arXiv subjects

Pankaj Kumar

Publications and source records attributed to Pankaj Kumar.

At least 19 recordsLinked to original sources

arXiv2026

Query-Side Attacks on GNN-Based KGQA: Tracing Failures from Entity Linking to Answer Generation

GNN-based Knowledge Graph Question Answering (KGQA) pipelines process queries through four discrete stages: entity linking, subgraph retrieval, GNN reasoning, and answer generation. Standard robustness evaluations conflate stage-level failures into a single end-to-end metric, obscuring both the source of brittleness and the appropriate mitigation target. We ask which stage fails, and why, when the pipeline is subjected to adversarial perturbations on the input question. We introduce a stage-isolation protocol with two answer-preserving adversarial perturbations verified against the knowledge graph: Compositional Restructuring (CR) and Relation Synonym Swap (RS) target distinct stages while leaving entity seeds intact. Evaluated across ComplexWebQuestions and WebQSP, the results run counter to prevailing assumptions: the GNN reasoning stage retains near-baseline accuracy when the subgraph is intact, while subgraph construction accounts for over 99\% of the end-to-end collapse under CR, occurring even when the gold answer is present in 74\% of retrieved subgraphs. This exposes a fundamental distinction between answer presence and answer reachability that end-to-end metrics cannot detect, and places the mitigation target firmly at the subgraph construction stage rather than the reasoning model. Perturbed datasets and evaluation infrastructure are released at https://anonymous.4open.science/r/atkgrag-E85C .

cs.CL↗

arXiv2026

Breakout/Interchange Reconnection as a driver of Jets, Fast CME, and Solar Energetic Particles

Understanding how energetic particles are accelerated and released from the low corona into the interplanetary medium during solar eruptions is crucial for space weather research. Here, we present multiwavelength observations of a solar eruption that are consistent with breakout reconnection playing an important role in driving a fast coronal mass ejection (CME) and the associated solar energetic particle (SEP) event. Extreme-ultraviolet and radio observations reveal evidence of breakout reconnection within a fan-spine topology. The filament eruption begins after the pre-eruption opening, accompanied by quasiperiodic jets and associated downflows near the null point, as well as recurrent faint Type III radio bursts during the ongoing slow breakout reconnection. Furthermore, the observations also reveal the formation of pre-eruption coronal rain via slow interchange reconnection near the null point. A large-scale circular ribbon, along with simultaneous four hard X-ray footpoint sources and intense Type III radio bursts, was observed during the explosive breakout reconnection that enabled opening of field lines and allowed energetic particles to escape into interplanetary space. In situ measurements by PSP and Wind confirm the prompt injection of electron beams consistent with the timing of the explosive breakout reconnection. A fast shock associated with the erupting flux rope during interchange reconnection played a major role in producing the gradual SEP event. These observations highlight the key role of breakout reconnection in producing a fast CME/shock with the SEP release and acceleration process. These results have broader implications for particle acceleration and release processes in multiscale null-point topologies, which produce a continuum of eruptions ranging from small-scale jets to large-scale CMEs.

astro-ph.SR↗

arXiv2026

A Nonsingular Logarithmic Bouncing Cosmology in $f(R,T)$ Gravity with Thermodynamic Viability

We present a nonsingular bouncing cosmological model in the framework of modified $f(R,T)$ gravity within a spatially flat Friedmann--Robertson--Walker universe. A logarithmic time-dependent scale factor is assumed to realize a smooth transition from a contracting phase to an expanding phase without encountering an initial singularity. Based on this assumption, the dynamical evolution of the Hubble parameter, deceleration parameter, energy density, and pressure is obtained for various choices of the model and the matter--geometry coupling parameter to confirm the occurrence of a successful bounce. The effective equation of state parameter is examined to characterize the cosmic fluid during different evolutionary phases. The violation of energy conditions, necessary for the realization of the bouncing behavior, is also discussed. The stability of the model is investigated using the squared speed of sound and is found to remain positive within the allowed parameter space, indicating classical stability. Furthermore, constraints on the matter--geometry coupling parameter are obtained by demanding positive energy density, negative pressure, and a viable cosmological evolution. The obtained cosmological constraint on the coupling parameter is also shown to be compatible with the currently available compact-object constraints. The thermodynamic behavior of the model is examined by testing the generalized second law of thermodynamics. The total entropy production rate remains negative during the contracting phase and changes its sign to positive during the expanding phase. However, it becomes singular at the bouncing point, reflecting the breakdown of the standard thermodynamic description during the transition phase.

gr-qc↗

arXiv2026

Running Vacuum Cosmology in f(R,T) Gravity: Observational Constraints and Thermodynamic Analysis

Despite the remarkable observational success of the $\Lambda$CDM model, the physical origin of dark energy and the cosmological constant remain unresolved, motivating the exploration of dynamical vacuum scenarios within modified theories of gravity. Inspired by the quantum field theoretical description of running vacuum energy, we investigate a running vacuum cosmology in the framework of linear $f(R,T)=R+\lambda T$ gravity. Unlike recent studies that establish an effective correspondence between running vacuum energy and modified gravity, our approach directly incorporates a quantum field theory motivated running vacuum scenario into the $f(R,T)$ framework and examines its cosmological implications. Exact analytical solutions of the modified field equations are obtained in a spatially flat Friedmann--Lema\^itre--Robertson--Walker spacetime, yielding an analytical expression for the Hubble parameter. The free model parameters are constrained using the latest Pantheon+SH0ES compilation, DESI baryon acoustic oscillation measurements, and their joint dataset through Markov Chain Monte Carlo analysis. The observationally constrained model successfully reproduces the observed expansion history, remains consistent with independent Cosmic Chronometer measurements, predicts an age of the universe within current observational bounds, and yields an effective equation of state compatible with current observations. Furthermore, the statefinder and $Om$ diagnostics indicate close agreement with the $\Lambda$CDM scenario at late times, while the generalized second law of thermodynamics remains valid throughout cosmic evolution. These results demonstrate that the proposed framework provides an observationally viable and thermodynamically consistent realization of running vacuum cosmology in $f(R,T)$ gravity, offering a theoretically motivated alternative for describing the present accelerated universe.

physics.gen-ph↗

arXiv2026

Attacking Graph Foundation Models Through Their Shared Representation

A graph foundation model generalizes across graph domains by mapping every input into one shared representation before any task reasoning. We call this map the alignment layer, the component that separates a graph foundation model from a graph neural network, and we show it is a distinct attack surface that prior work has not studied. We attack it at inference time, with no access to training, on six public models spanning spectral tokenizers, text embedding spaces, and a discrete codebook. A directed representation-space perturbation collapses every model, but at a budget comparable to the representation norm a plain graph network also needs, with one exception: OpenGraph, whose spectral tokenizer collapses at a fifth of that budget, an alignment-specific fragility a plain network does not share and which a same-representation control traces to the tokenizer rather than the decoder. A realizable input-space attack that edits edges, features, or text removes at least half the correct predictions on three of the six models at peak. How much of this fragility an input-access attacker realizes tracks how directly the decoder reads the representation, and not the clean accuracy a task leaves; we measure this carrier gain structurally from the decoder's local Lipschitz sensitivity, and report clean-accuracy headroom as a within-model ordering heuristic that does not survive on realizable attacks.

cs.AI↗

arXiv2026

Untangling 3D atomic reconstruction in twisted bilayer 2D crystals via dark field transmission electron microscopy

Reconstruction of the atomic crystal structure in twisted 2D materials has been demonstrated to be responsible for multiple exciting phenomena in van der Waals heterostructures, from the appearance of flat bands in twisted bilayer graphene to Wigner crystallization in transition metal dichalcogenides (TMDs). However, there are still no experimental methods for accessing the 3D atomic distributions nor models that describe the exact atomic shifts in such reconstructed structures, which significantly impedes the development of the field. Dark field (DF) transmission electron microscopy (TEM) has been conventionally employed to visualize the local in-plane atomic displacements. Here we expand this method to obtain a full description of the reconstructed atomic systems and demonstrate the quantitative relations between the local stacking and the intensity in the DF image. We show how local 3D atomic displacements and the interlayer distance can be extracted from a DF image.

cond-mat.mes-hall↗

arXiv2026

Confined exciton polaron in MoS$_2$ on twisted-hBN

The simple electrostatic picture of a trion is that of an excess charge inducing an exciton polarization and binding closer (farther) to the hole (electron) side of it. Trion formation can be forbidden when such spontaneous rearrangement of charges is not allowed by the application of external perturbation, such as electric field. Here we test this hypothesis experimentally using a non-monotonic electric field. We realize this scenario by imprinting the ferroelectric domains at the AA-stacked twisted-hBN (t-hBN) interface onto a monolayer of MoS2 placed over it. The spatially varying in-plane electric field around the domain wall serves the dual purpose of (a) confining and polarizing the 2D exciton in the domain wall, and (b) depleting the free charge carriers from the domain wall. We observe a large quantized exciton splitting confirming strong exciton confinement in the domain wall. Forced by the confining potential, the electron side of the polarized exciton lies closer to the domain with accumulated free electrons, which should ideally prevent any trion formation. Contrary to the laid hypothesis, we observe signatures of quantized charged exciton emission, with an inter-level splitting that mimics the level-splitting of the quantized excitons. This paradox is explained using the many-body picture of exciton polaron, where a conduction band hole attractively binds the polarized exciton and the electron Fermi sea. The results provide a definitive way to unambiguously discern exciton polaron from trion.

cond-mat.mes-hall↗

arXiv2026

On the Parameterized Complexity of Min-Sum-Radii

In the Min-Sum-Radii (MSR) clustering problem, we are given a finite set X of n points in a metric space. The objective is to find at most k clusters centered at a subset of these points such that every point of X is assigned to one of the clusters, minimizing the sum of the radii of the clusters. The problem is known to be NP-hard even on metrics induced by weighted planar graphs and metrics with constant doubling dimension, as shown by Gibson et al. (SWAT 2008). In this work, we investigate the parameterized complexity of MSR on metrics induced by undirected graphs. We distinguish between weighted graph metrics (with positive edge weights) and unweighted graph metrics (where all edges have unit weight). Weighted Graph Metrics: We show that MSR is W[1]-hard on metrics induced by weighted bipartite graphs, when parameterized by the combined parameter k (the number of clusters) and Delta (the cost of the clustering). We then investigate the structural parameterized complexity of the problem. Drexler et al. (arXiv:2310.02130) showed that the MSR problem admits an XP algorithm on metrics induced by weighted graphs when parameterized by treewidth, and asked whether this can be improved to fixed-parameter tractability. We first answer their question in the negative, and more strongly show that MSR stays W[1]-hard on metrics induced by undirected weighted bipartite graphs when parameterized by the vertex cover number plus k. We then turn our attention to parameters for dense graphs and show that MSR remains W[1]-hard when parameterized by k+Delta even on cliques and complete bipartite graphs. On the positive side, we employ the known XP algorithm parameterized by treewidth, to show that the MSR problem is FPT when parameterized by the parameter treewidth plus Delta.

cs.DS↗

arXiv2026

Thermodynamic implications and observational constraints of interacting $f(Q,\mathcal{T})$ gravity in FRW Universe

This work investigates the dynamical evolution of the universe within the framework of symmetric teleparallel $f(Q,\mathcal{T})$ gravity, where $Q$ is the non-metricity scalar and $\mathcal{T}$ is the trace of the energy-momentum tensor. We consider a spatially flat Friedmann-Robertson-Walker (FRW) metric and explore a specific functional form $f(Q,\mathcal{T}) = \alpha Q + \beta \mathcal{T}$ to derive the gravitational field equations. To characterize the late-time cosmic acceleration, we utilize a model-independent approach by adopting a particular Hubble parameter $H(z)$ parametrization. The model parameters are constrained using the latest observational datasets, including the Hubble ($H(z)$) measurements and Pantheon+ samples. Our results indicate a transition from a decelerated to an accelerated expansion phase. We further examine the physical viability of the model through various cosmological diagnostics such as energy density, the equation of state parameter and thermodynamic properties. The analysis demonstrates that $f(Q,\mathcal{T})$ gravity provides a consistent alternative to the $\Lambda$CDM model in explaining the current accelerated expansion of the universe.

gr-qc↗

arXiv2026

Diabetic Retinopathy Classification using Downscaling Algorithms and Deep Learning

Diabetic Retinopathy (DR) is an art and science of recording and classifying the retinal images of a diabetic patient. DR classification deals with classifying retinal fundus image into five stages on the basis of severity of diabetes. One of the major issue faced while dealing with DR classification problem is the large and varying size of images. In this paper we propose and explore the use of several downscaling algorithms before feeding the image data to a Deep Learning Network for classification. For improving training and testing; we amalgamate two datasets: Kaggle and Indian Diabetic Retinopathy Image Dataset. Our experiments have been performed on a novel Multi Channel Inception V3 architecture with a unique self crafted preprocessing phase. We report results of proposed approach using accuracy, specificity and sensitivity, which outperform the previous state of the art methods. Index Terms: Diabetic Retinopathy, Downscaling Algorithms, Multichannel CNN Architecture, Deep Learning

cs.CV↗

arXiv2026

Macroscopic Characteristics of Mixed Traffic Flow with Deep Reinforcement Learning Based Automated and Human-Driven Vehicles

Automated Vehicle (AV) control in mixed traffic, where AVs coexist with human-driven vehicles, poses significant challenges in balancing safety, efficiency, comfort, fuel efficiency, and compliance with traffic rules while capturing heterogeneous driver behavior. Traditional car-following models, such as the Intelligent Driver Model (IDM), often struggle to generalize across diverse traffic scenarios and typically do not account for fuel efficiency, motivating the use of learning-based approaches. Although Deep Reinforcement Learning (DRL) has shown strong microscopic performance in car-following conditions, its macroscopic traffic flow characteristics remain underexplored. This study focuses on analyzing the macroscopic traffic flow characteristics and fuel efficiency of DRL-based models in mixed traffic. A Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm is implemented for AVs' control and trained using the NGSIM highway dataset, enabling realistic interaction with human-driven vehicles. Traffic performance is evaluated using the Fundamental Diagram (FD) under varying driver heterogeneity, heterogeneous time-gap penetration levels, and different shares of RL-controlled vehicles. A macroscopic level comparison of fuel efficiency between the RL-based AV model and the IDM is also conducted. Results show that traffic performance is sensitive to the distribution of safe time gaps and the proportion of RL vehicles. Transitioning from fully human-driven to fully RL-controlled traffic can increase road capacity by approximately 7.52%. Further, RL-based AVs also improve average fuel efficiency by about 28.98% at higher speeds (above 50 km/h), and by 1.86% at lower speeds (below 50 km/h) compared to the IDM. Overall, the DRL framework enhances traffic capacity and fuel efficiency without compromising safety.

cs.AI↗

arXiv2026

Finite Volume Einstein Finsler Warped Product Manifolds of Non-positive or Non-negative Scalar Curvature

The notion of warped product plays an important role in Riemannian geometry moreover in geodesic metric spaces. The warped product was first introduced by Bishop and O'Neill to study Riemannian manifolds of negative curvature.Warped products have been mainly used to construct new examples of Riemannian manifolds with prescribed curvature conditions. This construction can be extended for Finslerian metrics with some minor restrictions. This is motivated by Asanov's papers, where some models of relativity theory are described through the warped product of Finsler metrics. These metrics are in the form of $(\alpha,\beta)$-metrics, which are the generalization of the Randers metrics; which are being asymmetric Finsler metrics in four-dimensional space-time. The product was later extended to the warped product case of Finsler manifolds by the work of Kozma, Peter and Verge.

math.DG↗

arXiv2026

Amalgamated CHIRP and OFDM for ISAC

Integrated Sensing and Communication (ISAC) requires the development of a waveform capable of efficiently supporting both communication and sensing functionalities. This paper proposes a novel waveform that combines the benefits of both the orthogonal frequency division multiplexing (OFDM) and the chirp waveforms to improve both the communication and sensing performance within an ISAC framework. Hence, a new architecture is proposed that utilizes the conventional communication framework while leveraging the parameters sensed at the receiver (Rx) for enhancing the communication performance. We demonstrate that the affine addition of OFDM and chirp signals results in a near constant-envelope OFDM waveform, which effectively reduces the peak-to-average power ratio (PAPR), a key limitation of traditional OFDM systems. Using the OFDM framework for sensing in the conventional fashion requires the allocation of some resources for sensing, which in turn reduces communication performance. As a remedy, the proposed affine amalgam facilitates sensing through the chirp waveform without consuming communication resources, thereby preserving communication efficiency. Furthermore, a novel technique of integrating the chirp signal into the OFDM framework at the slot-level is proposed to enhance the accuracy of range estimation. The results show that the OFDM signal incorporated with chirp has better autocorrelation properties, improved root mean square error (RMSE) of range and velocity, and lower PAPR. Finally, we characterize the trade-off between communications and sensing performance.

eess.SP↗

arXiv2026

Magnetic field decouples nodeless surface and nodal bulk orders in PdTe

Selective spectroscopic disentanglement of surface and bulk quantum orders remains an outstanding challenge in condensed matter physics. The candidate topological superconductor PdTe has recently been proposed to host a nodeless surface gap on top of a nodal bulk state, but their direct identification and mutual coupling remained experimentally elusive. Here, we employ magnetic-field-dependent Andreev reflection spectroscopy to spectroscopically disentangle these components. At zero magnetic field, the spectra exhibit a BCS-like gap structure, consistent with dominant transport through a fully gapped surface superconducting state. Strikingly, even a weak magnetic field leads to an abrupt suppression of the Andreev-enhanced conductance (AEC), while a residual AEC, attributable to the nodal bulk state, persists to much higher magnetic fields. The transition is accompanied by pronounced magnetic hysteresis pointing to the existence of vortex dynamics at low fields. Our findings suggest that the nodal bulk gap facilitates early vortex entry, which in turn disrupts the fragile surface superconductivity. These results establish a field-tunable decoupling of surface and bulk superconductivity and illustrate how distinct gap topologies can shape the global superconducting order in multichannel systems.

cond-mat.supr-con↗

arXiv2026

Interacting Ghost Dark Energy with Sign-Changeable Coupling in Brans-Dicke Cosmology

In this study, we analyze the ghost dark energy model in Brans-Dicke cosmology in the framework of a flat Friedmann-Lemaitre-Robertson-Walker universe. We consider an interaction between ghost dark energy and dark matter with a sign-changeable interaction term. To discuss the cosmological implications of the model, we consider a well-motivated logarithmic form of the Brans-Dicke scalar field. By deriving the cosmological evolution equations, we obtain the cosmological parameters such as the equation of state and deceleration parameters. We analyze the behavior of the cosmological parameters by plotting their graphs against the redshift parameter ($z$). We observe that the equation of state parameter shows quintessence-like behaviour during present and future epochs; however, phantom-like behavior is also possible for suitable values of the model parameters. Analysis of the deceleration parameter shows a smooth recent phase transition of the universe (deceleration to acceleration). An interesting result we observe is the decelerated expansion of the universe in the far future, i.e, the universe experiences another phase transition in the future. The physical significance of the well-known cosmological plane ($w_D-w_D'$ plane) is discussed in our model. We observe that the trajectories start in the freezing region with the same initial behavior, deviate from each other during the evolution and ends in the thawing region. Finally, we perform a detailed thermodynamic analysis and demonstrate that the generalized second law of thermodynamics is satisfied within the present interacting ghost dark energy model.

gr-qc↗

arXiv2025

A heavy-tail arctan-based mixture model for modelling and measuring actuarial risk

Heavy-tailed probability distributions are extremely useful and play a crucial role in modeling different types of financial data sets. This study presents a two-pronged methodology. First, a mixture probability distribution is created by combining Gaussian and Rayleigh distributions using the arctangent transformation, aimed at producing heavier-tailed features and enhancing alignment with real market data. Some statistical properties of the proposed model are also discussed. Furthermore, essential actuarial risk evaluation instruments, such as value-at-risk (VaR), tail value-at-risk (TVaR) and tail variance (TV) are employed for efficient risk management practices. Lastly, an application is provided using an insurance dataset to demonstrate the applicability of the proposed model. The proposed model demonstrates superior fitting performance compared to current baseline distributions, showcasing its practical value in financial risk evaluation. The combination of Gaussian and Rayleigh distributions through arctangent transformation is particularly successful in representing extreme market behaviour and tail dependencies that are frequently found in real-world financial data.

stat.AP↗

arXiv2025

Solar Eruption Onset and Particle Acceleration in Nested-Null Topologies

The magnetic breakout model explains a variety of solar eruptions, ranging from small-scale jets to large-scale coronal mass ejections (CMEs). Most of our previous studies are focused on jets and CMEs in single null-point topologies. Here, we investigate the initiation of CMEs and associated particle acceleration in a double null-point (or nested fan-spine) topology during multiple homologous M- and X-class flares from an active region. The initiation of the flare and associated eruption begins with inflow structures moving towards the inner null of the closed fan-spine topology. The first explosive breakout reconnection of the flux rope at the inner null produced a circular and remote ribbons along with successful eruption of the flux rope and associated fast EUV (shock) wave. Simultaneous flare reconnection beneath the erupting flux rope produced a typical two-ribbon flare along with two hard X-ray footpoint sources. When the flux rope (with shock) reaches the outer null, the second explosive breakout reconnection produces another large-scale remote ribbon. The radio observations reveal quasiperiodic Type III bursts (period=100-s) and a Type II burst during the breakout reconnection near the inner and outer nulls, along with gradual solar energetic particles (SEPs) observed at 1 AU for magnetically connected events. This study highlight the importance of two successive breakout reconnection in the initiation of CMEs in nested-null topologies and associated particle acceleration/release into the interplanetary medium. The particles are accelerated by the shock ahead of the flux rope, which formed during the inner breakout reconnection. These findings have significant implications for particle acceleration and escape processes in multi-scale null-point topologies that produce jets and CMEs.

astro-ph.SR↗

arXiv2025

Imaging and Radio Signatures of Shock-Plasmoid Interaction

Understanding how shocks interact with coronal structures is crucial for understanding the mechanisms of particle acceleration in the solar corona and inner heliosphere. Using simultaneous radio and white-light observations, we investigate the interaction between a CME-driven shock and a plasmoid. LASCO and STEREO-A COR-2 white-light images are analyzed to track the evolution of the plasmoid, CME and its associated shock, while the Wind/WAVES and STEREO/WAVES dynamic spectra provide complementary radio signatures of the shock-plasmoid interaction at $\approx$7 R$_\odot$. An interplanetary Type II radio burst was detected as the shock propagated through the plasmoid. The merging of the plasmoid into the CME was accompanied by interplanetary Type III radio bursts, suggesting escaping electron beams during the reconnection process. These observations clearly demonstrate that shock-plasmoid interactions can enhance the efficiency of particle acceleration associated with CMEs, with implications for electron acceleration in flare and heliospheric current sheets as well.

astro-ph.SR↗