Attention-based transformer models have become increasingly prevalent in collider analysis, offering enhanced performance for tasks such as jet tagging. However, they are computationally intensive and require substantial data for training. In this paper, we introduce a new jet classification network using an MLP mixer, where two subsequent MLP operations serve to transform particle and feature tokens over the jet constituents. The transformed particles are combined with subjet information using multi-head cross-attention so that the network is invariant under the permutation of the jet constituents. We utilize two clustering algorithms to identify subjets: the standard sequential recombination algorithms with fixed radius parameters and a new IRC-safe, density-based algorithm of dynamic radii based on HDBSCAN. The proposed network demonstrates comparable classification performance to state-of-the-art models while boosting computational efficiency drastically. Finally, we evaluate the network performance using various interpretable methods, including centred kernel alignment and attention maps, to highlight network efficacy in collider analysis tasks.

In light of a discrepancy of the direct $CP$ violation in $K\to\pi\pi$ decays, $\varepsilon'/\varepsilon_K$, we investigate gluino contributions to the electroweak penguin, where flavor violations are induced by squark trilinear couplings. Top-Yukawa contributions to $\Delta S = 2$ observables are taken into account, and vacuum stability conditions are evaluated in detail. It is found that this scenario can explain the discrepancy of $\varepsilon'/\varepsilon_K$ for the squark mass smaller than 5.6 TeV. We also show that the gluino contributions can amplify $\mathcal{B}(K \to \pi \nu \overline{\nu})$,$\mathcal{B}(K_S \to \mu^+ \mu^-)_{\rm eff}$and$\Delta A_{\rm CP}(b\to s\gamma)$. Such large effects could be measured in future experiments.

We deploy an advanced Machine Learning (ML) environment, leveraging a multi-scale cross-attention encoder for event classification, towards the identification of the $gg\to H\to hh\to b\bar b b\bar b$ process at the High Luminosity Large Hadron Collider (HL-LHC), where $h$ is the discovered Standard Model (SM)-like Higgs boson and $H$ a heavier version of it (with $m_H>2m_h$). In the ensuing boosted Higgs regime, the final state consists of two fat jets. Our multi-modal network can extract information from the jet substructure and the kinematics of the final state particles through self-attention transformer layers. The diverse learned information is subsequently integrated to improve classification performance using an additional transformer encoder with cross-attention heads. We ultimately prove that our approach surpasses in performance current alternative methods used to establish sensitivity to this process, whether solely based on kinematic analysis or else on a combination of this with mainstream ML approaches. Then, we employ various interpretive methods to evaluate the network results, including attention map analysis and visual representation of Gradient-weighted Class Activation Mapping (Grad-CAM). Finally, we note that the proposed network is generic and can be applied to analyse any process carrying information at different scales. Our code is publicly available for generic use.

The Fermilab Muon $g-2$ collaboration recently announced the first result of measurement of the muon anomalous magnetic moment ($g-2$), which confirmed the previous result at the Brookhaven National Laboratory and thus the discrepancy with its Standard Model prediction. We revisit low-scale supersymmetric models that are naturally capable to solve the muon $g-2$ anomaly, focusing on two distinct scenarios: chargino-contribution dominated and pure-bino-contribution dominated scenarios. It is shown that the slepton pair-production searches have excluded broad parameter spaces for both two scenarios, but they are not closed yet. For the chargino-dominated scenario, the models with $m_{\tilde{\mu}_{\rm L}}\gtrsim m_{\tilde{\chi}^{\pm}_1}$ are still widely allowed. For the bino-dominated scenario, we find that, although slightly non-trivial, the region with low $\tan \beta$ with heavy higgsinos is preferred. In the case of universal slepton masses, the low mass regions with $m_{\tilde{\mu}}\lesssim 230$GeV can explain the$g-2$ anomaly while satisfying the LHC constraints. Furthermore, we checked that the stau-bino coannihilation works properly to realize the bino thermal relic dark matter. We also investigate heavy staus case for the bino-dominated scenario, where the parameter region that can explain the muon $g-2$ anomaly is stretched to $m_{\tilde{\mu}}\lesssim 1.3$ TeV.

In this paper, we introduce IAFormer, a novel Transformer-based architecture that efficiently integrates pairwise particle interactions through a dynamic sparse attention mechanism. The IAformer has two new mechanisms within the model. First, the attention matrix depends on predefined boost invariant pairwise quantities, reducing the network parameter significantly from the original particle transformer models. Second, IAformer incorporate the sparse attention mechanism by utilizing the ``differential attention'', so that it can dynamically prioritizes relevant particle tokens while reducing computational overhead associated with less informative ones. This approach significantly lowers the model complexity without compromising performance. Despite being computationally efficient by more than an order of magnitude than the Particle Transformer network, IAFormer achieves state-of-the-art performance in classification tasks on the Top and quark-gluon datasets. Furthermore, we employ AI interpretability techniques, verifying that the model effectively captures physically meaningful information layer by layer through its sparse attention mechanism, building an efficient network output that is resistant to statistical fluctuations. IAformer highlights the need to sparse attention in any Transformer analysis to reduce the network size while improving its performance.

We explore the feasibility of measuring the CP properties of the Higgs boson coupling to $\tau$ leptons at the High Luminosity Large Hadron Collider (HL-LHC). Employing detailed Monte Carlo simulations, we analyze the reconstruction of the angle between $\tau$ lepton planes at the detector level, accounting for various hadronic $\tau$ decay modes. Considering standard model backgrounds and detector resolution effects, we employ three Deep Learning (DL) networks, Multi-Layer Perceptron (MLP), Graph Convolution Network (GCN), and Graph Transformer Network (GTN) to enhance signal-to-background separation. To incorporate CP-sensitive observables into Graph networks, we construct Heterogeneous graphs capable of integrating nodes and edges with different structures within the same framework. Our analysis demonstrates that GTN exhibits superior efficiency compared to GCN and MLP. Under a simplified detector simulation analysis, MLP can exclude CP mixing angle larger than $20^\circ$ at $68\%$ confidence level (CL), while GCN and GTN can achieve exclusions at $90\%$ CL and $95\%$ CL, respectively with $\sqrt{s}=14$~TeV and $\mathcal{L}$$=100\rm { fb}^{-1}$. Furthermore, the DL networks can achieve a significance of approximately $3\sigma$ in excluding the pure CP-odd state.

We report on the observations of two ultra metal-poor (UMP) stars with [Fe/H]~-4.0 including one new discovery. The two stars are studied in the on-going and quite efficient project to search for extremely metal-poor (EMP) stars with LAMOST and Subaru. Detailed abundances or upper limits of abundances have been derived for 15 elements from Li to Eu based on high-resolution spectra obtained with Subaru/HDS. The abundance patterns of both UMP stars are consistent with the "normal-population" among the low-metallicity stars. Both of the two program stars show carbon-enhancement without any excess of heavy neutron-capture elements, indicating that they belong to the subclass of CEMP-no stars, as is the case of most UMP stars previously studied. The [Sr/Ba] ratios of both CEMP-no UMP stars are above [Sr/Ba]~-0.4, suggesting the origin of the carbon-excess is not compatible with the mass transfer from an AGB companion where the s-process has operated. Lithium abundance is measured in the newly discovered UMP star LAMOST J125346.09+075343.1, making it the second UMP turnoff star with Li detection. The Li abundance of LAMOST J125346.09+075343.1 is slightly lower than the values obtained for less metal-poor stars with similar temperature, and provides a unique data point at [Fe/H]~-4.2 to support the "meltdown" of the Li Spite-plateau at extremely low metallicity. Comparison with the other two UMP and HMP (hyper metal-poor with [Fe/H]<-5.0) turnoff stars suggests that the difference in lighter elements such as CNO and Na might cause notable difference in lithium abundances among CEMP-no stars.

New physics contributions to the $Z$ penguin are revisited in the light of the recently-reported discrepancy of the direct CP violation in $K\to\pi\pi$. Interference effects between the standard model and new physics contributions to $\Delta S = 2$ observables are taken into account. Although the effects are overlooked in the literature, they make experimental bounds significantly severer. It is shown that the new physics contributions must be tuned to enhance $\mathcal{B}(K_L \to \pi^{0} \nu \bar{\nu})$, if the discrepancy of the direct CP violation is explained with satisfying the experimental constraints. The branching ratio can be as large as $6 \times 10^{-10}$ when the contributions are tuned at the 10 % level.

In this work, we investigated the electroweak vacuum instability during or after inflation. In the inflationary Universe, i.e., de Sitter space, the vacuum field fluctuations \left&lt; {\delta \phi }^{ 2 } \right&gt; enlarge in proportion to the Hubble scale $H^{2}$. Therefore, the large inflationary vacuum fluctuations of the Higgs field \left&lt; {\delta \phi }^{ 2 } \right&gt; are potentially catastrophic to trigger the vacuum transition to the negative-energy Planck-scale vacuum state and cause an immediate collapse of the Universe. However, the vacuum field fluctuations $\left< {\delta \phi }^{ 2 } \right>$, i.e., the vacuum expectation values have an ultraviolet divergence, and therefore a renormalization is necessary to estimate the physical effects of the vacuum transition. Thus, in this paper, we revisit the electroweak vacuum instability from the perspective of quantum field theory (QFT) in curved space-time, and discuss the dynamical behavior of the homogeneous Higgs field $\phi$ determined by the effective potential ${ V }_{\rm eff}\left( \phi \right)$in curved space-time and the renormalized vacuum fluctuations$\left< {\delta \phi }^{ 2 } \right>_{\rm ren}$ via adiabatic regularization and point-splitting regularization. We simply suppose that the Higgs field only couples the gravity via the non-minimal Higgs-gravity coupling $\xi(\mu)$. In this scenario, the electroweak vacuum stability is inevitably threatened by the dynamical behavior of the homogeneous Higgs field $\phi$, or the formations of AdS domains or bubbles unless the Hubble scale is small enough $H< \Lambda_{I} $.

We report on the elemental abundances of the carbon-enhanced metal-poor (CEMP) star J2217+2104 discovered by our metal-poor star survey with LAMOST and Subaru. This object is a red giant having extremely low Fe abundance ([Fe/H]=-4.0) and very large enhancement of C, N, and O with excesses of Na, Mg, Al, and Si. This star is a new example of a small group of such CEMP stars identified by previous studies. We find a very similar abundance pattern for O-Zn in this class of objects that shows enhancement of elements up to Si and normal abundance of Ca and Fe-group elements. Whereas the C/N ratio is different among these stars, the (C+N)/O ratio is similar. This suggests that C was also yielded with similar abundance ratios relative to O-Zn in progenitors, and was later affected by the CN-cycle. By contrast, the heavy neutron-capture elements Sr and Ba are deficient in J2217+2104, compared to the four objects in this class previously studied. This indicates that the neutron-capture process in the early Galaxy, presumably the r-process, has no direct connection to the phenomenon that has formed such CEMP stars. Comparisons of the abundance pattern well determined for such CEMP stars with those of supernova nucleosynthesis models constrain the progenitor mass to be about 25Msun, which is not particularly different from typical mass of progenitors expected for extremely metal-poor stars in general.

From a theoretical point of view, there is a strong motivation to consider an MeV-scale reheating temperature induced by long-lived massive particles with masses around the weak scale, decaying only through gravitational interaction. In this study, we investigate lower limits on the reheating temperature imposed by big-bang nucleosynthesis assuming both radiative and hadronic decays of such massive particles. For the first time, effects of neutrino self-interactions and oscillations are taken into account in the neutrino thermalization calculations. By requiring consistency between theoretical and observational values of light element abundances, we find that the reheating temperature should conservatively be $T_{\rm RH} \gtrsim 1.8$ MeV in the case of the 100% radiative decay, and $T_{\rm RH} \gtrsim$ 4-5 MeV in the case of the 100% hadronic decays for particle masses in the range of 10 GeV to 100 TeV.

Recently, it was pointed out that the electron and muon g-2 discrepancies can be explained simultaneously by a flavor-violating axion-like particle (ALP). We show that the parameter regions favored by the muon g-2 are already excluded by the muonium-antimuonium oscillation bound. In contrast, those for the electron g-2 can be consistent with this bound when the ALP is heavier than 1.5 GeV. We propose to search for a signature of the same-sign and same-flavor lepton pairs and the forward-backward muon asymmetry to test the model at the Belle II experiment.

Chemical compositions are determined based on high-resolution spectroscopy for 137 candidate extremely metal-poor (EMP) stars selected from the Sloan Digital Sky Survey (SDSS) and its first stellar extension, the Sloan Extension for Galactic Understanding and Exploration (SEGUE). High-resolution spectra with moderate signal-to-noise (S/N) ratios were obtained with the High Dispersion Spectrograph of the Subaru Telescope. Most of the sample (approximately 80%) are main-sequence turn-off stars, including dwarfs and subgiants. Four cool main-sequence stars, the most metal-deficient such stars known, are included in the remaining sample. Good agreement is found between effective temperatures estimated by the SEGUE stellar parameter pipeline, based on the SDSS/SEGUE medium-resolution spectra, and those estimated from the broadband $(V-K)_0$ and $(g-r)_0$ colors. Our abundance measurements reveal that 70 stars in our sample have [Fe/H] &lt; -3, adding a significant number of EMP stars to the currently known sample. Our analyses determine the abundances of eight elements (C, Na, Mg, Ca, Ti, Cr, Sr, and Ba) in addition to Fe. The fraction of carbon-enhanced metal-poor stars ([C/Fe]&gt; +0.7) among the 25 giants in our sample is as high as 36%, while only a lower limit on the fraction (9%) is estimated for turn-off stars. This paper is the first of a series of papers based on these observational results. The following papers in this series will discuss the higher-resolution and higher-S/N observations of a subset of this sample, the metallicity distribution function, binarity, and correlations between the chemical composition and kinematics of extremely metal-poor stars.

We consider the effects of the injections of energetic photon and electron (or positron) on the big-bang nucleosynthesis. We study the photodissociation of light elements in the early Universe paying particular attention to the case that the injection energy is sub-GeV and derive upper bounds on the primordial abundances of the massive decaying particle as a function of its lifetime. We also discuss a solution of the $^7$Li problem in this framework.

We present the results of the wide-field $^{12}$CO (1--0) observations of the nearby barred galaxy M83 carried out with the Nobeyama Millimeter Array (NMA). The interferometric data are combined with the data obtained with the Nobeyama 45-m telescope to recover the total-flux. The target fields of the observations cover the molecular bar and part of the spiral arms, with a spatial resolution of ~110 pc x 260 pc. By exploiting the resolution and sensitivity to extended CO emission, the impact of the galactic structures on the molecular gas content is investigated in terms of the gas kinematics and the star formation. By inspecting the gas kinematics, the pattern speed of the bar is estimated to be 57.4 $\pm$ 2.8 km s$^{-1}$ kpc$^{-1}$, which places the corotation radius to be about 1.7 times the semi-major radius of the bar. Within the observed field, HII regions brighter than 10$^{37.6}$ erg s$^{-1}$ in H{\alpha} luminosity are found to be preferentially located downstream of the CO emitting regions. Azimuthal angular offsets between molecular gas and star forming (SF) calculated with the angular cross-correlation method confirm the trend. By comparing with a cloud orbit model based on the derived pattern speed, the angular offsets are found to be in accordance with a time delay of about 10 Myr. Finally, to test whether the arm/bar promote star formation efficiency (SFE $\equiv$ Star Formation Rate (SFR)/H$_2$ mass), SFR is derived with the diffuse-background-subtracted H{alpha} and 24{\mu}m images. The arm-to-interarm ratio of the SFE is found to lie in the range of 2 to 5, while it is ~1 if no background-removal is performed. The CO-SF offsets and the enhancement of the SFE in the arm/bar found in the inner region of M83 are in agreement with the predictions of the classical galactic shock model.

Li-ion batteries are essential for the energy supply of satellites. The accurate estimation of their states is important for the reliable and safe operation in space. This paper introduces a new algorithm for the estimation of SOC and SOH. The multi-timescale algorithm combines Kalman filters and physics-based models for batteries. We use a P2D model combined with a degradation model that describes capacity fading due to SEI growth. The state estimation algorithm combines two extended Kalman filters for the two states evolving on different timescales, with one filter nested within the other one. We test the algorithm with synthetic data as well as with in-flight data from Japanese satellite REIMEI. The algorithm adequately estimates the SOC and SOH in both cases. Furthermore it gives insight into the reliability of the chosen model.

We have carried out the first very long baseline interferometry (VLBI) imaging of 44 GHz class I methanol maser (7_{0}-6_{1}A^{+}) associated with a millimeter core MM2 in a massive star-forming region IRAS 18151-1208 with KaVA (KVN and VERA Array), which is a newly combined array of KVN (Korean VLBI Network) and VERA (VLBI Exploration of Radio Astrometry). We have succeeded in imaging compact maser features with a synthesized beam size of 2.7 milliarcseconds x 1.5 milliarcseconds (mas). These features are detected at a limited number of baselines within the length of shorter than approximately 650 km corresponding to 100 Mlambda in the uv-coverage. The central velocity and the velocity width of the 44 GHz methanol maser are consistent with those of the quiescent gas rather than the outflow traced by the SiO thermal line. The minimum component size among the maser features is ~ 5 mas x 2 mas, which corresponds to the linear size of ~ 15 AU x 6 AU assuming a distance of 3 kpc. The brightness temperatures of these features range from ~ 3.5 x 10^{8} to 1.0 x 10^{10} K, which are higher than estimated lower limit from a previous Very Large Array observation with the highest spatial resolution of ~ 50 mas. The 44 GHz class I methanol maser in IRAS 18151-1208 is found to be associated with the MM2 core, which is thought to be less evolved than another millimeter core MM1 associated with the 6.7 GHz class II methanol maser.

Once all the sleptons as well as the Bino are observed at the ILC, the Bino contribution to the muon anomalous magnetic dipole moment (muon $g-2$) in supersymmetric (SUSY) models can be reconstructed. Motivated by the recently confirmed muon $g-2$ anomaly, we examine the reconstruction accuracy at the ILC with $\sqrt{s}$ = 500 GeV. For this purpose, measurements of stau parameters are important. We quantitatively study the determination of the mass and mixing parameters of the staus at the ILC. Furthermore, we discuss the implication of the stau study to the reconstruction of the SUSY contribution to the muon $g-2$. At the benchmark point of our choice, we find that the SUSY contribution to the muon $g-2$ can be determined with a precision of $\sim 1\%$ at the ILC.

The chemical abundances of very metal-poor stars provide important constraints on the nucleosynthesis of the first generation of stars and early chemical evolution of the Galaxy. We have obtained high-resolution spectra with the Subaru Telescope for candidates of very metal-poor stars selected with a large survey of Galactic stars carried out with LAMOST. In this series of papers, we report on the elemental abundances of about 400 very metal-poor stars and discuss the kinematics of the sample obtained by combining the radial velocities measured in this study and recent astrometry obtained with Gaia. This paper provides an overview of our survey and follow-up program, and reports radial velocities for the whole sample. We identify seven double-lined spectroscopic binaries from our high-resolution spectra, for which radial velocities of the components are reported. We discuss the frequency of such relatively short-period binaries at very low metallicity.