arXiv subjects

Chandan Kumar

Publications and source records attributed to Chandan Kumar.

At least 19 recordsLinked to original sources

arXiv2026

SAFAARI: Schema-Aware Framework for Accelerated Advertiser Response Intelligence

The evolution of customer support systems is rapidly advancing with agentic chatbots, yet these systems face significant limitations when accessing enterprise data without predefined API endpoints. This paper presents SAFAARI (Schema-Aware Framework for Accelerated Advertiser Response Intelligence), a multi-agent framework that addresses the critical bottleneck of schema linking in Natural Language to SQL (NL-to-SQL) systems through specialized content, metadata, and orchestration agents. We also introduce SEAL (Schema Evaluation and Accuracy in Language-to-SQL), a novel composite metric that holistically evaluates system performance while penalizing inconsistent results. Through systematic experimentation with five feature set configurations, SAFAARI achieves an 81.66% SEAL score (6.65% improvement over baseline), with notable gains in datapoint accuracy (5.51%) and schema-linking precision (4.69%). The framework's effectiveness is validated through human-in-the-loop evaluation with domain experts, which proves its adaptability across diverse support domains. By automating the labor-intensive process of schema linking and query generation, our framework demonstrates 8x reduction in development time while maintaining high accuracy. The solution streamlines API development and enhances self-service capabilities, particularly benefiting customer support enterprises with complex data ecosystems.

cs.AI↗

arXiv2026

Microscale bending plasticity and fracture behavior of amorphous aluminum oxide films

Recent work has demonstrated microscale compressive plasticity in pulse laser deposited (PLD) amorphous alumina (a-Al2O3). This work explores microscale bending plasticity and fracture behavior of a-Al2O3 films deposited using three different methods-PLD, atomic layer deposition (ALD) and sputter deposition (SD). The three deposition routes produced amorphous films with similar stoichiometric compositions. We demonstrate, for the first time, bending plasticity in PLD and ALD a-Al2O3 films at microscale using in situ microcantilever bending experiments at room temperature. All tested PLD a-Al2O3 microcantilevers showed substantial ductile behavior in bending by accommodating total strains >10% without fracture. Half of the tested ALD a-Al2O3 cantilevers exhibited elastic brittle fracture while the other half showed bending plasticity, indicating that the observed deformation behavior is strongly influenced by the presence and distribution of defects within the tested volume. All SD a-Al2O3 microcantilevers showed elastic brittle failure attributed to their columnar growth microstructure. The microscale bending response was found to be highly dependent on the film deposition method highlighting the role of defects in suppressing plasticity mechanisms. Notched microcantilever bending tests on all three films showed brittle failure with similar fracture toughness value of 3.1 +/- 0.2 MPa.m0.5, effectively ruling out any localized crack tip plasticity. These findings underscore the importance of minimizing defects during fabrication in order to develop damage tolerant amorphous oxides. Nonetheless, the observation of bending plasticity in both PLD and ALD microcantilevers, which include a tensile stress component as well, suggests that the plastic deformation mechanisms in amorphous alumina are more general and are not exclusively governed by the deposition method.

cond-mat.mtrl-sci↗

arXiv2026

Benchmarking Long Roll-outs of Auto-regressive Neural Operators for the Compressible Navier-Stokes Equations with Conserved Quantity Correction

Deep learning has been proposed as an efficient alternative for the numerical approximation of PDE solutions, offering fast, iterative simulation of PDEs through the approximation of solution operators. However, deep learning solutions have struggle to perform well over long prediction durations due to the accumulation of auto-regressive error, which is compounded by the inability of models to conserve physical quantities. In this work, we present conserved quantity correction, a model-agnostic technique for incorporation physical conservation criteria within deep learning models. Our results demonstrate consistent improvement in the long-term stability of auto-regressive neural operator models, regardless of the model architecture. Furthermore, we analyze the performance of neural operators from the spectral domain, highlighting significant limitations of present architectures. These results highlight the need for future work to consider architectures that place specific emphasis on high frequency components, which are integral to the understanding and modeling of turbulent flows.

cs.LG↗

arXiv2025

Tree Tensor Networks Methods for Efficient Calculation of Molecular Vibrational Spectra

We develop and employ general Tree Tensor Networks (TTNs) to compute the vibrational spectra for two model systems: a set of 64-dimensional coupled oscillators and acetonitrile. We explore various tree architectures, ranging from the simple linear structure of Matrix Product States (MPS), to trees where only the leaf nodes carry a physical leg -- as commonly seen in the underlying ansatz of the Multilayer Multiconfiguration Time-Dependent Hartree (ML-MCTDH) method -- and further to more general trees in which all nodes are allowed to possess a physical leg. In addition, we implement Locally Optimal Block Preconditioned Conjugate Gradient (LOBPCG) methods and Inverse Iteration methods as eigensolvers. Benchmarking runtime and accuracy shows that all tested topologies can reach high accuracy. For acetonitrile, inverse-iteration refinement brings all 84 computed states below 1~cm$^{-1}$ error, while the fork-4 tree, a comb-like tree with four backbone nodes, provides the best overall balance between accuracy and cost. MPS remains computationally attractive, whereas more connected trees generally improve accuracy at fixed bond dimension. All numerical simulations were performed using PyTreeNet, a Python package designed for flexible tensor network computations.

physics.chem-ph↗

arXiv2025

Stage Aware Diagnosis of Diabetic Retinopathy via Ordinal Regression

Diabetic Retinopathy (DR) has emerged as a major cause of preventable blindness in recent times. With timely screening and intervention, the condition can be prevented from causing irreversible damage. The work introduces a state-of-the-art Ordinal Regression-based DR Detection framework that uses the APTOS-2019 fundus image dataset. A widely accepted combination of preprocessing methods: Green Channel (GC) Extraction, Noise Masking, and CLAHE, was used to isolate the most relevant features for DR classification. Model performance was evaluated using the Quadratic Weighted Kappa, with a focus on agreement between results and clinical grading. Our Ordinal Regression approach attained a QWK score of 0.8992, setting a new benchmark on the APTOS dataset.

cs.CV↗

arXiv2025

Exploring AI in Steganography and Steganalysis: Trends, Clusters, and Sustainable Development Potential

Steganography and steganalysis are strongly related subjects of information security. Over the past decade, many powerful and efficient artificial intelligence (AI) - driven techniques have been designed and presented during research into steganography as well as steganalysis. This study presents a scientometric analysis of AI-driven steganography-based data hiding techniques using a thematic modelling approach. A total of 654 articles within the time span of 2017 to 2023 have been considered. Experimental evaluation of the study reveals that 69% of published articles are from Asian countries. The China is on top (TP:312), followed by India (TP-114). The study mainly identifies seven thematic clusters: steganographic image data hiding, deep image steganalysis, neural watermark robustness, linguistic steganography models, speech steganalysis algorithms, covert communication networks, and video steganography techniques. The proposed study also assesses the scope of AI-steganography under the purview of sustainable development goals (SDGs) to present the interdisciplinary reciprocity between them. It has been observed that only 18 of the 654 articles are aligned with one of the SDGs, which shows that limited studies conducted in alignment with SDG goals. SDG9 which is Industry, Innovation, and Infrastructure is leading among 18 SDGs mapped articles. To the top of our insight, this study is the unique one to present a scientometric study on AI-driven steganography-based data hiding techniques. In the context of descriptive statistics, the study breaks down the underlying causes of observed trends, including the influence of DL developments, trends in East Asia and maturity of foundational methods. The work also stresses upon the critical gaps in societal alignment, particularly the SDGs, ultimately working on unveiling the field's global impact on AI security challenges.

cs.CR↗

arXiv2025

Supervised Multilabel Image Classification Using Residual Networks with Probabilistic Reasoning

Multilabel image categorization has drawn interest recently because of its numerous computer vision applications. The proposed work introduces a novel method for classifying multilabel images using the COCO-2014 dataset and a modified ResNet-101 architecture. By simulating label dependencies and uncertainties, the approach uses probabilistic reasoning to improve prediction accuracy. Extensive tests show that the model outperforms earlier techniques and approaches to state-of-the-art outcomes in multilabel categorization. The work also thoroughly assesses the model's performance using metrics like precision-recall score and achieves 0.794 mAP on COCO-2014, outperforming ResNet-SRN (0.771) and Vision Transformer baselines (0.785). The novelty of the work lies in integrating probabilistic reasoning into deep learning models to effectively address the challenges presented by multilabel scenarios.

cs.CV↗

arXiv2025

DeiTFake: Deepfake Detection Model using DeiT Multi-Stage Training

Deepfakes are major threats to the integrity of digital media. We propose DeiTFake, a DeiT-based transformer and a novel two-stage progressive training strategy with increasing augmentation complexity. The approach applies an initial transfer-learning phase with standard augmentations followed by a fine-tuning phase using advanced affine and deepfake-specific augmentations. DeiT's knowledge distillation model captures subtle manipulation artifacts, increasing robustness of the detection model. Trained on the OpenForensics dataset (190,335 images), DeiTFake achieves 98.71\% accuracy after stage one and 99.22\% accuracy with an AUROC of 0.9997, after stage two, outperforming the latest OpenForensics baselines. We analyze augmentation impact and training schedules, and provide practical benchmarks for facial deepfake detection.

cs.CV↗

arXiv2025

Quantum non-Gaussian high Fock states of light pulses and their superpositions

The generation of high Fock states of light pulses and their superpositions with provable quantum non-Gaussian features is still very challenging, although the power of conditional methods to herald the approximate state from the available Gaussian states is growing. The atom-light interaction in the high-Q cavity has been considered a viable alternative to the heralded Fock states of the light pulses from nonlinear optics limited to three-photon Fock states for the last decade. Here, by optimizing the realistic protocol combining it with available optical delay elements we conclusively predict filtering of Fock states up to ten photons with a high success rate of $20\%$ using a hierarchy of quantum non-Gaussian criteria. Moreover, the filtering protocol enables the preparation of superposition of Fock states and we analyse this emerging case up to two photons with with provable quantum non-Gaussian coherence and high success rate $50\%$. To demonstrate their quality for applications, we evaluate the robustness of such features, the bunching capability in a linear network, and the sensing capability to estimate the magnitude of unknown force, noise and phase, These assessments outline the essential conditions and application criteria for realistic optical cavity QED interaction on light pulses to outperform photon detection methods.

quant-ph↗

arXiv2025

Probing Phonon Modes in Reconstructed twisted Homo and Hetero Bilayer System

Twist angle engineering in van der Waals homo and hetero-bilayers introduces profound modifications in their electronic, optical and mechanical properties due to lattice reconstruction. In these systems, the interlayer coupling and atomic rearrangement strongly depend on the twist angle, leading to the formation of periodic Moire superlattices. At small twist angles, significant lattice relaxation results in the emergence of domain structures separated by one dimensional soliton networks, influencing electronic band structures and phonon modes. Here we systematically investigate the impact of lattice reconstruction on phonon renormalization in twisted bilayer graphene (TBLG,homo) and graphene-hBN Moire superlattices(hetero). Using Raman spectroscopy, we identify distinct phonon behaviours across different twist angle regimes. In TBLG, we observe the evolution of the G peak, including broadening, splitting, and the emergence of additional peaks in the small angle range 0.3 to 1 degree, attributed to Moire modified phonon interactions. At large twist angles, the peaks gradually merge back into a single feature, reflecting the reduced impact of lattice reconstruction. Similarly, in hBN graphene Moire superlattices, we detect Moire induced Raman peaks above and below the G peak, while the central G peak remains largely invariant to twist angle variation. The theoretical calculations uncover Moire phonon modes originating from different stacking regions providing insights into phonon renormalization. Our results establish a direct link between twist angle, lattice reconstruction, Moire phonons, and interlayer coupling, offering a fundamental framework for understanding phonon engineering in twisted bilayer systems. These findings pave the way for controlling phononic, optoelectronic and heat flow properties in next generation van der Waals heterostructures.

cond-mat.mes-hall↗

arXiv2025

Exceptional point in a PT symmetric non-Hermitian terahertz plasmonic metasurface

In this paper, we experimentally demonstrate a non-Hermitian open PT-symmetric terahertz metasurface comprising complementary plasmonic structures capable of exhibiting an exceptional point (EP). The metasurface consists of two resonators of different sizes, representing effective gain and loss elements, placed orthogonally in close proximity to realize a non-Hermitian configuration leading to a PT symmetry state. A diagonal displacement of one resonator within this strongly coupled near-field configuration leads to the emergence of an exceptional point, where the system undergoes a sudden phase transition from a PT symmetric to a PT-asymmetric state. Terahertz time-domain spectroscopy (THz-TDS) is performed on the fabricated samples to experimentally validate the phase transition observed in numerical simulations. We employ coupled mode theory (CMT) to analyse and distinguish between the PT-symmetric, exceptional point, and PT-asymmetric states. This theoretical framework enables the calculation of eigenvalues, phase spectra, and eigenmodes associated with the metamaterial design, thereby corroborating the simulation results. Furthermore, we construct Poincare sphere to visualize the orientation of the polarization states of the eigenmodes, which further indicates the presence of the exceptional point. This comprehensive study of exceptional point in a plasmonic system holds potential for the development of practical, highly sensitive terahertz devices, addressing limitations of conventional PT-symmetric systems that rely on traditional gain and loss media.

physics.optics↗

arXiv2025

Quantitative agreement between experiment and theory for Vibrational Circular Dichroism enhanced by electronically excited states

Intensity enhancement in vibrational circular dichroism (VCD) arises in open-shell transition metal complexes from coupling between ground-state vibrational transitions and magnetic dipole-allowed transitions to low-lying excited states (LLESs). In this work we apply Nafie's vibronic coupling theory to M(II)-(-)-sparteine-Cl$_2$ (M=Zn, Co, Ni) complexes to investigate these enhancement effects. We show that the VCD intensity is extremely sensitive to the excitation energies that neither time-dependent density functional theory (TDDFT) nor state-averaged complete active space self consistent field (SA-CASSCF) calculations can predict with sufficient accuracy. We argue that instead of using more accurate quantum chemistry methods these excitation energies can be treated as parameters and optimized against experimental spectra. With this approach we obtain simulated VCD similarity scores above 0.4, a threshold considered reliable for absolute configuration assignment. The ability to quantitatively reproduce enhanced experimental spectra with computations opens up new research areas, offering amongst else unique possibilities for the study of chiral structure of systems such as transition metal complexes and metalloproteins that so far remained intractable.

physics.chem-ph↗

arXiv2025

Magnetic Charge State Controlled Spin-Wave Dynamics in Nanoscale Three-Dimensional Artificial Spin Ice

Three-dimensional (3D) magnetic nanostructures offer a versatile platform for exploring complex spin textures and spin-wave (SW) dynamics, with implications in next-generation spintronic and magnonic technologies. Advances in 3D nanofabrication have allowed a wide-range of structures and phenomena to be realized. Whilst the study of simple cylindrical magnetic nanowires allows the realization of ultrafast domain walls and a spin Cherenkov effect, placing such wires of complex cross-section into 3D arrangements allows one to produce magnetic metamaterials, known as artificial spin-ice (ASI), where the overall ground state and spin dynamics are governed by magnetostatic interactions between elements. Here, using Brillouin Light Scattering (BLS) we demonstrate the direct detection of magnetic charged states in a 3D-ASI system. The measured spin-wave modes in 3D-ASI are found to be directly controlled by the local magnetic charge configuration and the direction of the applied magnetic field. Micromagnetic simulations provide insight into the spatially selective excitation of spin waves and the evolution of magnetic microstates, uncovering a direct link to the field-dependent characteristics of the spin-wave spectrum. These findings make 3D-ASI architectures a promising system to realize reconfigurable, low-power magnonic devices with engineered collective dynamics.

cond-mat.mes-hall↗

arXiv2025

Husimi phase distribution in non-Gaussian operations

The Husimi phase distribution, an experimentally measurable quantity, is investigated for single-mode and two-mode squeezed vacuum states. The analysis highlights that non-Gaussian operations, i.e., photon subtraction (PS), photon addition (PA) and photon catalysis (PC), are effective tools for localizing phase distribution and enhancing phase robustness in the presence of noise, while PC enhances phase sensitivity but leads to greater delocalization. The work highlights the perspective that combined effects of squeezing, beam splitter transmittance, and environmental interactions must be carefully considered when quantum state engineering protocols are designed and phase properties provide a valuable insight into this endeavour.

quant-ph↗

arXiv2025

Continuous variable quantum teleportation, $U(2)$ invariant squeezing and non-Gaussian resource states

We investigate the role of quadrature squeezing in the quantum teleportation protocol for coherent states, using non-Gaussian resource states. For the two-mode systems, the non-Gaussian resource states that we use are obtained by an experimentally realizable scheme of photon subtraction, photon addition, and photon catalysis, on the two-mode squeezed vacuum, and two-mode squeezed thermal states. We first analyze the non-classical attribute of quadrature squeezing in these generated non-Gaussian states using the $U(2)$ invariant squeezing approach, which allows us to account for all possible quadratures. We then show that the presence of such non-classicality in non-Gaussian resource states is not necessary for successful quantum teleportation, a finding which is at variance with an earlier result in this direction. This result is important since it demonstrates how non-classicality other than quadrature squeezing present in the resource can be utilized for quantum teleportation.

quant-ph↗

arXiv2025

Provisioning Time-Based Subscription in NDN: A Secure and Efficient Access Control Scheme

This paper proposes a novel encryption-based access control mechanism for Named Data Networking (NDN). The scheme allows data producers to share their content in encrypted form before transmitting it to consumers. The encryption mechanism incorporates time-based subscription access policies directly into the encrypted content, enabling only consumers with valid subscriptions to decrypt it. This makes the scheme well-suited for real-world, subscription-based applications like Netflix. Additionally, the scheme introduces an anonymous and unlinkable signature-based authentication mechanism that empowers edge routers to block bogus content requests at the network's entry point, thereby mitigating Denial of Service (DoS) attacks. A formal security proof demonstrates the scheme's resistance to Chosen Plaintext Attacks (CPA). Performance analysis, using Mini-NDN-based emulation and a Charm library implementation, further confirms the practicality of the scheme. Moreover, it outperforms closely related works in terms of functionality, security, and communication overhead.

cs.CR↗

arXiv2024

Comparing on-off detector and single photon detector in photon subtraction based continuous variable quantum teleportation

We consider here two distinct photon detectors namely, single photon detector and on-off detector, to implement photon subtraction on a two-mode squeezed vacuum (TMSV) state. The two distinct photon subtracted TMSV states generated are utilized individually as resource states in continuous variable quantum teleportation. Owing to the fact that the two generated states have different success probabilities (of photon subtraction) and fidelities (of quantum teleportation), we consider the product of the success probability and fidelity enhancement as a figure of merit for the comparison of the two detectors. The results show that the single photon detector should be preferred over the on-off detector for the maximization of the considered figure of merit.

quant-ph↗

arXiv2024

Quantum Computing for Automotive Applications

Quantum computing could impact various industries, with the automotive industry with many computational challenges, from optimizing supply chains and manufacturing to vehicle engineering, being particularly promising. This chapter investigates state-of-the-art quantum algorithms to enhance efficiency, accuracy, and scalability across the automotive value chain. We explore recent advances in quantum optimization, machine learning, and numerical and chemistry simulations, highlighting their potential and limitations. We identify and discuss key challenges in near-term and fault-tolerant algorithms and their practical use in industrial applications. While quantum algorithms show potential in many application domains, current noisy intermediate-scale quantum hardware limits scale and, thus, business benefits. In the long term, fault-tolerant systems promise theoretical speedups; however, they also require further progress in hardware and software (e.g., related to error correction and data loading). We expect that with this progress, significant practical benefits will emerge eventually.

quant-ph↗