arXiv subjects

Naveen Kumar

Publications and source records attributed to Naveen Kumar.

At least 37 records · Page 2Linked to original sources

arXiv2023

Accelerating microstructure modelling via machine learning: a new method combining Autoencoder and ConvLSTM

Phase-field modeling is an elegant and versatile computation tool to predict microstructure evolution in materials in the mesoscale regime. However, these simulations require rigorous numerical solutions of differential equations, which are accurate but computationally expensive. To overcome this difficulty, we combine two popular machine learning techniques, autoencoder and convolutional long short-term memory (ConvLSTM), to accelerate the study of microstructural evolution without compromising the resolution of the microstructural representation. After training with phase-field generated microstructures of ten known compositions, the model can accurately predict the microstructure for the future nth frames based on previous m frames for an unknown composition. Replacing n phase-field steps with machine-learned microstructures can significantly accelerate the in silico study of microstructure evolution.

cond-mat.mtrl-sci↗

arXiv2023

Nonclassical properties of a deformed atom-cavity field state

We analyze here a nonclassical state produced by an atom-cavity field interaction. The two-level atom is passed through the single-mode electromagnetic cavity field. By deforming the field operators and introducing nonlinearity to the classic Jaynes-Cummings model, we explore the system in respect of a nonlinear Hamiltonian. Assuming that the atom is in an excited state and the field is in a coherent state initially, the analytic expression for the state vector of the entire system state vector is obtained. With the help of the derived state vector, we calculate the photon number distribution, Mandel's $Q_M$ parameter, Wigner function, anti-bunching, squeezing properties and $Q$ function etc.

quant-ph↗

arXiv2022

High brilliance $γ$-ray generation from the laser interaction in a carbon plasma channel

The generation of collimated, high brilliance $γ$-ray beams from a structured plasma channel target is studied by means of 2D PIC simulations. Simulation results reveal an optimum laser pulse pulse duration of $20\,\text{fs}$, for generating $γ$-photon beams of brilliances up to $10^{20}\,\text{s}^{-1}\text{mm}^{-1}\text{mrad}^{-2}\,(0.1\,\%\text{BW})^{-1}$ and photon energies well above $200$ MeV in the interaction of an ultra-intense laser (incident laser power $ P_L \geq 5$ PW) with a high-Z carbon structured plasma target. These results are aimed at employing the upcoming laser facilities with multi-petawatt (PW) laser powers to study the laser-driven nonlinear quantum electrodynamics processes in an all-optical laboratory setup.

physics.plasm-ph↗

arXiv2022

Energy fluctuations in one dimensional Zhang sandpile model

We consider the Zhang sandpile model in one-dimension (1D) with locally conservative (or dissipative) dynamics and examine its total energy fluctuations at the external drive time scale. The bulk-driven system leads to Lorentzian spectra, with a cutoff time $T$ growing linearly with the system size $L$. The fluctuations show $1/f^α$ behavior with $α\sim 1$ for the boundary drive, and the cutoff time varies non-linearly. For conservative local dynamics, the cutoff time shows a power-law growth $T \sim L^λ$ that differs from an exponential form $ \sim \exp(μL)$ observed for the nonconservative case. We suggest that the local dissipation is not a necessary ingredient of the system in 1D to get the $1/f$ noise, and the cutoff time can reveal the distinct nature of the local dynamics. We also discuss the energy fluctuations for locally nonconservative dynamics with random dissipation.

cond-mat.stat-mech↗

arXiv2021

Biomarker Gene Identification for Breast Cancer Classification

BACKGROUND: Breast cancer has emerged as one of the most prevalent cancers among women leading to a high mortality rate. Due to the heterogeneous nature of breast cancer, there is a need to identify differentially expressed genes associated with breast cancer subtypes for its timely diagnosis and treatment. OBJECTIVE: To identify a small gene set for each of the four breast cancer subtypes that could act as its signature, the paper proposes a novel algorithm for gene signature identification. METHODS: The present work uses interpretable AI methods to investigate the predictions made by the deep neural network employed for subtype classification to identify biomarkers using the TCGA breast cancer RNA Sequence data. RESULTS: The proposed algorithm led to the discovery of a set of 43 differentially expressed gene signatures. We achieved a competitive average 10-fold accuracy of 0.91, using neural network classifier. Further, gene set analysis revealed several relevant pathways, such as GRB7 events in ERBB2 and p53 signaling pathway. Using the Pearson correlation matrix, we noted that the subtype-specific genes are correlated within each subtype. CONCLUSIONS: The proposed technique enables us to find a concise and clinically relevant gene signature set.

cs.LG↗

arXiv2021

Radiation pressure acceleration of protons from structured thin-foil targets

The process of radiation pressure acceleration (RPA) of ions is investigated with the aim of suppressing the Rayleigh-Taylor like transverse instabilities in laser-foil interaction. This is achieved by imposing surface and density modulations on the target surface. We also study the efficacy of RPA of ions from density modulated and structured targets in the radiation dominated regime where the radiation reaction effects are important. We show that the use of density modulated and structured targets and the radiation reaction effects can help in achieving the twin goals of high ion energy (in GeV range) and lower energy spread.

physics.plasm-ph↗

arXiv2021

Deep Learning Based Model for Breast Cancer Subtype Classification

Breast cancer has long been a prominent cause of mortality among women. Diagnosis, therapy, and prognosis are now possible, thanks to the availability of RNA sequencing tools capable of recording gene expression data. Molecular subtyping being closely related to devising clinical strategy and prognosis, this paper focuses on the use of gene expression data for the classification of breast cancer into four subtypes, namely, Basal, Her2, LumA, and LumB. In stage 1, we suggested a deep learning-based model that uses an autoencoder to reduce dimensionality. The size of the feature set is reduced from 20,530 gene expression values to 500 by using an autoencoder. This encoded representation is passed to the deep neural network of the second stage for the classification of patients into four molecular subtypes of breast cancer. By deploying the combined network of stages 1 and 2, we have been able to attain a mean 10-fold test accuracy of 0.907 on the TCGA breast cancer dataset. The proposed framework is fairly robust throughout 10 different runs, as shown by the boxplot for classification accuracy. Compared to related work reported in the literature, we have achieved a competitive outcome. In conclusion, the proposed two-stage deep learning-based model is able to accurately classify four breast cancer subtypes, highlighting the autoencoder's capacity to deduce the compact representation and the neural network classifier's ability to correctly label breast cancer patients.

cs.LG↗

arXiv2021

Non-thermal particle acceleration at highly oblique non-relativistic shocks

The non-thermal acceleration of electrons and ions at an oblique, non-relativistic shock is studied using large scale particle-in-cell (PIC) simulations in one spatial dimension. Physical parameters are selected to highlight the role of electron pre-heating in injection and subsequent acceleration of particles at high Mach number shocks. Simulation results show evidence for the early onset of the diffusive shock acceleration process for both electrons and ions at a highly oblique subluminal shock. Ion acceleration efficiencies of $\lesssim 5\%$ are measured at late times, though this is not the saturated value.

physics.plasm-ph↗

arXiv2021

COV-ELM classifier: An Extreme Learning Machine based identification of COVID-19 using Chest X-Ray Images

Coronaviruses constitute a family of viruses that gives rise to respiratory diseases. As COVID-19 is highly contagious, early diagnosis of COVID-19 is crucial for an effective treatment strategy. However, the RT-PCR test which is considered to be a gold standard in the diagnosis of COVID-19 suffers from a high false-negative rate. Chest X-ray (CXR) image analysis has emerged as a feasible and effective diagnostic technique towards this objective. In this work, we propose the COVID-19 classification problem as a three-class classification problem to distinguish between COVID-19, normal, and pneumonia classes. We propose a three-stage framework, named COV-ELM. Stage one deals with preprocessing and transformation while stage two deals with feature extraction. These extracted features are passed as an input to the ELM at the third stage, resulting in the identification of COVID-19. The choice of ELM in this work has been motivated by its faster convergence, better generalization capability, and shorter training time in comparison to the conventional gradient-based learning algorithms. As bigger and diverse datasets become available, ELM can be quickly retrained as compared to its gradient-based competitor models. The proposed model achieved a macro average F1-score of 0.95 and the overall sensitivity of ${0.94 \pm 0.02} at a 95% confidence interval. When compared to state-of-the-art machine learning algorithms, the COV-ELM is found to outperform its competitors in this three-class classification scenario. Further, LIME has been integrated with the proposed COV-ELM model to generate annotated CXR images. The annotations are based on the superpixels that have contributed to distinguish between the different classes. It was observed that the superpixels correspond to the regions of the human lungs that are clinically observed in COVID-19 and Pneumonia cases.

eess.IV↗

arXiv2021

Linking space-time correlations for a class of self-organized critical systems

The hypothesis of self-organized criticality explains the existence of long-range `space-time' correlations, observed inseparably in many natural dynamical systems. A simple link between these correlations is yet unclear, particularly in fluctuations at `external drive' time scales. As an example, we consider a class of sandpile models displaying non-trivial correlations. Employing the scaling methods, we demonstrate the computation of spatial correlation by establishing a link between local and global temporal correlations.

cond-mat.stat-mech↗

arXiv2021

Scaling theory for the $1/f$ noise

We show that in a broad class of processes that show a $1/f^α$ spectrum, the power also explicitly depends on the characteristic time scale. Despite an enormous amount of work, this generic behavior remains so far overlooked and poorly understood. An intriguing example is how the power spectrum of a simple random walk on a ring with $L$ sites shows $1/f^{3/2}$ not $1/f^2$ behavior in the frequency range $1/L^2 \ll f \ll 1/2$. We address the fundamental issue by a scaling method and discuss a class of solvable processes covering physically relevant applications.

cond-mat.stat-mech↗

arXiv2021

Exploring the limits of Concurrency in ML Training on Google TPUs

Recent results in language understanding using neural networks have required training hardware of unprecedentedscale, with thousands of chips cooperating on a single training run. This paper presents techniques to scaleML models on the Google TPU Multipod, a mesh with 4096 TPU-v3 chips. We discuss model parallelism toovercome scaling limitations from the fixed batch size in data parallelism, communication/collective optimizations,distributed evaluation of training metrics, and host input processing scaling optimizations. These techniques aredemonstrated in both the TensorFlow and JAX programming frameworks. We also present performance resultsfrom the recent Google submission to the MLPerf-v0.7 benchmark contest, achieving record training times from16 to 28 seconds in four MLPerf models on the Google TPU-v3 Multipod machine.

cs.LG↗

arXiv2020

RNN based Incremental Online Spoken Language Understanding

Spoken Language Understanding (SLU) typically comprises of an automatic speech recognition (ASR) followed by a natural language understanding (NLU) module. The two modules process signals in a blocking sequential fashion, i.e., the NLU often has to wait for the ASR to finish processing on an utterance basis, potentially leading to high latencies that render the spoken interaction less natural. In this paper, we propose recurrent neural network (RNN) based incremental processing towards the SLU task of intent detection. The proposed methodology offers lower latencies than a typical SLU system, without any significant reduction in system accuracy. We introduce and analyze different recurrent neural network architectures for incremental and online processing of the ASR transcripts and compare it to the existing offline systems. A lexical End-of-Sentence (EOS) detector is proposed for segmenting the stream of transcript into sentences for intent classification. Intent detection experiments are conducted on benchmark ATIS, Snips and Facebook's multilingual task oriented dialog datasets modified to emulate a continuous incremental stream of words with no utterance demarcation. We also analyze the prospects of early intent detection, before EOS, with our proposed system.

cs.CL↗

arXiv2020

Pair-beam propagation in a magnetised plasma for modelling the polarized radiation emission from gamma-ray bursts in laboratory astrophysics experiments

The propagation of a relativistic electron-positron beam in a magnetized electron-ion plasma is studied, focusing on the polarization of the radiation generated in this case. Special emphasis is laid on investigating the polarization of the generated radiation for a range of beam-plasma parameters, transverse and longitudinal beam sizes, and the external magnetic fields. Our results not only help in understanding the high degrees of circular polarization observed in gamma-rays bursts but they also help in distinguishing the different modes associated with the filamentation dynamics of the pair-beam in laboratory astrophysics experiments.

physics.plasm-ph↗

arXiv2020

Electronic parametric instabilities of an ultrarelativistic laser pulse in a plasma

Electronic parametric instabilities of an ultrarelativistic circularly polarized laser pulse propagating in underdense plasmas are studied by numerically solving the dispersion relation which includes the effect of the radiation reaction force in laser-driven plasma dynamics. Emphasis is placed on studying the different modes in the laser-plasma system and identifying the absolute and convective nature of the unstable modes in a parameter map spanned by the normalized laser vector potential and the plasma density. Implications for the ultraintense laser-plasma experiments are pointed out.

physics.plasm-ph↗

arXiv2020

MLPerf Training Benchmark

Machine learning (ML) needs industry-standard performance benchmarks to support design and competitive evaluation of the many emerging software and hardware solutions for ML. But ML training presents three unique benchmarking challenges absent from other domains: optimizations that improve training throughput can increase the time to solution, training is stochastic and time to solution exhibits high variance, and software and hardware systems are so diverse that fair benchmarking with the same binary, code, and even hyperparameters is difficult. We therefore present MLPerf, an ML benchmark that overcomes these challenges. Our analysis quantitatively evaluates MLPerf's efficacy at driving performance and scalability improvements across two rounds of results from multiple vendors.

cs.LG↗

arXiv2019

Quantized deep learning models on low-power edge devices for robotic systems

In this work, we present a quantized deep neural network deployed on a low-power edge device, inferring learned motor-movements of a suspended robot in a defined space. This serves as the fundamental building block for the original setup, a robotic system for farms or greenhouses aimed at a wide range of agricultural tasks. Deep learning on edge devices and its implications could have a substantial impact on farming systems in the developing world, leading not only to sustainable food production and income, but also increased data privacy and autonomy.

eess.SP↗

arXiv2019

Scale MLPerf-0.6 models on Google TPU-v3 Pods

The recent submission of Google TPU-v3 Pods to the industry wide MLPerf v0.6 training benchmark demonstrates the scalability of a suite of industry relevant ML models. MLPerf defines a suite of models, datasets and rules to follow when benchmarking to ensure results are comparable across hardware, frameworks and companies. Using this suite of models, we discuss the optimizations and techniques including choice of optimizer, spatial partitioning and weight update sharding necessary to scale to 1024 TPU chips. Furthermore, we identify properties of models that make scaling them challenging, such as limited data parallelism and unscaled weights. These optimizations contribute to record performance in transformer, Resnet-50 and SSD in the Google MLPerf-0.6 submission.

cs.LG↗