arXiv subjects

Krishnendu Ghosh

Publications and source records attributed to Krishnendu Ghosh.

10 recordsLinked to original sources

arXiv2025

Sharing the Learned Knowledge-base to Estimate Convolutional Filter Parameters for Continual Image Restoration

Continual learning is an emerging topic in the field of deep learning, where a model is expected to learn continuously for new upcoming tasks without forgetting previous experiences. This field has witnessed numerous advancements, but few works have been attempted in the direction of image restoration. Handling large image sizes and the divergent nature of various degradation poses a unique challenge in the restoration domain. However, existing works require heavily engineered architectural modifications for new task adaptation, resulting in significant computational overhead. Regularization-based methods are unsuitable for restoration, as different restoration challenges require different kinds of feature processing. In this direction, we propose a simple modification of the convolution layer to adapt the knowledge from previous restoration tasks without touching the main backbone architecture. Therefore, it can be seamlessly applied to any deep architecture without any structural modifications. Unlike other approaches, we demonstrate that our model can increase the number of trainable parameters without significantly increasing computational overhead or inference time. Experimental validation demonstrates that new restoration tasks can be introduced without compromising the performance of existing tasks. We also show that performance on new restoration tasks improves by adapting the knowledge from the knowledge base created by previous restoration tasks. The code is available at https://github.com/aupendu/continual-restore.

cs.CV↗

arXiv2025

Artificially Generated Visual Scanpath Improves Multi-label Thoracic Disease Classification in Chest X-Ray Images

Expert radiologists visually scan Chest X-Ray (CXR) images, sequentially fixating on anatomical structures to perform disease diagnosis. An automatic multi-label classifier of diseases in CXR images can benefit by incorporating aspects of the radiologists' approach. Recorded visual scanpaths of radiologists on CXR images can be used for the said purpose. But, such scanpaths are not available for most CXR images, which creates a gap even for modern deep learning based classifiers. This paper proposes to mitigate this gap by generating effective artificial visual scanpaths using a visual scanpath prediction model for CXR images. Further, a multi-class multi-label classifier framework is proposed that uses a generated scanpath and visual image features to classify diseases in CXR images. While the scanpath predictor is based on a recurrent neural network, the multi-label classifier involves a novel iterative sequential model with an attention module. We show that our scanpath predictor generates human-like visual scanpaths. We also demonstrate that the use of artificial visual scanpaths improves multi-class multi-label disease classification results on CXR images. The above observations are made from experiments involving around 0.2 million CXR images from 2 widely-used datasets considering the multi-label classification of 14 pathological findings. Code link: https://github.com/ashishverma03/SDC

eess.IV↗

arXiv2021

Spin-spin interactions in solids from mixed all-electron and pseudopotential calculations $-$ a path to screening materials for spin qubits

Understanding the quantum dynamics of spin defects and their coherence properties requires accurate modeling of spin-spin interaction in solids and molecules, for example by using spin Hamiltonians with parameters obtained from first-principles calculations. We present a real-space approach based on density functional theory for the calculation of spin-Hamiltonian parameters, where only selected atoms are treated at the all-electron level, while the rest of the system is described with the pseudopotential approximation. Our approach permits calculations for systems containing more than 1000 atoms, as demonstrated for defects in diamond and silicon carbide. We show that only a small number of atoms surrounding the defect needs to be treated at the all-electron level, in order to obtain an overall all-electron accuracy for hyperfine and zero-field splitting tensors. We also present results for coherence times, computed with the cluster correlation expansion method, highlighting the importance of accurate spin-Hamiltonian parameters for quantitative predictions of spin dynamics.

cond-mat.mtrl-sci↗

arXiv2019

DFT-FE -- A massively parallel adaptive finite-element code for large-scale density functional theory calculations

We present an accurate, efficient and massively parallel finite-element code, DFT-FE, for large-scale ab-initio calculations (reaching $\sim 100,000$ electrons) using Kohn-Sham density functional theory (DFT). DFT-FE is based on a local real-space variational formulation of the Kohn-Sham DFT energy functional that is discretized using a higher-order adaptive spectral finite-element (FE) basis, and treats pseudopotential and all-electron calculations in the same framework, while accommodating non-periodic, semi-periodic and periodic boundary conditions. We discuss the main aspects of the code, which include, the various strategies of adaptive FE basis generation, and the different approaches employed in the numerical implementation of the solution of the discrete Kohn-Sham problem that are focused on significantly reducing the floating point operations, communication costs and latency. We demonstrate the accuracy of DFT-FE by comparing the energies, ionic forces and periodic cell stresses on a wide range of problems with popularly used DFT codes. Further, we demonstrate that DFT-FE significantly outperforms widely used plane-wave codes---both in CPU-times and wall-times, and on both non-periodic and periodic systems---at systems sizes beyond a few thousand electrons, with over $5-10$ fold speedups in systems with more than 10,000 electrons. The benchmark studies also highlight the excellent parallel scalability of DFT-FE, with strong scaling demonstrated on up to 192,000 MPI tasks.

physics.comp-ph↗

arXiv2019

All-electron density functional calculations for electron and nuclear spin interactions in molecules and solids

The interaction between electronic and nuclear spins in the presence of external magnetic fields can be described by a spin Hamiltonian, with parameters obtained from first principles, electronic structure calculations. We describe an approach to compute these parameters, applicable to both molecules and solids, which is based on Density Functional Theory (DFT) and real-space, all-electron calculations using finite elements (FE). We report results for hyperfine tensors, zero field splitting tensors (spin-spin component) and nuclear quadrupole tensors of a series of molecules and of the nitrogen-vacancy center in diamond. We compare our results with those of calculations using Gaussian orbitals and plane-wave basis sets, and we discuss their numerical accuracy. We show that calculations based on FE can be systematically converged with respect to the basis set, thus allowing one to establish reference values for the spin Hamiltonian parameters, at a given level of DFT.

cond-mat.mtrl-sci↗

arXiv2018

Impact Ionization in $β-Ga_2O_3$

A theoretical investigation of extremely high field transport in an emerging wide-bandgap material $β-Ga_2O_3$ is reported from first principles. The signature high-field effect explored here is impact ionization. Interaction between a valence-band electron and an excited electron is computed from the matrix elements of a screened Coulomb operator. Maximally localized Wannier functions (MLWF) are utilized in computing the impact ionization rates. A full-band Monte Carlo (FBMC) simulation is carried out incorporating the impact ionization rates, and electron-phonon scattering rates. This work brings out valuable insights on the impact ionization coefficient (IIC) of electrons in $β-Ga_2O_3$. The isolation of the $Γ$ point conduction band minimum by a significantly high energy from other satellite band pockets play a vital role in determining ionization co-efficients. IICs are calculated for electric fields ranging up to 8 MV/cm for different crystal directions. A Chynoweth fitting of the computed IICs is done to calibrate ionization models in device simulators.

cond-mat.mtrl-sci↗

arXiv2017

Electron Mobility in Monoclinic β-Ga2O3 - Effect of Plasmon-phonon Coupling, Anisotropy, and Confinement

This work reports an investigation of electron transport in monoclinic β-Ga2O3 based on a combination of density functional perturbation theory based lattice dynamical computations, coupling calculation of lattice modes with collective plasmon oscillations and Boltzmann theory based transport calculations. The strong entanglement of the plasmon with the different longitudinal optical (LO) modes make the role LO-plasmon coupling crucial for transport. The electron density dependence of the electron mobility in β-Ga2O3 is studied in bulk material form and also in the form of two-dimensional electron gas. Under high electron density a bulk mobility of 182 cm2/ V.s is predicted while in 2DEG form the corresponding mobility is about 418 cm2/V.s when remote impurities are present at the interface and improves further as the remote impurity center moves away from the interface. The trend of the electron mobility shows promise for realizing high electron mobility in dopant isolated electron channels. The experimentally observed small anisotropy in mobility is traced through a transient Monte Carlo simulation. It is found that the anisotropy of the IR active phonon modes is responsible for giving rise to the anisotropy in low-field electron mobility.

cond-mat.mtrl-sci↗

arXiv2017

Ab Initio Velocity-Field Curves in Monoclinic $β$-Ga\textsubscript{2}O\textsubscript{3}}

We investigate the high-field transport in monoclinic $β$-Ga\textsubscript{2}O\textsubscript{3}} using a combination of ab initio calculations and full band Monte Carlo (FBMC) simulation. Scattering rate calculation and the final state selection in the FBMC simulation use complete wave-vector (both electron and phonon) and crystal direction dependent electron phonon interaction (EPI) elements. We propose and implement a semi-coarse version of the Wannier-Fourier interpolation method [F. Giustino, M. L. Cohen, and S. G. Louie, Physical Review B, vol. 76, no. 16, 2007] for short-range non-polar optical phonon (EPI) elements in order to ease the computational requirement in FBMC simulation. During the interpolation of the EPI, the inverse Fourier sum over the real-space electronic grids is done on a coarse mesh while the unitary rotations are done on a fine mesh. This paper reports the high field transport in monoclinic $β$-Ga\textsubscript{2}O\textsubscript{3}} with deep insight on the contribution of electron-phonon interactions, and velocity-field characteristics for electric fields ranging up to 450 kV/cm in different crystal directions. A peak velocity of $2\times10^{7}& cm/s is estimated at an electric field of 200 kV/cm.

cond-mat.mtrl-sci↗

arXiv2015

Thermoelectric Transport Coefficients in Mono-layer MoS2 and WSe2 Role of Substrate, Interface Phonons, Plasmon, and Dynamic Screening

The thermoelectric transport coefficients of electrons in two recently emerged transition metal dichalcogenides(TMD), MoS2 and WSe2, are calculated by solving Boltzmann Transport equation and coupled electrical and thermal current equations using Rode iterative technique. Scattering from localized donor impurities, acoustic deformation potential, longitudinal optical (LO) phonons, and substrate induced remote phonon modes are taken into account. Hybridization of TMD plasmon with remote phonon modes is investigated. Dynamic screening under linear polarization response is explored in TMDs sitting on a dielectric environment and the screened electron-phonon coupling matrix elements are calculated. The effect of screening and substrate induced remote phonon mediated scattering on the transport coefficients of the mentioned materials is explained. The transport coefficients are obtained for a varying range of temperature and doping density for three different types of substrates SiO2, Al2O3, and HfO2. The thermoelectric properties of interest including Seebeck coefficient, Peltier coefficient, and electronic thermal conductivity are calculated.

cond-mat.mtrl-sci↗

arXiv2013

Design and Analysis of High Frequency InN Tunnel Transistors

This work reports the design and analysis of an n-type tunneling field effect transistor based on InN. The tunneling current is evaluated from the fundamental principles of quantum mechanical tunneling and semiclassical carrier transport. We investigate the RF performance of the device. High transconductance of 2 mS/um and current gain cut-off frequency of around 460 GHz makes the device suitable for THz applications. A significant reduction in gate to drain capacitance is observed under relatively higher drain bias. In this regard, the avalanche breakdown phenomenon in highly doped InN junctions is analyzed quantitatively for the first time and is compared to that of Si and InAs.

cond-mat.mtrl-sci↗