arXiv subjects

Michael Geller

Publications and source records attributed to Michael Geller.

At least 19 recordsLinked to original sources

arXiv2026

Post-Recombination Fluctuations from a Sequestered Dark Sector

We develop a formalism to characterize the imprints of late-time sources of cosmological fluctuations under the sole assumption that the injection occurs on timescales short compared to the horizon. For post-recombination injections, we derive the general modification of photon geodesics in the presence of scalar, vector, and tensor perturbations, and compute the resulting impact on the Cosmic Microwave Background through the integrated Sachs-Wolfe effect. We show that the signal is generically dominated by instantaneous injections of anisotropic stress. As an application, we consider first-order phase transitions in a sequestered dark sector and show that current observations constrain fractional energy injections at the permille level.

astro-ph.CO↗

arXiv2026

Dissipative Losses In Black Hole-Induced Vacuum Decay

We address the long-standing puzzle of false vacuum decay catalyzed by black holes. Naively, small black holes with large Hawking temperatures can generate highly-boosted true vacuum bubbles in the early universe and trigger vacuum decay without any exponential suppression. Working in the thin-wall regime of the $\phi^4$ and sine-Gordon models, we show that radiative losses play a crucial role in decelerating these bubbles and preventing runaway vacuum decay. We find that while the production rate is enhanced compared with vacuum tunneling in some parts of the parameter space, it is always exponentially suppressed.

hep-ph↗

arXiv2025

Natural Ultralight Dark Matter: The Quadratic Twin

Scalar ultralight dark matter (ULDM) is uniquely accessible to tabletop experiments such as clocks and interferometers, and its search has been the focus of a vast experimental effort. However, the scalar ULDM mass is not protected from radiative corrections, and the entirety of the parameter space within reach of experiments suffers from a severe naturalness problem. In this paper, we propose a new twin mechanism that protects the mass of the scalar ULDM. Our scalar ULDM is a pseudo-Nambu-Goldstone boson with quadratic couplings to the Standard Model (SM) and to a twin copy of the SM, with a mirror $\mathbb{Z}_2$ symmetry exchanging each SM particle with its twin. Due to the mirror symmetry, the leading-order mass correction is quadratic in the (tiny) coupling while the linear order is canceled. This opens up vast regions of parameter space for natural quadratically coupled ultralight dark matter, within the sensitivity of existing and future experiments.

hep-ph↗

arXiv2025

Ultralight Dilatonic Dark Matter

The dilaton, a pseudo-Nambu-Goldstone boson (pNGB) of broken scale invariance, is an appealing ultralight dark matter (DM) candidate. Its mass is protected by conformal invariance and it can be searched for in tabletop experiments. However, contrary to standard pNGBs of internal symmetries, the dilaton generically has a large non-derivative self-coupling, leading to radiative contributions to its mass of the order of its decay constant. Hence typical ultralight dilatons should also have sub-eV decay constants, which would incur significant deviations from standard DM behavior at structure formation times, in severe tension with observations. Therefore, a fine-tuning is required to generate a hierarchy between the mass and the decay constant. In this work, we consider whether supersymmetry (SUSY) can be used to protect this hierarchy from quantum corrections. To ensure an ultralight dilaton mass robust against realistic SUSY-breaking contributions, we must consider a novel dilaton stabilization mechanism. The observed DM abundance can be produced by the misalignment mechanism for dilaton masses ranging from $10^{-11}$ to $1$ eV. Unfortunately, irreducible SUSY-breaking corrections due to gravity restrict the couplings between the dilaton and the Standard Model to be extremely small, beyond the reach of any current or proposed experiments. Our work demonstrates that constructing a consistent model of ultralight dilaton DM is quite involved.

hep-ph↗

arXiv2025

Advancing Personalized Federated Learning: Integrative Approaches with AI for Enhanced Privacy and Customization

In the age of data-driven decision making, preserving privacy while providing personalized experiences has become paramount. Personalized Federated Learning (PFL) offers a promising framework by decentralizing the learning process, thus ensuring data privacy and reducing reliance on centralized data repositories. However, the integration of advanced Artificial Intelligence (AI) techniques within PFL remains underexplored. This paper proposes a novel approach that enhances PFL with cutting-edge AI methodologies including adaptive optimization, transfer learning, and differential privacy. We present a model that not only boosts the performance of individual client models but also ensures robust privacy-preserving mechanisms and efficient resource utilization across heterogeneous networks. Empirical results demonstrate significant improvements in model accuracy and personalization, along with stringent privacy adherence, as compared to conventional federated learning models. This work paves the way for a new era of truly personalized and privacy-conscious AI systems, offering significant implications for industries requiring compliance with stringent data protection regulations.

cs.LG↗

arXiv2025

Tunable Multilayer Surface Plasmon Resonance Biosensor for Trace-Level Toxin Detection

This paper presents a comprehensive study on a novel multilayer surface plasmon resonance (SPR) biosensor designed for detecting trace-level toxins in liquid samples with exceptional precision and efficiency. Leveraging the Kretschmann configuration, the proposed design integrates advanced two-dimensional materials, including black phosphorus (BP) and transition metal dichalcogenides (TMDs), to significantly enhance the performance metrics of the sensor. Key innovations include the optimization of sensitivity through precise material layering, minimization of full-width at half-maximum (FWHM) to improve signal resolution, and maximization of the figure of merit (FoM) for superior detection accuracy. Numerical simulations are employed to validate the structural and functional enhancements of the biosensor. The results demonstrate improved interaction between the evanescent field and the analyte, enabling detection at trace concentrations with higher specificity. This biosensor is poised to contribute to advancements in biochemical sensing, environmental monitoring, and other critical applications requiring high-sensitivity toxin detection.

eess.SY↗

arXiv2025

Meta-Federated Learning: A Novel Approach for Real-Time Traffic Flow Management

Efficient management of traffic flow in urban environments presents a significant challenge, exacerbated by dynamic changes and the sheer volume of data generated by modern transportation networks. Traditional centralized traffic management systems often struggle with scalability and privacy concerns, hindering their effectiveness. This paper introduces a novel approach by combining Federated Learning (FL) and Meta-Learning (ML) to create a decentralized, scalable, and adaptive traffic management system. Our approach, termed Meta-Federated Learning, leverages the distributed nature of FL to process data locally at the edge, thereby enhancing privacy and reducing latency. Simultaneously, ML enables the system to quickly adapt to new traffic conditions without the need for extensive retraining. We implement our model across a simulated network of smart traffic devices, demonstrating that Meta-Federated Learning significantly outperforms traditional models in terms of prediction accuracy and response time. Furthermore, our approach shows remarkable adaptability to sudden changes in traffic patterns, suggesting a scalable solution for real-time traffic management in smart cities. This study not only paves the way for more resilient urban traffic systems but also exemplifies the potential of integrated FL and ML in other real-world applications.

cs.LG↗

arXiv2025

Integrating Personalized Federated Learning with Control Systems for Enhanced Performance

In the expanding field of machine learning, federated learning has emerged as a pivotal methodology for distributed data environments, ensuring privacy while leveraging decentralized data sources. However, the heterogeneity of client data and the need for tailored models necessitate the integration of personalization techniques to enhance learning efficacy and model performance. This paper introduces a novel framework that amalgamates personalized federated learning with robust control systems, aimed at optimizing both the learning process and the control of data flow across diverse networked environments. Our approach harnesses personalized algorithms that adapt to the unique characteristics of each client's data, thereby improving the relevance and accuracy of the model for individual nodes without compromising the overall system performance. To manage and control the learning process across the network, we employ a sophisticated control system that dynamically adjusts the parameters based on real-time feedback and system states, ensuring stability and efficiency. Through rigorous experimentation, we demonstrate that our integrated system not only outperforms standard federated learning models in terms of accuracy and learning speed but also maintains system integrity and robustness in face of varying network conditions and data distributions. The experimental results, obtained from a multi-client simulated environment with non-IID data distributions, underscore the benefits of integrating control systems into personalized federated learning frameworks, particularly in scenarios demanding high reliability and precision.

cs.LG↗

arXiv2024

Probing Dark Matter Isocurvature with Primordial Non-Gaussianity

Multiple fields can become dynamical during the inflationary epoch. We consider an example where a light field acquires isocurvature fluctuations during inflation and contributes to the dark matter abundance at late times. Interactions between the light field and the adiabatic sector contribute to mixed adiabatic-isocurvature non-Gaussianity (NG). We show the resulting form of NG has a different kinematic dependence than the 'local shape' commonly considered, and highlight the parameter space where a dedicated search is expected to significantly improve the current $\textit{Planck}$ sensitivity. We interpret our results in the context of the QCD axion and illustrate how the proposed NG searches can improve upon the existing searches for isocurvature power spectrum and bispectrum.

astro-ph.CO↗

arXiv2024

Challenges in Interpreting the NANOGrav 15-Year Data Set as Early Universe Gravitational Waves Produced by ALP Induced Instability

In this paper, we study a possible early universe source for the recent observation of a stochastic gravitational wave background at the NANOGrav pulsar timing array. The source is a tachyonic instability in a dark gauge field induced by an axion-like particle (ALP), a known source for gravitational waves. We find that relative to the previous analysis with the NANOGrav 12.5-year data set, the current 15-year data set favors parameter space with a relatively larger axion mass and decay constant. This favored parameter space is heavily constrained by $ΔN_{\rm eff}$ and overproduction of ALP dark matter. While there are potential mechanisms for avoiding the second problem, evading the $ΔN_{\rm eff}$ constraint remains highly challenging. In particular, we find that the gravitational wave magnitude is significantly suppressed with respect to the gauge boson dark radiation, which implies that successfully explaining the NANOGrav observation requires a large additional dark radiation, violating the cosmological constraints. Satisfying the $ΔN_{\rm eff}$ constraint will limit the potential contribution from this mechanism to the observed signal to at most a percent level.

hep-ph↗

arXiv2023

Relevant Dilaton Stabilization

We propose a simple modification of the Goldberger-Wise mechanism for stabilizing the scale of spontaneously broken conformal theories. The source of explicit conformal symmetry breaking is a relevant operator with a small coefficient, as opposed to the usual mechanism of an almost marginal operator with an order-one coefficient. In the warped 5D picture this relevant stabilization corresponds to a small tadpole for the bulk scalar on the UV brane, which can be technically natural if it is the only source for the breaking of a symmetry (for example, a discrete $Z_2$). This modification of the stabilization mechanism has significant consequences for the nature of the conformal phase transition, since the radion/dilaton potential is no longer shallow. The bounce action is significantly reduced, leading to a weaker first-order phase transition instead of the supercooled and strongly first-order transition seen in Goldberger-Wise stabilization. This also leads to reduction of gravitational wave signals which, however, may still be observable at future detectors. We present numerical and analytical studies of the phase transition and the resulting gravitational wave signal strength, assuming that the effective dilaton potential provides a good leading approximation. While the dilaton is not expected to be generically light in this setup, in order to keep perturbative control over the effective theory one needs to mildly tune the dilaton quartic to be somewhat small.

hep-ph↗

arXiv2023

A new source for light dark matter isocurvature in low scale inflation

Light scalar and pseudoscalar particles are compelling dark matter candidates, with a vast running experimental program to discover them. Previous studies have shown that these light fields can generate sizable isocurvature perturbations in high scale inflationary models. Thereby, dark matter existence and cosmic microwave background measurements impose an upper bound on the inflationary scale. In this work, focusing on the axion case, we point out that light fields present during inflation can generate important isocurvature perturbations also in scenarios of low-scale inflation. In our mechanism, the axion field starts with some non-zero field value during inflation and rolls along its potential. Since inflation has a different duration in different patches of the universe, different regions will then have different values of the axion field, generating cold dark matter isocurvature modes. These modes are fully correlated with the adiabatic ones and share the same spectral index. In this scenario, the axion mass determines a lower bound on the Hubble parameter during inflation.

hep-ph↗

arXiv2023

Boosting Asymmetric Charged DM via Thermalization

We consider a dark sector scenario with two dark matter species with opposite dark $U(1)$ charges and an asymmetric population comprising some fraction of the dark matter abundance. A new mechanism for boosting dark matter is introduced, arising from the large mass hierarchy between the two particles. In the galaxy, the two species thermalize efficiently through dark Rutherford scattering greatly boosting the lighter dark matter particle, far above the virial and escape velocities in the galaxy, while the dark charge prevents it from escaping. We study the consequences of this scenario for direct-detection experiments, assuming a kinetic mixing between the dark photon and the photon. If the charged dark sector makes up 5% of the total DM mass in our galaxy and the mass ratio is between $10^3-10^4$, we find that current and future experiments may probe the boosted light dark matter for masses down to 100 keV, in a hitherto unexplored parameter range.

hep-ph↗

arXiv2022

Gravitational Waves from Incomplete Inflationary Phase Transitions

We study the observable implications of an incomplete first order phase transition during inflation. In such a phase transition, the nucleated bubbles do not percolate and instead are continuously produced until the onset of reheating. The process creates an inhomogeneity with a distinct power spectrum that depends on both the physics of the phase transition and the inflationary dynamics. Upon horizon re-entry, this spectrum generates gravitational waves through non-linear effects. This stochastic gravitational wave background is predicted to have unique signatures that may be detectable by future experiments spanning a wide frequency range. The discovery of such a gravitational wave signal would shed a light on the detailed dynamics of inflation.

hep-ph↗

arXiv2021

B-modes from Post-inflationary Gravitational Waves Sourced by Axionic Instabilities at Cosmic Reionization

We show that axion-like particles that only couple to invisible dark photons can generate visible B-mode signals around the reionization epoch. The axion field starts rolling shortly before reionization, resulting in a tachyonic instability for the dark photons. This generates an exponential growth of the dark photon quanta sourcing both scalar metric modes and gravitational waves that leave an imprint on the reionized baryons. The tensor modes modify the cosmic microwave background (CMB) polarization at reionization, generating visible B-mode signatures for the next generation of CMB experiments for parameter ranges that satisfy the current experimental constraints.

hep-ph↗

arXiv2021

A Custodial Symmetry for Muon g-2

We discuss the recent results on the muon anomalous magnetic moment in the context of new physics models with light scalars. We propose a model in which the one-loop contributions to g-2 of a scalar and a pseudoscalar naturally cancel in the massless limit due to the symmetry structure of the model. This model allows to interpolate between two possible interpretations. In the first interpretation, the results provide a strong evidence of the existence of new physics, dominated by the positive contribution of a CP-even scalar. In the second one, supported by the recent lattice result, the data provides a strong upper bound on new physics, specifically in the case of (negative) pseudoscalar contributions. We emphasize that tree-level signatures of the new degrees of freedom of the model are enhanced relative to conventional explanations of the discrepancy. As a result, this model can be tested in the near future with accelerator-based experiments and possibly also at the precision frontier.

hep-ph↗

arXiv2021

Direct Detection of Mirror Matter in Twin Higgs Models

We explore the possibility of discovering the mirror baryons and electrons of the Mirror Twin Higgs model in direct detection experiments, in a scenario in which these particles constitute a subcomponent of the observed DM. We consider a framework in which the mirror fermions are sub-nano-charged, as a consequence of kinetic mixing between the photon and its mirror counterpart. We consider both nuclear recoil and electron recoil experiments. The event rates depend on the fraction of mirror DM that is ionized, and also on its distribution in the galaxy. Since mirror DM is dissipative, at the location of the Earth it may be in the form of a halo or may have collapsed into a disk, depending on the cooling rate. For a given mirror DM abundance we determine the expected event rates in direct detection experiments for the limiting cases of an ionized halo, an ionized disk, an atomic halo and an atomic disk. We find that by taking advantage of the complementarity of the different experiments, it may be possible to establish not just the multi-component nature of mirror dark matter, but also its distribution in the galaxy. In addition, a study of the recoil energies may be able to determine the masses and charges of the constituents of the mirror sector. By showing that the mass and charge of mirror helium are integer multiples of those of mirror hydrogen, these experiments have the potential to distinguish the mirror nature of the theory. We also carefully consider mirror plasma screening effects, showing that the capture of mirror dark matter particles in the Earth has at most a modest effect on direct detection signals.

hep-ph↗

arXiv2020

Detecting Dark Photons from Atomic Rearrangement in the Galaxy

We study a new dark sector signature for an atomic process of "rearrangement" in the galaxy. In this process, a hydrogen-like atomic dark matter state together with its anti-particle can rearrange to form a highly-excited bound state. This bound state will then de-excite into the ground state emitting a large number of dark photons that can be measured in experiments on Earth through their kinetic mixing with the photon. We find that for DM masses in the GeV range, the dark photons have enough energy to pass the thresholds of neutrino observatories such as Borexino and Super-Kamiokande that can probe for our scenario even when our atomic states constitute a small fraction of the total DM abundance. We study the corresponding bounds on the parameters of our model from current data as well as the prospects for future detectors.

hep-ph↗