arXiv subjects

Alpan Raval

Publications and source records attributed to Alpan Raval.

At least 19 recordsLinked to original sources

arXiv2024

NurtureNet: A Multi-task Video-based Approach for Newborn Anthropometry

Malnutrition among newborns is a top public health concern in developing countries. Identification and subsequent growth monitoring are key to successful interventions. However, this is challenging in rural communities where health systems tend to be inaccessible and under-equipped, with poor adherence to protocol. Our goal is to equip health workers and public health systems with a solution for contactless newborn anthropometry in the community. We propose NurtureNet, a multi-task model that fuses visual information (a video taken with a low-cost smartphone) with tabular inputs to regress multiple anthropometry estimates including weight, length, head circumference, and chest circumference. We show that visual proxy tasks of segmentation and keypoint prediction further improve performance. We establish the efficacy of the model through several experiments and achieve a relative error of 3.9% and mean absolute error of 114.3 g for weight estimation. Model compression to 15 MB also allows offline deployment to low-cost smartphones.

cs.CV↗

arXiv2022

Predicting Treatment Adherence of Tuberculosis Patients at Scale

Tuberculosis (TB), an infectious bacterial disease, is a significant cause of death, especially in low-income countries, with an estimated ten million new cases reported globally in $2020$. While TB is treatable, non-adherence to the medication regimen is a significant cause of morbidity and mortality. Thus, proactively identifying patients at risk of dropping off their medication regimen enables corrective measures to mitigate adverse outcomes. Using a proxy measure of extreme non-adherence and a dataset of nearly $700,000$ patients from four states in India, we formulate and solve the machine learning (ML) problem of early prediction of non-adherence based on a custom rank-based metric. We train ML models and evaluate against baselines, achieving a $\sim 100\%$ lift over rule-based baselines and $\sim 214\%$ over a random classifier, taking into account country-wide large-scale future deployment. We deal with various issues in the process, including data quality, high-cardinality categorical data, low target prevalence, distribution shift, variation across cohorts, algorithmic fairness, and the need for robustness and explainability. Our findings indicate that risk stratification of non-adherent patients is a viable, deployable-at-scale ML solution. As the official AI partner of India's Central TB Division, we are working on multiple city and state-level pilots with the goal of pan-India deployment.

cs.LG↗

arXiv2021

Interpretability of Epidemiological Models : The Curse of Non-Identifiability

Interpretability of epidemiological models is a key consideration, especially when these models are used in a public health setting. Interpretability is strongly linked to the identifiability of the underlying model parameters, i.e., the ability to estimate parameter values with high confidence given observations. In this paper, we define three separate notions of identifiability that explore the different roles played by the model definition, the loss function, the fitting methodology, and the quality and quantity of data. We define an epidemiological compartmental model framework in which we highlight these non-identifiability issues and their mitigation.

stat.ME↗

arXiv2007

Molecular Clock on a Neutral Network

The number of fixed mutations accumulated in an evolving population often displays a variance that is significantly larger than the mean (the overdispersed molecular clock). By examining a generic evolutionary process on a neutral network of high-fitness genotypes, we establish a formalism for computing all cumulants of the full probability distribution of accumulated mutations in terms of graph properties of the neutral network, and use the formalism to prove overdispersion of the molecular clock. We further show that significant overdispersion arises naturally in evolution when the neutral network is highly sparse, exhibits large global fluctuations in neutrality, and small local fluctuations in neutrality. The results are also relevant for elucidating the topological structure of a neutral network from empirical measurements of the substitution process.

q-bio.PE↗

arXiv2007

Evolution favors protein mutational robustness in sufficiently large populations

BACKGROUND: An important question is whether evolution favors properties such as mutational robustness or evolvability that do not directly benefit any individual, but can influence the course of future evolution. Functionally similar proteins can differ substantially in their robustness to mutations and capacity to evolve new functions, but it has remained unclear whether any of these differences might be due to evolutionary selection for these properties. RESULTS: Here we use laboratory experiments to demonstrate that evolution favors protein mutational robustness if the evolving population is sufficiently large. We neutrally evolve cytochrome P450 proteins under identical selection pressures and mutation rates in populations of different sizes, and show that proteins from the larger and thus more polymorphic population tend towards higher mutational robustness. Proteins from the larger population also evolve greater stability, a biophysical property that is known to enhance both mutational robustness and evolvability. The excess mutational robustness and stability is well described by existing mathematical theories, and can be quantitatively related to the way that the proteins occupy their neutral network. CONCLUSIONS: Our work is the first experimental demonstration of the general tendency of evolution to favor mutational robustness and protein stability in highly polymorphic populations. We suggest that this phenomenon may contribute to the mutational robustness and evolvability of viruses and bacteria that exist in large populations.

q-bio.PE↗

arXiv2007

Thermodynamics of Neutral Protein Evolution

Naturally evolving proteins gradually accumulate mutations while continuing to fold to thermodynamically stable native structures. This process of neutral protein evolution is an important mode of genetic change, and forms the basis for the molecular clock. Here we present a mathematical theory that predicts the number of accumulated mutations, the index of dispersion, and the distribution of stabilities in an evolving protein population from knowledge of the stability effects (ddG values) for single mutations. Our theory quantitatively describes how neutral evolution leads to marginally stable proteins, and provides formulae for calculating how fluctuations in stability cause an overdispersion of the molecular clock. It also shows that the structural influences on the rate of sequence evolution that have been observed in earlier simulations can be calculated using only the single-mutation ddG values. We consider both the case when the product of the population size and mutation rate is small and the case when this product is large, and show that in the latter case proteins evolve excess mutational robustness that is manifested by extra stability and increases the rate of sequence evolution. Our basic method is to treat protein evolution as a Markov process constrained by a minimal requirement for stable folding, enabling an evolutionary description of the proteins solely in terms of the experimentally measureable ddG values. All of our theoretical predictions are confirmed by simulations with model lattice proteins. Our work provides a mathematical foundation for understanding how protein biophysics helps shape the process of evolution.

q-bio.PE↗

arXiv2005

A single determinant for the rate of yeast protein evolution

A gene's rate of sequence evolution is among the most fundamental evolutionary quantities in common use, but what determines evolutionary rates has remained unclear. Here, we show that the two most commonly used methods to disentangle the determinants of evolutionary rate, partial correlation analysis and ordinary multivariate regression, produce misleading or spurious results when applied to noisy biological data. To overcome these difficulties, we employ an alternative method, principal component regression, which is a multivariate regression of evolutionary rate against the principal components of the predictor variables. We carry out the first combined analysis of seven predictors (gene expression level, dispensability, protein abundance, codon adaptation index, gene length, number of protein-protein interactions, and the gene's centrality in the interaction network). Strikingly, our analysis reveals a single dominant component which explains 40-fold more variation in evolutionary rate than any other, suggesting that protein evolutionary rate has a single determinant among the seven predictors. The dominant component explains nearly half the variation in the rate of synonymous and protein evolution. Our results support the hypothesis that selection against the cost of translation-error-induced protein misfolding governs the rate of synonymous and protein sequence evolution in yeast.

q-bio.PE↗

arXiv2003

Some asymptotic properties of duplication graphs

Duplication graphs are graphs that grow by duplication of existing vertices, and are important models of biological networks, including protein-protein interaction networks and gene regulatory networks. Three models of graph growth are studied: pure duplication growth, and two two-parameter models in which duplication forms one element of the growth dynamics. A power-law degree distribution is found to emerge in all three models. However, the parameter space of the latter two models is characterized by a range of parameter values for which duplication is the predominant mechanism of graph growth. For parameter values that lie in this ``duplication-dominated'' regime, it is shown that the degree distribution either approaches zero asymptotically, or approaches a non-zero power-law degree distribution very slowly. In either case, the approach to the true asymptotic degree distribution is characterized by a dependence of the scaling exponent on properties of the initial degree distribution. It is therefore conjectured that duplication-dominated, scale-free networks may contain identifiable remnants of their early structure. This feature is inherited from the idealized model of pure duplication growth, for which the exact finite-size degree distribution is found and its asymptotic properties studied.

cond-mat.stat-mech↗

arXiv2002

Black Hole Fluctuations and Backreaction in Stochastic Gravity

We present a framework for analyzing black hole backreaction from the point of view of quantum open systems using influence functional formalism. We focus on the model of a black hole described by a radially perturbed quasi-static metric and Hawking radiation by a conformally coupled massless quantum scalar field. It is shown that the closed-time-path (CTP) effective action yields a non-local dissipation term as well as a stochastic noise term in the equation of motion, the Einstein-Langevin equation. Once the thermal Green's function in a Schwarzschild background becomes available to the required accuracy the strategy described here can be applied to obtain concrete results on backreaction. We also present an alternative derivation of the CTP effective action in terms of the Bogolyubov coefficients, thus making a connection with the interpretation of the noise term as measuring the difference in particle production in alternative histories.

gr-qc↗

arXiv2000

Is there emitted radiation in Unruh effect?

The thermal radiance felt by a uniformly accelerated detector/oscillator/atom--the Unruh effect-- is often mistaken to be some emitted radiation detectable by an observer/probe/sensor. Here we show by an explicit calculation of the energy momentum tensor of a quantum scalar field that, at least in 1+1 dimension, while a polarization cloud is found to exist around the particle trajectory, there is no emitted radiation from a uniformly accelerated oscillator in equilibrium conditions. Under nonequilibrium conditions which can prevail for non-uniformly accelerated trajectories or before the atom or oscillator reaches equilibrium, there is conceivably radiation emitted, but that is not what Unruh effect entails.

quant-ph↗

arXiv2000

New quantum aspects of a vacuum-dominated universe

In a new model that we proposed, nonperturbative vacuum contributions to the effective action of a free quantized massive scalar field lead to a cosmological solution in which the scalar curvature becomes constant after a time $t_j$ (when the redshift $z \sim 1$) that depends on the mass of the scalar field and its curvature coupling. This spatially-flat solution implies an accelerating universe at the present time and gives a good one-parameter fit to high-redshift Type Ia supernovae (SNe-Ia) data, and the present age and energy density of the universe. Here we show that the imaginary part of the nonperturbative curvature term that causes the cosmological acceleration, implies a particle production rate that agrees with predictions of other methods and extends them to non-zero mass fields. The particle production rate is very small after the transition and is not expected to alter the nature of the cosmological solution. We also show that the equation of state of our model undergoes a transition at $t_j$ from an equation of state dominated by non-relativistic pressureless matter (without a cosmological constant) to an effective equation of state of mixed radiation and cosmological constant, and we derive the equation of state of the vacuum. Finally, we explain why nonperturbative vacuum effects of this ultralow-mass particle do not significantly change standard early universe cosmology.

gr-qc↗

arXiv1999

Vacuum-driven Metamorphosis

We show that nonperturbative vacuum effects can produce a vacuum-driven transition from a matter-dominated universe to one in which the effective equation of state is that of radiation plus cosmological constant. The actual material content of the universe after the transition remains that of non-relativistic matter. This metamorphosis of the equation of state can be traced to nonperturbative vacuum effects that cause the scalar curvature to remain nearly constant at a well-defined value after the transition, and is responsible for the observed acceleration of the recent expansion of the universe.

gr-qc↗

arXiv1999

Vacuum effects of ultra-low mass particle account for Recent Acceleration of Universe

In recent work, we showed that non-perturbative vacuum effects of a very low mass particle could induce, at a redshift of order 1, a transition from a matter-dominated to an accelerating universe. In that work, we used the simplification of a sudden transition out of the matter-dominated stage and were able to fit the Type Ia supernovae (SNe-Ia) data points with a spatially-open universe. In the present work, we find a more accurate, smooth {\it spatially-flat} analytic solution to the quantum-corrected Einstein equations. This solution gives a good fit to the SNe-Ia data with a particle mass parameter $m_h$ in the range $6.40 \times 10^{-33}$ eV to $7.25 \times 10^{-33}$ eV. It follows that the ratio of total matter density (including dark matter) to critical density, $Ø_0$, is in the range 0.58 to 0.15, and the age $t_0$ of the universe is in the range $8.10 h^{-1}$ Gyr to $12.2 h^{-1}$ Gyr, where $h$ is the present value of the Hubble constant, measured as a fraction of the value 100 km/(s Mpc). This spatially-flat model agrees with estimates of the position of the first acoustic peak in the small angular scale fluctuations of the cosmic background radiation, and with light-element abundances of standard big-bang nucleosynthesis. Our model has only a single free parameter, $m_h$, and does not require that we live at a special time in the evolution of the universe.

gr-qc↗

arXiv1999

Non-perturbative effects of vacuum energy on the recent expansion of the universe

We show that the vacuum energy of a free quantized field of very low mass can significantly alter the recent expansion of the universe. The effective action of the theory is obtained from a non-perturbative sum of scalar curvature terms in the propagator. We numerically investigate the semiclassical Einstein equations derived from it. As a result of non-perturbative quantum effects, the scalar curvature of the matter-dominated universe stops decreasing and approaches a constant value. The universe in our model evolves from an open matter-dominated epoch to a mildly inflating de Sitter expansion. The Hubble constant during the present de Sitter epoch, as well as the time at which the transition occurs from matter-dominated to de Sitter expansion, are determined by the mass of the field and by the present matter density. The model provides a theoretical explanation of the observed recent acceleration of the universe, and gives a good fit to data from high-redshift Type Ia supernovae, with a mass of about 10^{-33} eV, and a current ratio of matter density to critical density, Omega_0 <0.4 . The age of the universe then follows with no further free parameters in the theory, and turns out to be greater than 13 Gyr. The model is spatially open and consistent with the possibility of inflation in the very early universe. Furthermore, our model arises from the standard renormalizable theory of a free quantum field in curved spacetime, and does not require a cosmological constant or the associated fine-tuning.

gr-qc↗

arXiv1999

Notes on Black Hole Fluctuations and Backreaction

In these notes we prepare the ground for a systematic investigation into the issues of black hole fluctuations and backreaction by discussing the formulation of the problem, commenting on possible advantages and shortcomings of existing works, and introducing our own approach via a stochastic semiclassical theory of gravity based on the Einstein-Langevin equation and the fluctuation-dissipation relation for a self-consistent description of metric fluctuations and dissipative dynamics of the black hole with backreaction of its Hawking radiance.

gr-qc↗

arXiv1998

Relation Between Einstein And Quantum Field Equations

We show that there exists a choice of scalar field modes, such that the evolution of the quantum field in the zero-mass and large-mass limits is consistent with the Einstein equations for the background geometry. This choice of modes is also consistent with zero production of these particles and thus corresponds to a preferred vacuum state preserved by the evolution. In the zero-mass limit, we find that the quantum field equation implies the Einstein equation for the scale factor of a radiation-dominated universe; in the large-mass case, it implies the corresponding Einstein equation for a matter-dominated universe. Conversely, if the classical radiation-dominated or matter-dominated Einstein equations hold, there is no production of scalar particles in the zero and large mass limits, respectively. The suppression of particle production in the large mass limit is over and above the expected suppression at large mass. Our results hold for a certain class of conformally ultrastatic background geometries and therefore generalize previous results by one of us for spatially flat Robertson-Walker background geometries. In these geometries, we find that the temporal part of the graviton equations reduces to the temporal equation for a massless minimally coupled scalar field, and therefore the results for massless particle production hold also for gravitons. Within the class of modes we study, we also find that the requirement of zero production of massless scalar particles is not consistent with a non-zero cosmological constant. Possible implications are discussed.

gr-qc↗

arXiv1997

Coherence and Fluctuations in the Interaction between Moving Atoms and a Quantum Field

Mesoscopic physics deals with three fundamental issues: quantum coherence, fluctuations and correlations. Here we analyze these issues for atom optics, using a simplified model of an assembly of atoms (or detectors, which are particles with some internal degree of freedom) moving in arbitrary trajectories in a quantum field. Employing the influence functional formalism, we study the self-consistent effect of the field on the atoms, and their mutual interactions via coupling to the field. We derive the coupled Langevin equations for the atom assemblage and analyze the relation of dissipative dynamics of the atoms with the correlation and fluctuations of the quantum field. This provides a useful theoretical framework for analysing the coherent properties of atom-field systems.

quant-ph↗

arXiv1997

Thermal Particle Creation in Cosmological Spacetimes: A Stochastic Approach

The stochastic method based on the influence functional formalism introduced in an earlier paper to treat particle creation in near-uniformly accelerated detectors and collapsing masses is applied here to treat thermal and near-thermal radiance in certain types of cosmological expansions. It is indicated how the appearance of thermal radiance in different cosmological spacetimes and in the two apparently distinct classes of black hole and cosmological spacetimes can be understood under a unifying conceptual and methodological framework.

gr-qc↗