arXiv subjects

Anirudha Ghosh

Publications and source records attributed to Anirudha Ghosh.

11 recordsLinked to original sources

arXiv2026

Coherent spin waves in a maximal entropy phase

In solids, disorder is conventionally regarded as detrimental to coherence. It typically localizes and dampens collective excitations, as exemplified by Anderson localization or the broadening of magnetic modes in systems lacking long-range order. While high-entropy materials are specifically designed to harness disorder and stabilize homogeneous mixed-phase structures that can display unique properties, this same disorder is nonetheless expected to preclude the formation of coherent magnetic excitations. To test the limits of this picture, we selected the antiferromagnetic system YBaCuFeO5, as it features two distinct transition metal atoms with significantly different magnetic moments, rendering its spin dynamics exceptionally sensitive to local atomic ordering. Combining resonant inelastic x-ray scattering and linear spin wave theory, we reveal a surprising paradox: YBaCuFeO5 exhibits an unexpected, entropy-driven mixed phase, in which disorder, rather than reducing the lifetime of the collective excitations, favors coherence. In this mixed phase, the spin waves remain dispersive, markedly distinct from those expected for an ordered ground state, and exhibit well-defined acoustic and optical branches separated by a large optical gap. These results demonstrate that in entropy-stabilized magnets, disorder can favor coherent collective modes previously thought to be exclusive to low-entropy systems.

cond-mat.str-el↗

arXiv2026

Context-Aware Pesticide Recommendation via Few-Shot Pest Recognition for Precision Agriculture

Effective pest management is crucial for enhancing agricultural productivity, especially for crops such as sugarcane and wheat that are highly vulnerable to pest infestations. Traditional pest management methods depend heavily on manual field inspections and the use of chemical pesticides. These approaches are often costly, time-consuming, labor-intensive, and can have a negative impact on the environment. To overcome these challenges, this study presents a lightweight framework for pest detection and pesticide recommendation, designed for low-resource devices such as smartphones and drones, making it suitable for use by small and marginal farmers. The proposed framework includes two main components. The first is a Pest Detection Module that uses a compact, lightweight convolutional neural network (CNN) combined with prototypical meta-learning to accurately identify pests even when only a few training samples are available. The second is a Pesticide Recommendation Module that incorporates environmental factors like crop type and growth stage to suggest safe and eco-friendly pesticide recommendations. To train and evaluate our framework, a comprehensive pest image dataset was developed by combining multiple publicly available datasets. The final dataset contains samples with different viewing angles, pest sizes, and background conditions to ensure strong generalization. Experimental results show that the proposed lightweight CNN achieves high accuracy, comparable to state-of-the-art models, while significantly reducing computational complexity. The Decision Support System additionally improves pest management by reducing dependence on traditional chemical pesticides and encouraging sustainable practices, demonstrating its potential for real-time applications in precision agriculture.

cs.CV↗

arXiv2025

DemoBias: An Empirical Study to Trace Demographic Biases in Vision Foundation Models

Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities across various downstream tasks, including biometric face recognition (FR) with description. However, demographic biases remain a critical concern in FR, as these foundation models often fail to perform equitably across diverse demographic groups, considering ethnicity/race, gender, and age. Therefore, through our work DemoBias, we conduct an empirical evaluation to investigate the extent of demographic biases in LVLMs for biometric FR with textual token generation tasks. We fine-tuned and evaluated three widely used pre-trained LVLMs: LLaVA, BLIP-2, and PaliGemma on our own generated demographic-balanced dataset. We utilize several evaluation metrics, like group-specific BERTScores and the Fairness Discrepancy Rate, to quantify and trace the performance disparities. The experimental results deliver compelling insights into the fairness and reliability of LVLMs across diverse demographic groups. Our empirical study uncovered demographic biases in LVLMs, with PaliGemma and LLaVA exhibiting higher disparities for Hispanic/Latino, Caucasian, and South Asian groups, whereas BLIP-2 demonstrated comparably consistent. Repository: https://github.com/Sufianlab/DemoBias.

cs.CV↗

arXiv2025

Can Vision Transformers with ResNet's Global Features Fairly Authenticate Demographic Faces?

Biometric face authentication is crucial in computer vision, but ensuring fairness and generalization across demographic groups remains a big challenge. Therefore, we investigated whether Vision Transformer (ViT) and ResNet, leveraging pre-trained global features, can fairly authenticate different demographic faces while relying minimally on local features. In this investigation, we used three pre-trained state-of-the-art (SOTA) ViT foundation models from Facebook, Google, and Microsoft for global features as well as ResNet-18. We concatenated the features from ViT and ResNet, passed them through two fully connected layers, and trained on customized face image datasets to capture the local features. Then, we designed a novel few-shot prototype network with backbone features embedding. We also developed new demographic face image support and query datasets for this empirical study. The network's testing was conducted on this dataset in one-shot, three-shot, and five-shot scenarios to assess how performance improves as the size of the support set increases. We observed results across datasets with varying races/ethnicities, genders, and age groups. The Microsoft Swin Transformer backbone performed better among the three SOTA ViT for this task. The code and data are available at: https://github.com/Sufianlab/FairVitBio.

cs.CV↗

arXiv2025

Intermolecular Radiative Decay: A non-local decay mechanism providing an insider's view of the solvation shell

Aqueous solutions are crucial in chemistry, biology, environmental science, and technology. The chemistry of solutes is influenced by the surrounding solvation shell of water molecules, which have different chemical properties than bulk water due to their different electronic and geometric structure. It is an experimental challenge to selectively investigate this property-determining electronic and geometric structure. Here, we report experimental results on a novel non-local X-ray emission process, Intermolecular Radiative Decay (IRD), for the prototypical ions Na$^{+}$ and Mg$^{2+}$ in water. We show that, in IRD, an electron from the solvation shell fills a core hole in the solute, and the released energy is emitted as an X-ray photon. We analyze the underlying mechanism using theoretical calculations, and show how IRD will allow us to meet the challenge of chemically selective probing of solvation shells from within.

physics.chem-ph↗

arXiv2022

Resonant Soft X-Ray Scattering on LaPt$_2$Si$_2$

X-ray absorption (XAS) and Resonant Inelastic X-ray Scattering (RIXS) spectra of LaPt$_2$Si$_2$ single crystal at the Si L and La N edges are presented. The data are interpreted in terms of density functional theory, showing that the Si spectra can be described in terms of Si $s$ and $d$ local partial density of states (LPDOS), and the La spectra are due to quasi-atomic local $4f$ excitations. Calculations show that Pt $d$-LPDOS dominates the occupied states, and a sharp localized La $f$ state is found in the unoccupied states, in line with the observations.

cond-mat.mtrl-sci↗

arXiv2022

Intra-Atomic and Local Exchange Fields in the Van der Waals Magnet CrI3

We report on a combined experimental and theoretical study on CrI3 single crystals by employing the polarization dependence of resonant inelastic X-ray scattering (RIXS). Our investigations reveal multiple Cr 3d orbital splitting (dd excitations) as well as magnetic dichroism (MD) in the RIXS spectra which is evidence of spin-flip in the dd excitation. Interestingly, the dd excitation energies are similar on both sides of the ferromagnetic transition temperature, TC of 61 K, although MD in RIXS is predominant at 0.4 tesla magnetic field below TC. This demonstrates that the ferromagnetic superexchange interaction that is responsible for the intra-atomic exchange field, is vanishingly small compared to local exchange field that comes from exchange and correlation interaction among the interacting Cr 3d orbitals. The investigation presented here demonstrate that the electronic structure of bulk CrI3 is complex in the sense that dynamical electron correlations are significant. The recorded RIXS spectra reported here reveal clearly resolved Cr 3d intra-orbital dd excitations that represent transitions between electronic levels that are heavily influenced by multi-configuration effects. Our calculations employing the crystal field TTmultiplet theory taking into account the Cr 3d hybridization with the ligand valence states and the full multiplet structure due to intra-atomic and crystal field interactions in Oh and D3d symmetry, clearly reproduced the dichroic trend in experimental RIXS spectra.

cond-mat.str-el↗

arXiv2021

Exotic Magnetic and Electronic Properties of Layered CrI3 Single Crystals Under High Pressure

Through advanced experimental techniques on CrI$_{3}$ single crystals, we derive a previously not discussed pressure-temperature phase diagram. We find that $T_{c}$ increases to $\sim$ 66\,K with pressure up to $\sim$ 3\,GPa followed by a decrease to $\sim$ 10\,K at 21.2\,GPa. The experimental results are reproduced by theoretical calculations based on density functional theory where electron-electron interactions are treated by a static on-site Hubbard U on Cr 3$d$ orbitals. The origin of the pressure induced reduction of the ordering temperature is associated with a decrease of the calculated bond angle, from 95$^{\circ}$ at ambient pressure to $\sim$ 85$^{\circ}$ at 25\,GPa. Above 22\,GPa, the magnetically ordered state is essentially quenched, possibly driving the system to a Kitaev spin-liquid state at low temperature, thereby opening up the possibility of further exploration of long-range quantum entanglement between spins. The pressure-induced semiconductor-to-metal phase transition was revealed by high-pressure resistivity that is accompanied by a transition from a robust ferromagnetic state to gradually more dominating anti-ferromagnetic interactions and was consistent with theoretical modeling.

cond-mat.str-el↗

arXiv2021

Understanding blue shift of the longitudinal surface plasmon resonance during growth of gold nanorods

We have investigated in detail the growth dynamics of gold nanorods with various aspect ratios in different surrounding environments. Surprisingly, a blue shift in the temporal evolution of colloidal gold nanorods in aqueous medium has been observed during the growth of nanorods by UV visible absorption spectroscopy. The longitudinal surface plasmon resonance peak evolves as soon as the nanorods start to grow from spheres, and the system undergoes a blue shift in the absorption spectra. Although a red-shift is expected as a natural phenomenon during the growth process of all nanosystems, our blue shift observation is regarded as a consequence of competition between the parameters of growth solution and actual growth of nanorods. The growth of nanorods contributes to the red-shift which is hidden under the dominating contribution of the growth solution responsible for the observed massive blue shift.

physics.app-ph↗

arXiv2020

BDNet: Bengali Handwritten Numeral Digit Recognition based on Densely connected Convolutional Neural Networks

Images of handwritten digits are different from natural images as the orientation of a digit, as well as similarity of features of different digits, makes confusion. On the other hand, deep convolutional neural networks are achieving huge success in computer vision problems, especially in image classification. BDNet is a densely connected deep convolutional neural network model used to classify (recognize) Bengali handwritten numeral digits. It is end-to-end trained using ISI Bengali handwritten numeral dataset. During training, untraditional data preprocessing and augmentation techniques are used so that the trained model works on a different dataset. The model has achieved the test accuracy of 99.775%(baseline was 99.40%) on the test dataset of ISI Bengali handwritten numerals. So, the BDNet model gives 62.5% error reduction compared to previous state-of-the-art models. Here we have also created a dataset of 1000 images of Bengali handwritten numerals to test the trained model, and it giving promising results. Codes, trained model and our own dataset are available at: {https://github.com/Sufianlab/BDNet}.

cs.CV↗

arXiv2015

Complete polarization characterization of single plasmonic nanoparticle enabled by a novel Dark-field Mueller matrix spectroscopy system

Information on the polarization properties of scattered light from plasmonic systems are of paramount importance due to fundamental interest and potential applications. However, such studies are severely compromised due to the experimental difficulties in recording full polarization response of plasmonic nanostructures. Here, we report on a novel Mueller matrix spectroscopic system capable of acquiring complete polarization information from single isolated plasmonic nanoparticle/nanostructure. The outstanding issues pertaining to reliable measurements of full 4X4 spectroscopic scattering Mueller matrices from single nanoparticle/nanostructures are overcome by integrating an efficient Mueller matrix measurement scheme and a robust calibration method with a dark-field microscopic spectroscopy arrangement.The spectral polarization responses of the required polarization state generator, analyzer units, the imaging and the detection systemsare taken care off by eigenvalue calibration, thus enabling recording of the spectral polarization response (Mueller matrix) exclusively of the plasmonic system. Feasibility of quantitative Mueller matrix polarimetryand its potential utility is illustrated on a simple plasmonic system, that of gold nanorods. The demonstrated novel ability to record full polarization information over a broad wavelength range and to quantify the intrinsic plasmon polarimetry characteristics via Mueller matrix inverse analysis should lead to a novel route towards quantitative understanding, analysis / interpretation of a number of intricate plasmonic effects and may also prove useful towards development of polarization-controlled novel sensing schemes.

physics.optics↗