arXiv subjects

Atul Sharma

Publications and source records attributed to Atul Sharma.

At least 19 recordsLinked to original sources

arXiv2026

Schwarzschild black holes from twistor space

Twistor theory forms the basis for many surprising advances in areas ranging from dynamical systems to quantum field theory. Yet for almost fifty years, one of the main drawbacks of twistor theory has been its inability to give non-perturbative descriptions of non-chiral (or non-self-dual) field configurations. This difficulty is known as 'the googly problem.' In this paper, we provide a resolution of the googly problem for a particular solution of the vacuum Einstein equations: the Schwarzschild metric. We start with the twistor space of the self-dual Taub-NUT Euclidean gravitational instanton, expressed in Kerr-Schild form. Within this twistor space, we then consider a quadric which corresponds to the anti-self-dual Taub-NUT metric. While the full quadric is not holomorphic with respect to the complex structure of the self-dual Taub-NUT twistor space, its holomorphic locus still has complex dimension two. This 'coincidence locus' -- points in twistor space on the holomorphic portion of the quadric -- inherits a complex structure from the twistor space and a K\"ahler form from the quadric itself. Remarkably, these structures are compatible, giving rise to a non-self-dual, four-dimensional K\"ahler metric which is conformal to Schwarzschild (in Lorentzian or Euclidean signature). This is the first instance of a non-self-dual Einstein metric constructed entirely from holomorphic data in a twistor space.

hep-th↗

arXiv2026

Physics Informed Neural Network-based Computational Method for Accelerating Time-Periodic Unsteady CFD Simulations

Presently, there is a steady state approach in Computational fluid dynamics (CFD) to obtain a steady solution directly from the steady state governing equations. Whereas, for obtaining a time-periodic flow solution, the present unsteady governing equations-based CFD approach starts from an initial condition and requires a large computational time during the initial non-periodic transient phase before reaching the periodic state. For obtaining the periodic flow directly, without transient simulations that may not be of interest, our objective is to propose a Physics Informed Neural Network (PINN)-based periodic CFD approach. The motivation is a substantial reduction in computational time by a meshless PINN-based periodic CFD solver as compared to the present mesh-based transient-to-periodic solver. Proof-of-concept, for the periodic CFD approach, is demonstrated here for 2D periodic heat diffusion and fluid flow problems. The proposed PINN-based periodic solver primarily focuses on the time-periodic state, optimizing the neural network model's trainable parameters to precisely fit a smaller time window (one time-period) rather than the temporal domain starting from the initial condition. After presenting a verification study, effect of the PINN-related various hyperparameters such as the number of collocation points, neural network architecture, and point spacing for numerical differentiation, on computational time and accuracy are presented. Our results demonstrate that the PINN-based periodic solver takes substantially less computational time to achieve almost same accuracy as that obtained by the traditional transient-to-periodic solver.

physics.comp-ph↗

arXiv2026

Observer-Based Fixed-Time Nested Sliding-Mode Control for Tip-Position Regulation of a Single-Link Flexible Manipulator

This paper presents a novel position control strategy for a single-link flexible manipulator, tailored for applications where precise position must be achieved within strict time constraints. To accomplish this objective, firstly, a nested non-singular terminal sliding mode controller is designed for the system, enabling precise and robust control. Furthermore, a fixed-time sliding mode observer is designed to estimate unmeasured system states accurately in a fixed time, thereby enabling closed-loop control implementation. A stability analysis is presented to guarantee the robustness and efficacy of the proposed composite control algorithm. The effectiveness of the proposed fixed-time controller is demonstrated through numerical simulation on accuracy, stability, and convergence speed. The proposed controller's performance is also compared with that of other state-of-the-art control schemes. The proposed controller is further validated through experiments conducted on a real hardware setup.

eess.SY↗

arXiv2026

Celestial 1-form symmetries

The $S$-algebra originally arose as a chiral algebra of asymptotic symmetries of Yang-Mills theory. We show that in the self-dual sector of Yang-Mills, the $S$-algebra gets upgraded to an infinite-dimensional algebra of $1$-form symmetries in the bulk. The associated 2-form currents encode the integrability and hierarchies of self-dual Yang-Mills. As an application, we prove the equality of Carrollian corner charges with modes of the celestial chiral algebra by expressing them as integrals of the same 2-form currents over homologous 2-cycles.

hep-th↗

arXiv2025

Chiral holography

Using twistor string theory, we engineer a gravitational holographic dual of the self-dual sector of $\mathcal{N}=4$ super Yang-Mills theory. This provides a top-down example of holography at zero 't Hooft coupling. Our holographic dictionary operates at the level of D-branes wrapping twistor lines and generalizes the D-instanton prescription used to compute scattering amplitudes. As an application, we study the correlators of determinant operators that act as generating functions of half-BPS operators in the self-dual theory. We show that these correlators are dual to configurations of giant gravitons in the bulk.

hep-th↗

arXiv2025

MAUI: Reconstructing Private Client Data in Federated Transfer Learning

Recent works in federated learning (FL) have shown the utility of leveraging transfer learning for balancing the benefits of FL and centralized learning. In this setting, federated training happens after a stable point has been reached through conventional training. Global model weights are first centrally pretrained by the server on a public dataset following which only the last few linear layers (the classification head) of the model are finetuned across clients. In this scenario, existing data reconstruction attacks (DRAs) in FL show two key weaknesses. First, strongly input-correlated gradient information from the initial model layers is never shared, significantly degrading reconstruction accuracy. Second, DRAs in which the server makes highly specific, handcrafted manipulations to the model structure or parameters (for e.g., layers with all zero weights, identity mappings and rows with identical weight patterns) are easily detectable by an active client. Improving on these, we propose MAUI, a stealthy DRA that does not require any overt manipulations to the model architecture or weights, and relies solely on the gradients of the classification head. MAUI first extracts "robust" feature representations of the input batch from the gradients of the classification head and subsequently inverts these representations to the original inputs. We report highly accurate reconstructions on the CIFAR10 and ImageNet datasets on a variety of model architectures including convolution networks (CNN, VGG11), ResNets (18, 50), ShuffleNet-V2 and Vision Transformer (ViT B-32), regardless of the batch size. MAUI significantly outperforms prior DRAs in reconstruction quality, achieving 40-120% higher PSNR scores.

cs.CR↗

arXiv2025

In Planta Tattoo and Kirigami Sensors for Self-Powered Monitoring of Vapor Pressure Deficit and Growth Dynamics

We report a scalable, self-powered in planta sensor platform for continuous monitoring of plant hydration and growth. The system integrates two components a leaf mounted tattoo sensor for estimating vapor pressure deficit and a kirigami inspired strain sensor for tracking radial stem growth. Uniquely, the tattoo sensor serves a dual function measuring temperature and humidity beneath the leaf surface while simultaneously harvesting power from ambient moisture via a vanadium pentoxide nanosheet membrane. This moist-electric generator configuration enables energy-autonomous operation, delivering a power density of 0.1114 miroW per square cm. The V2O5 based sensor exhibits high sensitivity to humidity and temperature, enabling accurate VPD estimation for over 10 days until leaf senescence. The eutectogel based kirigami strain sensor, wrapped around the stem, offers a gauge factor of 1.5 and immunity to unrelated mechanical disturbances, allowing continuous growth tracking for more than 20 days. Both sensors are fabricated via cleanroom-free, roll to roll compatible methods, underscoring their potential for large-scale agricultural deployment to monitor abiotic stress and improve crop management.

eess.SP↗

arXiv2025

Graviton scattering on self-dual black holes

The computation of gravitational wave scattering on black hole spacetimes is an extremely hard problem, typically requiring approximation schemes that either treat the black hole perturbatively or are only amenable to numerical techniques. In this paper, we consider linearised gravitational waves (or gravitons) scattering on the self-dual analogue of a black hole: namely, the self-dual Taub-NUT metric. Using the hidden integrability of the self-dual sector, we solve the linearised Einstein equations on these self-dual black hole backgrounds exactly in terms of simple, explicit quasi-momentum eigenstates. Using a description of the self-dual Taub-NUT metric and its gravitons in terms of twistor theory, we obtain an explicit formula, exact in the background, for the tree-level maximal helicity violating (MHV) graviton scattering amplitude at arbitrary multiplicity, with and without spin. This is obtained from the description of the MHV amplitudes in terms of the perturbation theory of a chiral sigma model whose target is the twistor space of the background. The incorporation of spin effects on these backgrounds is a straightforward application of the Newman-Janis shift. We also demonstrate that the holomorphic collinear splitting functions in the self-dual background are equal to those in flat space so that the celestial symmetry algebra is undeformed.

hep-th↗

arXiv2025

S-algebra in Gauge Theory: Twistor, Spacetime and Holographic Perspectives

The celestial $S$-algebra arose from a reinterpretation of collinear limits of the Yang-Mills S-matrix as OPEs in celestial holography. It was subsequently represented via asymptotic charge aspects defined in the Yang-Mills radiative phase space defined at null infinity on the one hand, and via a twisted holography vertex algebra construction in twistor space on the other. Here we first identify it with the traditional symmetry algebra of self-dual Yang-Mills theory as an integrable system via its hierarchies of conserved quantities and associated flows; the self-dual phase space can be canonically identified with that of full Yang-Mills at null infinity $\mathscr{I}$. We derive the associated canonical generators from the twistor space action, identifying two infinite towers of charges corresponding to the two gluon helicities. These expressions are translated into spacetime data at null infinity using twistor integral formulae. Examining the charge algebra at spacelike infinity reveals the vertex algebras studied in the context of twisted holography. Our discussion extends directly to the celestial LHam$(\mathbb{C}^2)$ symmetries of self-dual gravity. This analysis provides a unified framework for celestial symmetries, connecting twistor, spacetime, and holographic approaches and culminating in a nonlinear extrapolate dictionary for self-dual gauge theory.

hep-th↗

arXiv2025

Chiral higher-spin theories from twistor space

We reformulate chiral higher-spin Yang-Mills and gravity on $\mathbb{R}^4$ as 'CR-holomorphic' theories of Chern-Simons type; in the most general case, these are Moyal deformed to become non-commutative. They are defined on the space of non-projective twistors of unit length. These spaces carry $S^7$, $S^3\times \mathbb{R}^4$ or AdS$_{3+4}$ metrics but are also endowed with a Cauchy-Riemann structure, an odd-dimensional analogue of a complex structure, with respect to which the theories are holomorphic. They are circle bundles over standard projective twistor spaces and the higher spin fields arise naturally as Kaluza-Klein modes. We give a perturbative analysis to identify the spectrum and three-point vertices on spacetime and, for flat space, in momentum space. These vertices can have helicities $(+++)$, $(++-)$ or $(+--)$, but are nevertheless all of $\overline{\text{MHV}}$ type in the sense that they are supported on momenta with proportional anti-self-dual spinors. On reduction to spacetime, there are higher valence vertices but these appear to be a gauge artifact. Further generalizations are discussed.

hep-th↗

arXiv2025

Hubs and Spokes Learning: Efficient and Scalable Collaborative Machine Learning

We introduce the Hubs and Spokes Learning (HSL) framework, a novel paradigm for collaborative machine learning that combines the strengths of Federated Learning (FL) and Decentralized Learning (P2PL). HSL employs a two-tier communication structure that avoids the single point of failure inherent in FL and outperforms the state-of-the-art P2PL framework, Epidemic Learning Local (ELL). At equal communication budgets (total edges), HSL achieves higher performance than ELL, while at significantly lower communication budgets, it can match ELL's performance. For instance, with only 400 edges, HSL reaches the same test accuracy that ELL achieves with 1000 edges for 100 peers (spokes) on CIFAR-10, demonstrating its suitability for resource-constrained systems. HSL also achieves stronger consensus among nodes after mixing, resulting in improved performance with fewer training rounds. We substantiate these claims through rigorous theoretical analyses and extensive experimental results, showcasing HSL's practicality for large-scale collaborative learning.

cs.LG↗

arXiv2025

Towards nonlinear thermohydrodynamic simulations via the Onsager-Regularized Lattice Boltzmann Method

This work presents a generalized, assumption-free, and stencil-independent theoretical analyses of the recently proposed Onsager-Regularized (OReg) lattice Boltzmann (LB) method [Jonnalagadda et al., Phys. Rev. E 104, 015313 (2021)] and demonstrates its ability to mitigate spurious errors associated with the insufficient isotropy of standard first-neighbor lattices without the inclusion of any external correction terms. The hydrodynamic limit recovered by the OReg scheme is derived for two equilibrium distribution functions, namely the so-called thermal guided equilibrium and the popular second order polynomial equilibrium, to show that the OReg scheme yields macroscopic dynamics that are O(u) times more accurate than that of the bare BGK collision model. Specifically, we show that, with the guided equilibrium on the D2Q9 standard lattice, the OReg scheme inherently compensates for the insufficient lattice isotropy of the standard D2Q9 lattice by automatically adjusting the lattice viscosity, yielding O(u4) and O(u2) accurate kinetic models when operated at the reference and arbitrary temperatures respectively. Further, we also show that the OReg scheme presents an O(u3) accurate kinetic model for the D2Q9 lattice when used with the second order polynomial equilibrium formulation. Thereafter, the accuracy of the OReg-guided-equilibrium kinetic model is numerically demonstrated for quasi-one-dimensional simulations of the rotated decaying shear wave and isothermal shocktube problems. The present work lays the theoretical foundation of a generic framework which can enable fully local, correction-free, nonlinear thermohydrodynamic LB simulations on standard lattices, thereby facilitating scalable simulations of physically challenging fluid flows.

physics.comp-ph↗

arXiv2024

No-Boundary State for Klein Space

Analytic continuation from $(3,1)$ signature Minkowski to $(2,2)$ signature Klein space has emerged as a useful tool for the understanding of scattering amplitudes and flat space holography. Under this continuation, past and future null infinity merge into a single boundary ($\mathcal{J}$) which is the product of a null line with a $(1,1)$ signature torus. The Minkowskian $\mathcal{S}$-matrix continues to a Kleinian $\mathcal{S}$-vector which in turn may be represented by a Poincar\'e-invariant vacuum state $|\mathcal{C}\rangle$ in the Hilbert space built on $\mathcal{J}$. $|\mathcal{C} \rangle$ contains all information about $\mathcal{S}$ in a novel, repackaged form. We give an explicit construction of $|\mathcal{C}\rangle$ in a Lorentz/conformal basis for a free massless scalar. $\mathcal{J}$ separates into two halves $\mathcal{J}_\pm $ which are the asymptotic null boundaries of the regions timelike and spacelike separated from the origin. $|\mathcal{C}\rangle$ is shown to be a maximally entangled state in the product of the $\mathcal{J}_\pm $ Hilbert spaces.

hep-th↗

arXiv2024

Saliva-Sensing Dental Floss: A Novel Tool for Assessing Daily Stress via On-Demand Salivary Cortisol Measurement Integrating Molecularly Imprinted Polymer and Thread Microfluidics

Elevated stress negatively impacts health, increasing the risk of cardiovascular diseases, weak immunity, anxiety, and cognitive decline. Current lab-based methods lack real-time stress measurement, while saliva diagnostics offers a convenient and non-invasive alternative. This work proposes repurposing dental floss as a thread-analytical saliva diagnostic device. This method uses flossing to collect and transport saliva to a flexible electrochemical sensor via capillary microfluidics, where cortisol, a stress biomarker, is measured. By integrating an electropolymerized molecularly imprinted polymer specific to cortisol (cort-eMIP) on a porous-laser-engraved-graphene (PLEG) working electrode, the saliva-floss platform shows a detection range of 0.10-10000 pg mL-1 (R2 = 0.9916) with a detection limit of 0.023 pg mL-1 for cortisol (in buffer medium). The saliva-sensing dental floss provides results within 5 minutes. The thread-based microfluidic design minimizes interference and ensures consistent repeatability when testing artificial and actual human saliva samples, yielding 98.64 to 102.4 percent recoveries with a relative standard deviation of 5.01%, demonstrating high accuracy and precision. Validated using human saliva samples in a stress study, it showed a high correlation (r= 0.9910) with conventional ELISA assays. Combined with a wireless readout, this saliva floss offers a convenient way to monitor daily stress levels and can be extended to detect other critical salivary biomarkers.

physics.med-ph↗

arXiv2024

Gluon scattering on the self-dual dyon

The computation of scattering amplitudes in the presence of non-trivial background gauge fields is an important but extremely difficult problem in quantum field theory. In even the simplest backgrounds, obtaining explicit formulae for processes involving more than a few external particles is often intractable. Recently, it has been shown that remarkable progress can be made by considering background fields which are chiral in nature. In this paper, we obtain a compact expression for the tree-level, maximal helicity violating (MHV) scattering amplitude of an arbitrary number of gluons in the background of a self-dual dyon. This is a Cartan-valued, complex gauge field sourced by a point particle with equal electric and magnetic charges, and can be viewed as the self-dual version of a Coulomb field. Twistor theory enables us to manifest the underlying integrability of the self-dual dyon, trivializing the perturbative expansion in the MHV sector. The formula contains a single position-space integral over a spatial slice, which can be evaluated explicitly in simple cases. As an application of the formula, we show that the holomorphic collinear splitting functions of gluons in the self-dual dyon background are un-deformed from a trivial background, meaning that holomorphic celestial OPE coefficients and the associated chiral algebra are similarly un-deformed. We also comment on extensions of our MHV formula to the full tree-level gluon S-matrix.

hep-th↗

arXiv2024

A Celestial Dual for MHV Amplitudes

It is shown that a 2D CFT consisting of a central charge $c$ Liouville theory, a chiral level one, rank $N$ Kac-Moody algebra and a weight $-3/2$ free fermion holographically generate 4D MHV tree-level scattering amplitudes. The correlators of this 2D CFT give directly the 4D leaf amplitudes associated to a single hyperbolic slice of flat space. The 4D celestial amplitudes arise in a large-$N$ and semiclassical large-$c$ limit, according to the holographic dictionary, as a translationally-invariant combination of leaf amplitudes. A step in the demonstration is showing that the semiclassical limit of Liouville correlators are given by contact AdS$_3$ Witten diagrams.

hep-th↗

arXiv2024

Leak and Learn: An Attacker's Cookbook to Train Using Leaked Data from Federated Learning

Federated learning is a decentralized learning paradigm introduced to preserve privacy of client data. Despite this, prior work has shown that an attacker at the server can still reconstruct the private training data using only the client updates. These attacks are known as data reconstruction attacks and fall into two major categories: gradient inversion (GI) and linear layer leakage attacks (LLL). However, despite demonstrating the effectiveness of these attacks in breaching privacy, prior work has not investigated the usefulness of the reconstructed data for downstream tasks. In this work, we explore data reconstruction attacks through the lens of training and improving models with leaked data. We demonstrate the effectiveness of both GI and LLL attacks in maliciously training models using the leaked data more accurately than a benign federated learning strategy. Counter-intuitively, this bump in training quality can occur despite limited reconstruction quality or a small total number of leaked images. Finally, we show the limitations of these attacks for downstream training, individually for GI attacks and for LLL attacks.

cs.CR↗

arXiv2024

Soft Algebras for Leaf Amplitudes

Celestial MHV amplitudes are comprised of non-distributional leaf amplitudes associated to an AdS$_3$ leaf of a foliation of flat spacetime. It is shown here that the leaf amplitudes are governed by the same infinite-dimensional soft `$S$-algebra' as their celestial counterparts. Moreover, taking the soft limit of the smooth three-point MHV leaf amplitude yields a nondegenerate minus-minus two-point leaf amplitude. The two- and three-point MHV leaf amplitudes are used to compute the plus-minus-minus leaf operator product coefficients.

hep-th↗