arXiv subjects

Kunal Kumar

Publications and source records attributed to Kunal Kumar.

At least 19 recordsLinked to original sources

arXiv2026

Broken-Symmetry Reference States for Sample-Based Quantum Diagonalization with Near-DMRG Accuracy

Sample-based quantum diagonalization (SQD) accuracy remains limited for strongly correlated systems like iron-sulfur clusters. This accuracy heavily depends on the mean-field reference state, which determines the sampling ansatz and configuration-recovery target. We extend the qcsc-prefect workflow to support a broken-symmetry unrestricted Hartree-Fock (BS-UHF) starting point, derived from atom-localized antiferromagnetic initial guesses. For the $[Fe_4S_4(SCH_3)_4]^{2-}$ cubane (54 electrons, 36 orbitals, 72 data qubits; 88 total qubits), this lowers the mean-field reference by $\approx 534$ mHa relative to the Restricted Hartree-Fock (RHF) solution (-327.0809 vs. -326.5474 Ha). Combined with deep multi-step configuration recovery on ROQUO GPU nodes at subspace dimension $d = 3 \times 10^{10}$, we reach $E = -327.235076$ Ha. This is 4.3 mHa above the finite-$D = 4000$ DMRG value (-327.2393952 Ha) and $\sim 512-517$ mHa below a previous same-active-space RHF-reference SQD calculation. Extrapolating to the $d \to \infty$ limit via a Gaussian-in-log $d$ fitting ansatz gives an asymptotic estimate $E_\infty = -327.244468$ Ha $\pm 1.05$ mHa, just 0.33 mHa above our discarded-weight-extrapolated DMRG estimate ($E_{D\to\infty} = -327.244797$ Ha $\pm 0.46$ mHa), well within their combined $1σ$ uncertainty. We characterize the determinant-subspace composition driving convergence, show matching validation on $[Fe_2S_2(SCH_3)_4]^{2-}$, and report multi-HPC execution on Fugaku, Miyabi-G, and ROQUO. We conclude that reference-state preparation, rather than subspace size alone, is a key driver of SQD energy accuracy under strong electronic correlation.

quant-ph↗

arXiv2026

Vibrational Sample-Based Quantum Diagonalization with Unitary-Cluster-Jastrow Ansätze on IBM QPUs

We extend sample-based quantum diagonalization (SQD) to anharmonic vibrational spectra and use it to benchmark how vibrational ansätze scale on present-day hardware under a direct one-hot encoding (up to 72 qubits). Five particle-conserving circuits are warm-started through vibrational coupled-cluster (VCC) t1, t2 amplitudes: UVCCSD, the compact heuristic circuit (CHC), and three unitary-cluster-Jastrow ansätze introduced here, VLUCJ, Vg-uCJ, and VIm-uCJ. A self-consistent configuration-recovery loop reconstructs correlated states from hardware samples in the low-retention regime where naïve post-selection breaks down, enabling the calculation of ground-state energies for H$_2$O, CH$_2$O, and CH$_2$ClF (3-9 modes). We compare harmonic-oscillator and vibrational self-consistent-field (VSCF, "modal") reference bases. In the small-modal-basis regime relevant for near-term devices, the modal basis is typically more accurate than the harmonic-oscillator representation; empirically, it also yields substantially better-converged hardware sampling for deep circuits at these qubit counts. Across ansätze, we find a clear retention/accuracy tradeoff when scaling: the new VLUCJ circuit is the most hardware-robust, retaining the largest fraction of physically valid one-hot outcomes, whereas the more expressive uCJ variants and UVCCSD can recover more correlation when retention is sufficient. Finally, we extend the recovery protocol to excited states by targeting a chosen per-mode VSCF occupation and warm-starting it with a state-specific VCC calculation, and we demonstrate the calculation of one representative fundamental vibrational frequency per molecule on IBM QPUs.

quant-ph↗

arXiv2026

A Reynolds-like Criterion for the Onset of Universal Active Turbulence

Beyond a critical Toner--Tu drive, bacterial turbulence exhibits universal statistics, including a parameter-independent energy spectrum $E(k)\sim k^{-3/2}$. We show that the same universal state can be reached not only by increasing activity, but also by strengthening nonlinear advection, weakening nonlinear saturation, or tuning the Swift--Hohenberg instability. Motivated by the competition between these processes, we identify the independent dimensionless groups governing the dynamics and construct a Reynolds-like control parameter from them. A single critical value, calibrated from the known activity-controlled transition, predicts the onset of universality along the other parameter routes and collapses them onto a common threshold. These results identify a dynamical organizing parameter for the onset of universal active turbulence and provide a Reynolds-like measure for the Toner--Tu--Swift--Hohenberg universality class.

physics.flu-dyn↗

arXiv2026

Geometry, elasticity, and activity in the transport of self-propelled filaments in turbulence

We investigate the transport of elastic active filaments in two-dimensional turbulence, focusing on how propulsion geometry and elasticity determine vortex trapping and transport. Using a bead-spring model with activity applied at the filament head, we compare propulsion that follows the instantaneous filament conformation with propulsion imposed along a fixed external direction. We find that activity does not generically enhance transport: when propulsion remains coupled to the filament backbone, vortex trapping remains dominant and motion stays effectively diffusive, whereas fixed-direction propulsion enables persistent excursions across flow structures and leads to superdiffusive transport. In both cases, activity shifts filament conformations toward more extended states, effectively opposing elastic relaxation without eliminating preferential sampling of coherent vortical regions. At low Weissenberg number, this conformational change is amplified: activity cooperates with elasticity to enhance preferential sampling of vortical regions and strengthen vortex trapping. Transport therefore emerges from a competition between activity, elasticity, and flow-induced deformation, with elasticity determining how effectively activity-induced extensions can persist against turbulent trapping. These results establish propulsion geometry as the key control parameter for transport, with elasticity and activity acting cooperatively rather than independently to shape filament dynamics in turbulent flows.

physics.flu-dyn↗

arXiv2026

Turbulence-Induced Fluctuating Interfaces in Heterogeneously-Active Suspensions

We investigate the effects of heterogeneous (spatially varying) activity in a hydrodynamical model for dense bacterial suspensions, confining ourselves to experimentally realizable, simple, quenched, activity patterns. We show that the evolution of the bacterial velocity field under such activity patterning leads to the emergence of hydrodynamic interfaces separating spatially localized turbulence from jammed frictional surroundings. We characterise the intermittent and multiscale fluctuations of this interface and also investigate how heterogeneity influences mixing via the residence times of Lagrangian tracers. This work reveals how naturally occurring heterogeneities could decisively steer active flows into more complex configurations than those typically studied, opening up parallels to droplet dynamics, front propagation and turbulent mixing layers.

cond-mat.soft↗

arXiv2026

A Mixture-of-Experts Framework for Practical Hybrid-Quantum Models in Credit Card Fraud Detection

This paper investigates whether hybrid quantum-classical machine learning can deliver practical improvements in financial fraud detection performance for card-based and other payment transactions. Building on a Guided Quantum Compressor architecture, the approach integrates an autoencoder, a variational quantum circuit, and a classical neural head, and then embeds this hybrid model into a mixture-of-experts framework including a state-of-the-art gradient-boosted tree classifier. Using a European credit card dataset with severe class imbalance, the routed hybrid architecture with 0.6 threshold achieves average precision scores of $0.793\pm0.085$ compared to $0.770\pm0.096$ of XGBoost on 3 repeated 5-fold cross-validation benchmarks. Precision and recall comparisons reveals a possible trade-off of fraud and nominal detections with a reduction in false positives at the cost of a small reduction in fraud detections. The improvements are achieved while adding only 7 to 21 minutes of extra inference time depending on the choice of hyperparameters. These results indicate that selectively routing transactions to quantum-classical models can enhance fraud detection while remaining compatible with the latency and operational constraints of modern financial institutions.

quant-ph↗

arXiv2026

DeepContext: Stateful Real-Time Detection of Multi-Turn Adversarial Intent Drift in LLMs

While Large Language Model (LLM) capabilities have scaled, safety guardrails remain largely stateless, treating multi-turn dialogues as a series of disconnected events. This lack of temporal awareness facilitates a "Safety Gap" where adversarial tactics, like Crescendo and ActorAttack, slowly bleed malicious intent across turn boundaries to bypass stateless filters. We introduce DeepContext, a stateful monitoring framework designed to map the temporal trajectory of user intent. DeepContext discards the isolated evaluation model in favor of a Recurrent Neural Network (RNN) architecture that ingests a sequence of fine-tuned turn-level embeddings. By propagating a hidden state across the conversation, DeepContext captures the incremental accumulation of risk that stateless models overlook. Our evaluation demonstrates that DeepContext significantly outperforms existing baselines in multi-turn jailbreak detection, achieving a state-of-the-art F1 score of 0.84, which represents a substantial improvement over both hyperscaler cloud-provider guardrails and leading open-weight models such as Llama-Prompt-Guard-2 (0.67) and Granite-Guardian (0.67). Furthermore, DeepContext maintains a sub-20ms inference overhead on a T4 GPU, ensuring viability for real-time applications. These results suggest that modeling the sequential evolution of intent is a more effective and computationally efficient alternative to deploying massive, stateless models.

cs.AI↗

arXiv2025

A Narrative Review of Clinical Decision Support Systems in Offloading Footwear for Diabetes-Related Foot Ulcers

Offloading footwear helps prevent and treat diabetic foot ulcers (DFUs) by lowering plantar pressure (PP), yet prescription decisions remain fragmented: feature selection varies, personalization is limited, and evaluation practices differ. We performed a narrative review of 45 studies (12 guidelines/protocols, 25 knowledge-based systems, 8 machine-learning applications) published to Aug 2025. We thematically analyzed knowledge type, decision logic, evaluation methods, and enabling technologies. Guidelines emphasize PP thresholds (<=200 kPa or >=25--30\% reduction) but rarely yield actionable, feature-level outputs. Knowledge-based systems use rule- and sensor-driven logic, integrating PP monitoring, adherence tracking, and usability testing. ML work introduces predictive, optimization, and generative models with high computational accuracy but limited explainability and clinical validation. Evaluation remains fragmented: protocols prioritize biomechanical tests; knowledge-based systems assess usability/adherence; ML studies focus on technical accuracy with weak linkage to long-term outcomes. From this synthesis we propose a five-part CDSS framework: (1) a minimum viable dataset; (2) a hybrid architecture combining rules, optimization, and explainable ML; (3) structured feature-level outputs; (4) continuous validation and evaluation; and (5) integration with clinical and telehealth workflows. This framework aims to enable scalable, patient-centered CDSSs for DFU care; prioritizing interoperable datasets, explainable models, and outcome-focused evaluation will be key to clinical adoption.

cs.LG↗

arXiv2025

Integrated Airline Fleet and Crew Recovery through Local Search

Airline operations are prone to delays and disruptions, since the schedules are generally tight and depend on a lot of resources. When disruptions occur, the flight schedule needs to be adjusted such that the operation can continue. Since this happens during the day of operations, this needs to be done as close to real time as possible, posing a challenge with respect to computation time. Moreover, to limit the impact of disruptions, we want a solution with minimal cost and passenger impact. Since airline operations include many interlinked decisions, an integrated approach leads to better overall solutions. We specifically look at resolving these disruptions in both the aircraft and crew schedules. Resolving these disruptions is complex, especially when it is done in an integrated way, i.e. including multiple different resources. To solve this problem in an integrated manner, we developed a fast simulated annealing approach. To the best of our knowledge, we are the first to develop a local search approach to resolve airline disruptions in an integrated way. This approach is compared with traditional approaches, and an experimental study is done to evaluate different neighbour generation methods, and to investigate different recovery scenarios and strategies. The comparison is done using real world data from KLM Royal Dutch Airlines. Here, we show that our approach resolves disruptions quickly and in a cost-efficient manner, and that it outperforms traditional approaches. Compared to naive delay propagation, our method saves 40% in non-performance costs. Moreover, while most airlines use tools that consider resources separately, our approach shows that integrated disruption management is possible within 30 seconds.

math.OC↗

arXiv2025

Fringe around a Beet Slice: Wetting-induced Dimple in a Thin Liquid Film

When a slice of beet is placed on a plate with a thin layer of beet juice, one can observe a clear fringe around the beet, where the color is more translucent than the rest of the juice. The hypotheses in literature were inconsistent and limited, which motivated us to revisit this phenomenon. Using a motorized confocal displacement sensor, we measured the temporal evolution of the liquid surface profile across the fringe. Our findings suggest that a suction flow, induced by the capillary rise of the contact line, causes a dimple - a small concave depression - to form on the liquid surface. While surface tension and gravity tends to smooth out the dimple, viscous drag acts against them if the liquid film is sufficiently thin. Our scaling analysis correctly estimates the dependence of dimple lifetime on liquid properties and film thickness. We also capture the dimple formation dynamics by numerically solving the lubrication equation with the Young-Laplace equation. This work provides a new interpretation for a common phenomenon.

cond-mat.soft↗

arXiv2025

Onset of Intermittency and Multiscaling in Active Turbulence

Recent results suggest that highly active, chaotic, non-equilibrium states of living fluids might share much in common with high Reynolds number, inertial turbulence. We now show, by using a hydrodynamical model, the onset of intermittency and the consequent multiscaling of Eulerian and Lagrangian structure functions as a function of the bacterial activity. Our results bridge the worlds of low and high Reynolds number flows as well as open up intriguing possibilities of what makes flows intermittent.

cond-mat.soft↗

arXiv2014

GMCALC: a calculator for the Georgi-Machacek model

The Georgi-Machacek model adds scalar triplets to the Standard Model Higgs sector in such a way as to preserve custodial SU(2) symmetry in the scalar potential. This allows the triplets to have a non-negligible vacuum expectation value while satisfying constraints from the rho parameter. Depending on the parameters, the 125~GeV neutral Higgs particle can have couplings to WW and ZZ larger than in the Standard Model due to mixing with the triplets. The model also contains singly- and doubly-charged Higgs particles that couple to vector boson pairs at tree level (WZ and like-sign WW, respectively). GMCALC is a self-contained FORTRAN program that, given a set of input parameters, calculates the particle spectrum and tree-level couplings in the Georgi-Machacek model, checks theoretical and indirect constraints, and computes the branching ratios and total widths of the scalars. It also generates a param_card.dat file for MadGraph5 to be used with the corresponding FeynRules model implementation.

hep-ph↗

arXiv2014

Indirect constraints on the Georgi-Machacek model and implications for Higgs couplings

We update the indirect constraints on the Georgi-Machacek model from $B$-physics and electroweak precision observables, including new constraints from $b \to s γ$ and $B^0_s \to μ^+ μ^-$. We illustrate the effect of these constraints on the couplings of the Standard Model-like Higgs boson by performing scans using the most general scalar potential, subject to vacuum stability and perturbativity constraints. We find that simultaneous enhancements of all the Higgs production cross sections by up to 39\% are still allowed after imposing these constraints. LHC rate measurements on the Higgs pole could be blind to these enhancements if unobserved non-standard Higgs decays are present.

hep-ph↗

arXiv2014

The decoupling limit in the Georgi-Machacek model

We study the most general scalar potential of the Georgi-Machacek model, which adds isospin-triplet scalars to the Standard Model (SM) in a way that preserves custodial SU(2) symmetry. We show that this model possesses a decoupling limit, in which the predominantly-triplet states become heavy and degenerate while the couplings of the remaining light neutral scalar approach those of the SM Higgs boson. We find that the SM-like Higgs boson couplings to fermion pairs and gauge boson pairs can deviate from their SM values by corrections as large as $\mathcal{O}(v^2/M_{\rm new}^2)$, where $v$ is the SM Higgs vacuum expectation value and $M_{\rm new}$ is the mass scale of the predominantly-triplet states. In particular, the SM-like Higgs boson couplings to $W$ and $Z$ boson pairs can decouple much more slowly than in two Higgs doublet models, in which they deviate from their SM values like $\mathcal{O}(v^4/M_{\rm new}^4)$. Furthermore, near the decoupling limit the SM-like Higgs boson couplings to $W$ and $Z$ pairs are always larger than their SM values, which cannot occur in two Higgs doublet models. As such, a precision measurement of Higgs couplings to $W$ and $Z$ pairs may provide an effective method of distinguishing the Georgi-Machacek model from two Higgs doublet models. Using numerical scans, we show that the coupling deviations can reach 10% for $M_{\rm new}$ as large as 800~GeV.

hep-ph↗

arXiv2013

Learning what the Higgs boson is mixed with

The Standard Model Higgs boson may be mixed with another scalar that does not couple to fermions. The electroweak quantum numbers of such an additional scalar can be determined by measuring the quartic Higgs-Higgs-vector-vector couplings, which contribute---along with the coveted triple Higgs coupling---to double Higgs production in $e^+e^-$ collisions. We show that simultaneous sensitivity to the quartic Higgs-Higgs-vector-vector coupling and the triple Higgs coupling can be obtained using measurements of the double Higgs production cross section at two different e+e- center-of-mass energies. Kinematic distributions of the two Higgs bosons in the final state could provide additional discriminating power.

hep-ph↗

arXiv2013

Effects from New Colored States and the Higgs Portal on Gluon Fusion and Higgs Decays

We study effects from new colored states and the Higgs portal on gluon fusion production. We isolate possible loop contributions from new colored scalars, fermions, and vectors, incorporating effects from Higgs portal-induced scalar mixing, leading to dramatic effects on gluon fusion and branching fractions. Higgs identification must generally allow for these effects, and using our results, possible tensions from fits to the Standard Model expectation can be relieved by inclusion of New Physics effects.

hep-ph↗

arXiv2012

Can the 126 GeV boson be a pseudoscalar?

We test the possibility that the newly-discovered 126 GeV boson is a pseudoscalar by examining the correlations among the loop-induced pseudoscalar decay branching fractions to $γγ$, $ZZ^*$, $Zγ$, and $WW^*$ final states in a model-independent way. These four decays are controlled by only two effective operators, so that the rates in $Zγ$ and $WW^*$ are predicted now that the rates in $γγ$ and $ZZ^*,Zγ^* \to 4 \ell$ have been measured. We find that the pseudoscalar possibility is disfavored but not conclusively excluded. Experimental exclusion of the $Zγ$ decay to well below $σ/σ_{\rm SM} \sim 170$ or conclusive observation of the $WW^*$ decay near the Standard Model rate would eliminate the pseudoscalar possibility. The $Zγ$ exclusion should be possible using existing data. The only loophole in our argument is the possibility that the $4\ell$ signal comes from pseudoscalar decays to a pair of new neutral gauge bosons with mass near the $Z$ pole.

hep-ph↗

arXiv2012

Magnetic Fluffy Dark Matter

We explore extensions of inelastic Dark Matter and Magnetic inelastic Dark Matter where the WIMP can scatter to a tower of heavier states. We assume a WIMP mass $m_χ\sim \mathcal{O}(1-100)$ GeV and a constant splitting between successive states $δ\sim\mathcal{O}(1 - 100)$ keV. For the spin-independent scattering scenario we find that the direct experiments CDMS and XENON strongly constrain most of the DAMA/LIBRA preferred parameter space, while for WIMPs that interact with nuclei via their magnetic moment a region of parameter space corresponding to $m_χ\sim 11 $ GeV and $δ< 15$ keV is allowed by all the present direct detection constraints.

hep-ph↗