Searcharxiv⌕ Search

arXiv subjects

Manoj Kumar

Publications and source records attributed to Manoj Kumar.

At least 73 records · Page 4Linked to original sources
arXiv2020

Age structured SIR model for the spread of infectious diseases through indirect contacts

In this article, we discuss an age-structured SIR model in which disease not only spread through direct person to person contacts for e.g. infection due to surface contamination but it can also spread through indirect contacts. It is evident that age also plays a crucial role in SARS virus infection including COVID-19 infection. We formulate our model as an abstract semilinear Cauchy problem in an appropriate Banach space to show the existence of solution and also show the existence of steady states. It is assumed in this work that the population is in a demographic stationary state and show that there is no disease-free equilibrium point as long as there is a transmission of infection due to the indirect contacts in the environment.

q-bio.PE↗
arXiv2020

Meta-learning with Latent Space Clustering in Generative Adversarial Network for Speaker Diarization

The performance of most speaker diarization systems with x-vector embeddings is both vulnerable to noisy environments and lacks domain robustness. Earlier work on speaker diarization using generative adversarial network (GAN) with an encoder network (ClusterGAN) to project input x-vectors into a latent space has shown promising performance on meeting data. In this paper, we extend the ClusterGAN network to improve diarization robustness and enable rapid generalization across various challenging domains. To this end, we fetch the pre-trained encoder from the ClusterGAN and fine-tune it by using prototypical loss (meta-ClusterGAN or MCGAN) under the meta-learning paradigm. Experiments are conducted on CALLHOME telephonic conversations, AMI meeting data, DIHARD II (dev set) which includes challenging multi-domain corpus, and two child-clinician interaction corpora (ADOS, BOSCC) related to the autism spectrum disorder domain. Extensive analyses of the experimental data are done to investigate the effectiveness of the proposed ClusterGAN and MCGAN embeddings over x-vectors. The results show that the proposed embeddings with normalized maximum eigengap spectral clustering (NME-SC) back-end consistently outperform Kaldi state-of-the-art z-vector diarization system. Finally, we employ embedding fusion with x-vectors to provide further improvement in diarization performance. We achieve a relative diarization error rate (DER) improvement of 6.67% to 53.93% on the aforementioned datasets using the proposed fused embeddings over x-vectors. Besides, the MCGAN embeddings provide better performance in the number of speakers estimation and short speech segment diarization as compared to x-vectors and ClusterGAN in telephonic data.

eess.AS↗
arXiv2020

On the comparison of optimization algorithms for the random-field Potts model

For many systems with quenched disorder the study of ground states can crucially contribute to a thorough understanding of the physics at play, be it for the critical behavior if that is governed by a zero-temperature fixed point or for uncovering properties of the ordered phase. While ground states can in principle be computed using general-purpose optimization algorithms such as simulated annealing or genetic algorithms, it is often much more efficient to use exact or approximate techniques specifically tailored to the problem at hand. For certain systems with discrete degrees of freedom such as the random-field Ising model, there are polynomial-time methods to compute exact ground states. But even as the number of states increases beyond two as in the random-field Potts model, the problem becomes NP hard and one cannot hope to find exact ground states for relevant system sizes. Here, we compare a number of approximate techniques for this problem and evaluate their performance.

cond-mat.dis-nn↗
arXiv2020

Growth Kinetics and Aging Phenomena in a Frustrated System

We study numerically the ordering kinetics in a two-dimensional Ising model with random coupling where the fraction of antiferromagnetic links $a$ can be gradually tuned. We show that, upon increasing such fraction, the behavior changes in a radical way. Small $a$ does not prevent the system from a complete ordering, but this occurs in an extremely (logarithmically) slow manner. However, larger values of this parameter destroy complete ordering, due to frustration, and the evolution is comparatively faster (algebraic). Our study shows a precise correspondence between the kind of developing order, ferromagnetic versus frustrated, and the speed of evolution. The aging properties of the system are studied by focusing on the scaling properties of two-time quantities, the autocorrelation and linear response functions. We find that the contribution of equilibrium and an aging part to these functions occurs differently in the various regions of the phase diagram of the model. When quenching inside the ferromagnetic phase, the two-time quantities are obtained by the addition of these parts. Instead, in the paramagnetic phase, these two contributions enter multiplicatively. Both of the scaling forms are shown with excellent accuracy, and the corresponding scaling functions and exponents have been determined and discussed.

cond-mat.stat-mech↗
arXiv2020

Auto-Tuning Spectral Clustering for Speaker Diarization Using Normalized Maximum Eigengap

In this study, we propose a new spectral clustering framework that can auto-tune the parameters of the clustering algorithm in the context of speaker diarization. The proposed framework uses normalized maximum eigengap (NME) values to estimate the number of clusters and the parameters for the threshold of the elements of each row in an affinity matrix during spectral clustering, without the use of parameter tuning on the development set. Even through this hands-off approach, we achieve a comparable or better performance across various evaluation sets than the results found using traditional clustering methods that apply careful parameter tuning and development data. A relative improvement of 17% in the speaker error rate on the well-known CALLHOME evaluation set shows the effectiveness of our proposed spectral clustering with auto-tuning.

eess.AS↗
arXiv2020

VideoFlow: A Conditional Flow-Based Model for Stochastic Video Generation

Generative models that can model and predict sequences of future events can, in principle, learn to capture complex real-world phenomena, such as physical interactions. However, a central challenge in video prediction is that the future is highly uncertain: a sequence of past observations of events can imply many possible futures. Although a number of recent works have studied probabilistic models that can represent uncertain futures, such models are either extremely expensive computationally as in the case of pixel-level autoregressive models, or do not directly optimize the likelihood of the data. To our knowledge, our work is the first to propose multi-frame video prediction with normalizing flows, which allows for direct optimization of the data likelihood, and produces high-quality stochastic predictions. We describe an approach for modeling the latent space dynamics, and demonstrate that flow-based generative models offer a viable and competitive approach to generative modelling of video.

cs.CV↗
arXiv2019

Meta-learning for robust child-adult classification from speech

Computational modeling of naturalistic conversations in clinical applications has seen growing interest in the past decade. An important use-case involves child-adult interactions within the autism diagnosis and intervention domain. In this paper, we address a specific sub-problem of speaker diarization, namely child-adult speaker classification in such dyadic conversations with specified roles. Training a speaker classification system robust to speaker and channel conditions is challenging due to inherent variability in the speech within children and the adult interlocutors. In this work, we propose the use of meta-learning, in particular, prototypical networks which optimize a metric space across multiple tasks. By modeling every child-adult pair in the training set as a separate task during meta-training, we learn a representation with improved generalizability compared to conventional supervised learning. We demonstrate improvements over state-of-the-art speaker embeddings (x-vectors) under two evaluation settings: weakly supervised classification (up to 14.53% relative improvement in F1-scores) and clustering (up to relative 9.66% improvement in cluster purity). Our results show that protonets can potentially extract robust speaker embeddings for child-adult classification from speech.

eess.AS↗
arXiv2019

Learning Domain Invariant Representations for Child-Adult Classification from Speech

Diagnostic procedures for ASD (autism spectrum disorder) involve semi-naturalistic interactions between the child and a clinician. Computational methods to analyze these sessions require an end-to-end speech and language processing pipeline that go from raw audio to clinically-meaningful behavioral features. An important component of this pipeline is the ability to automatically detect who is speaking when i.e., perform child-adult speaker classification. This binary classification task is often confounded due to variability associated with the participants' speech and background conditions. Further, scarcity of training data often restricts direct application of conventional deep learning methods. In this work, we address two major sources of variability - age of the child and data source collection location - using domain adversarial learning which does not require labeled target domain data. We use two methods, generative adversarial training with inverted label loss and gradient reversal layer to learn speaker embeddings invariant to the above sources of variability, and analyze different conditions under which the proposed techniques improve over conventional learning methods. Using a large corpus of ADOS-2 (autism diagnostic observation schedule, 2nd edition) sessions, we demonstrate upto 13.45% and 6.44% relative improvements over conventional learning methods.

eess.AS↗
arXiv2019

Speaker diarization using latent space clustering in generative adversarial network

In this work, we propose deep latent space clustering for speaker diarization using generative adversarial network (GAN) backprojection with the help of an encoder network. The proposed diarization system is trained jointly with GAN loss, latent variable recovery loss, and a clustering-specific loss. It uses x-vector speaker embeddings at the input, while the latent variables are sampled from a combination of continuous random variables and discrete one-hot encoded variables using the original speaker labels. We benchmark our proposed system on the AMI meeting corpus, and two child-clinician interaction corpora (ADOS and BOSCC) from the autism diagnosis domain. ADOS and BOSCC contain diagnostic and treatment outcome sessions respectively obtained in clinical settings for verbal children and adolescents with autism. Experimental results show that our proposed system significantly outperform the state-of-the-art x-vector based diarization system on these databases. Further, we perform embedding fusion with x-vectors to achieve a relative DER improvement of 31%, 36% and 49% on AMI eval, ADOS and BOSCC corpora respectively, when compared to the x-vector baseline using oracle speech segmentation.

eess.AS↗
arXiv2019

A study of semi-supervised speaker diarization system using gan mixture model

We propose a new speaker diarization system based on a recently introduced unsupervised clustering technique namely, generative adversarial network mixture model (GANMM). The proposed system uses x-vectors as front-end representation. Spectral embedding is used for dimensionality reduction followed by k-means initialization during GANMM pre-training. GANMM performs unsupervised speaker clustering by efficiently capturing complex data distributions. Experimental results on the AMI meeting corpus show that the proposed semi-supervised diarization system matches or exceeds the performance of competitive baselines. On an evaluation set containing fifty sessions with varying durations, the best achieved average diarization error rate (DER) is 17.11%, a relative improvement of 33% over the information bottleneck baseline and comparable to xvector baseline.

eess.AS↗
arXiv2019

Graphene Quantum Dot as a Probe for DNA Nucleobase Detection: A First Principle Study

In this study, structural stability, electronic, optical and vibrational properties of DNA nucleobase adsorbed Graphene Quantum Dot (GQD) has been investigated using density functional theory. Based on state-of-art electronic structure calculations, we predict order of GQD sensitivity for DNA nucleobase as Thymine > Cytosine > Guanine > Adenine. An interaction of GQD with DNA nucleobase leads to modulation in electronic energy gap. Our calculated UV/vis and IR vibrational spectra show unique spectral band that can be used fingerprints in next generation DNA sequencing diagnostics.

cond-mat.mes-hall↗
arXiv2019

Fourier analysis associated to a vector measure on a compact group

In this paper, we introduce and study the Fourier transform of functions which are integrable with respect to a vector measure on a compact group (not necessarily abelian). We also study the Fourier transform of vector measures. We also introduce and study the convolution of functions from $L^p$-spaces associated to a vector measure. We prove some analogues of the classical Young's inequalities. Similarly, we also study convolution of a scalar measure and a vector measure.

math.FA↗
arXiv2019

Convolution structures for an Orlicz space with respect to vector measures on a compact group

The aim of this paper is to present some results about the space L^Φ(ν), where νis a vector measure on a compact (not necessarily abelian) group and Φis a Young function. We show that under certain conditions, the space L^Φ(ν) becomes an L^1(G)-module with respect to the usual convolution of functions. We also define one more convolution structure on L^Φ(ν).

math.FA↗
arXiv2019

Effects of Frustration on Fluctuation-dissipation Relations

We study numerically the aging properties of the two-dimensional Ising model with quenched disorder considered in our recent paper [Phys. Rev. E 95, 062136 (2017)], where frustration can be tuned by varying the fraction a of antiferromagnetic interactions. Specifically we focus on the scaling properties of the autocorrelation and linear response functions after a quench of the model to a low temperature. We find that the interplay between equilibrium and aging occurs differently in the various regions of the phase diagram of the model. When the quench is made into the ferromagnetic phase the two-time quantities are made by the sum of an equilibrium and an aging part, whereas in the paramagnetic phase these parts combine in a multiplicative way. Scaling forms are shown to be obeyed with good accuracy, and the corresponding exponents and scaling functions are determined and discussed in the framework of what is known in clean and disordered systems.

cond-mat.stat-mech↗
arXiv2018

Exact solutions of fractional partial differential equations by Sumudu transform iterative method

Developing analytical methods for solving fractional partial differential equations (FPDEs) is an active area of research. Especially finding exact solutions of FPDEs is a challenging task. In the present paper we extend Sumudu transform iterative method (STIM) to solve a variety of time and space FPDEs as well as systems of them. We demonstrate the utility of the method by finding exact solutions to a large number of FPDEs.

math.AP↗
arXiv2018

Measuring Conversational Productivity in Child Forensic Interviews

Child Forensic Interviewing (FI) presents a challenge for effective information retrieval and decision making. The high stakes associated with the process demand that expert legal interviewers are able to effectively establish a channel of communication and elicit substantive knowledge from the child-client while minimizing potential for experiencing trauma. As a first step toward computationally modeling and producing quality spoken interviewing strategies and a generalized understanding of interview dynamics, we propose a novel methodology to computationally model effectiveness criteria, by applying summarization and topic modeling techniques to objectively measure and rank the responsiveness and conversational productivity of a child during FI. We score information retrieval by constructing an agenda to represent general topics of interest and measuring alignment with a given response and leveraging lexical entrainment for responsiveness. For comparison, we present our methods along with traditional metrics of evaluation and discuss the use of prior information for generating situational awareness.

cs.CL↗
arXiv2018

Parallel Architecture and Hyperparameter Search via Successive Halving and Classification

We present a simple and powerful algorithm for parallel black box optimization called Successive Halving and Classification (SHAC). The algorithm operates in $K$ stages of parallel function evaluations and trains a cascade of binary classifiers to iteratively cull the undesirable regions of the search space. SHAC is easy to implement, requires no tuning of its own configuration parameters, is invariant to the scale of the objective function and can be built using any choice of binary classifier. We adopt tree-based classifiers within SHAC and achieve competitive performance against several strong baselines for optimizing synthetic functions, hyperparameters and architectures.

cs.CV↗