-activation ResNets based on a $G$-only closure hierarchy and diagnose its finite validity window. Exploiting the exact conditional Gaussianity of residual increments, we derive an exact stochastic recursion for $G$. Applying Gaussian approximations systematically yields a continuous-depth ODE system for the mean kernel $K_0$, the kernel covariance $V_4$, and the $1/n$ mean correction $K_{1,\mathrm{EFT}}$, which emerges diagrammatically as a one-loop tadpole correction. Numerically, $K_0$ remains accurate at all depths. However, the $V_4$ equation residual accumulates to an $O(1)$ error at finite time, primarily driven by approximation errors in the $G$-only transport term. Furthermore, $K_{1,\mathrm{EFT}}$ fails due to the breakdown of the source closure, which exhibits a systematic mismatch even at initialization. These findings highlight the limitations of $G$-only state-space reduction and suggest extending the state space to incorporate the sigma-kernel.

cs.LG↗

arXiv2011

Light Neutralino Dark Matter Scenario in Supersymmetric Four-Higgs Doublet Model

We examine a possibility to explain the signals of dark matter direct detection from DAMA and CoGeNT experiments, under the framework of supersymmetric extension of the standard model. For this purpose, we introduce four Higgs doublet fields and assume that one pair of Higgs fields which have very small vacuum expectation values gives sizable effects for annihilation and scattering processes of dark matter. We show that the preferred parameter regions for DAMA and CoGeNT results can be simultaneously explained by supersymmetric four-Higgs doublet model with the parameters consistent with the observed value of dark matter relic abundance. The extra Higgs fields introduced as an explanation of the light dark matter scenario also explain the Wjj anomaly reported by CDF.

hep-ph↗

arXiv2010

Flavor structure of E6 GUT models

It has been pointed out that in E6 grand unified theory with SU(2) family symmetry, the spontaneous CP violation can solve the supersymmetric CP problem. The scenario predicts V_{ub} \sim O(λ^4) rather than O(λ^3) which is naively expected value, because of a cancellation at the leading order. Since the experimental value of V_{ub} is O(λ^4), it must be important to consider the reason and the conditions for the cancellation. In this paper, we give a simple reason for the cancellation and show that in some E6 models such a cancellation requires that the vacuum expectation value (VEV) of the adjoint Higgs does not break U(1)_{B-L}. Note that this direction of the VEV plays an important role in solving the doublet-triplet splitting problem by Dimopoulos-Wilczek mechanism. In this E6 models, the experiments may measure the direction of the adjoint Higgs VEV by measuring the size of V_{ub} \sim O(λ^4).

hep-ph↗

arXiv2010

Gauge-mediated supersymmetry breaking with generalized messenger sector at LHC

We consider the generalized gauge mediated supersymmetry breaking (GMSB) models with the messenger fields which do not form the complete multiplets of SU(5) GUT symmetry. Such a situation may happen in the anomalous U(1) GUT scenario because the mass spectrum of the superheavy particle does not respect SU(5) GUT symmetry, although the success of the gauge coupling unification can be explained. In this paper, we assume that one pair of the messenger fields gives the dominant contribution, and the LHC signature for the two possible messengers, X + \bar{X} and Q + \bar{Q}, are examined. We investigate the possibility to measure the deviation from the usual GUT relation of the gaugino masses which is one of the most important features of these scenarios.

hep-ph↗