Papers by Pavlos Protopapas
arXiv (Cornell University), May 21, 2009
Catalogs of periodic variable stars contain large numbers of periodic light-curves (photometric t... more Catalogs of periodic variable stars contain large numbers of periodic light-curves (photometric time series data from the astrophysics domain). Separating anomalous objects from well-known classes is an important step towards the discovery of new classes of astronomical objects. Most anomaly detection methods for time series data assume either a single continuous time series or a set of time series whose periods are aligned. Light-curve data precludes the use of these methods as the periods of any given pair of light-curves may be out of sync. One may use an existing anomaly detection method if, prior to similarity calculation, one performs the costly act of aligning two light-curves, an operation that scales poorly to massive data sets. This paper presents PCAD, an unsupervised anomaly detection method for large sets of unsynchronized periodic time-series data, that outputs a ranked list of both global and local anomalies. It calculates its anomaly score for each light-curve in relation to a set of centroids produced by a modified k-means clustering algorithm. Our method is able to scale to large data sets through the use of sampling. We validate our method on both light-curve data and other time series data sets. We demonstrate its effectiveness at finding known anomalies, and discuss the effect of sample size and number of centroids on our results. We compare our method to naive solutions and existing time series anomaly detection Editor: Tow Fawcett.
ArXiv, 2018
In this paper we propose a data augmentation method for time series with irregular sampling, Time... more In this paper we propose a data augmentation method for time series with irregular sampling, Time-Conditional Generative Adversarial Network (T-CGAN). Our approach is based on Conditional Generative Adversarial Networks (CGAN), where the generative step is implemented by a deconvolutional NN and the discriminative step by a convolutional NN. Both the generator and the discriminator are conditioned on the sampling timestamps, to learn the hidden relationship between data and timestamps, and consequently to generate new time series. We evaluate our model with synthetic and real-world datasets. For the synthetic data, we compare the performance of a classifier trained with T-CGAN-generated data, against the performance of the same classifier trained on the original data. Results show that classifiers trained on T-CGAN-generated data perform the same as classifiers trained on real data, even with very short time series and small training sets. For the real world datasets, we compare our...
arXiv (Cornell University), Nov 7, 2021
Uncertainty quantification (UQ) helps to make trustworthy predictions based on collected observat... more Uncertainty quantification (UQ) helps to make trustworthy predictions based on collected observations and uncertain domain knowledge. With increased usage of deep learning in various applications, the need for efficient UQ methods that can make deep models more reliable has increased as well. Among applications that can benefit from effective handling of uncertainty are the deep learning based differential equation (DE) solvers. We adapt several state-of-the-art UQ methods to get the predictive uncertainty for DE solutions and show the results on four different DE types.
Monthly Notices of the Royal Astronomical Society, 2019
ABSTRACTIn the last years, automatic classification of variable stars has received substantial at... more ABSTRACTIn the last years, automatic classification of variable stars has received substantial attention. Using machine learning techniques for this task has proven to be quite useful. Typically, machine learning classifiers used for this task require to have a fixed training set, and the training process is performed offline. Upcoming surveys such as the Large Synoptic Survey Telescope will generate new observations daily, where an automatic classification system able to create alerts online will be mandatory. A system with those characteristics must be able to update itself incrementally. Unfortunately, after training, most machine learning classifiers do not support the inclusion of new observations in light curves, they need to re-train from scratch. Naively re-training from scratch is not an option in streaming settings, mainly because of the expensive pre-processing routines required to obtain a vector representation of light curves (features) each time we include new observat...
The Astrophysical Journal, 2016
The need for the development of automatic tools to explore astronomical databases has been recogn... more The need for the development of automatic tools to explore astronomical databases has been recognized since the inception of CCDs and modern computers. Astronomers already have developed solutions to tackle several science problems, such as automatic classification of stellar objects, outlier detection, and globular clusters identification, among others. New scientific problems emerge, and it is critical to be able to reuse the models learned before, without rebuilding everything from the beginning when the sciencientific problem changes. In this paper, we propose a new meta-model that automatically integrates existing classification models of variable stars. The proposed meta-model incorporates existing models that are trained in a different context, answering different questions and using different representations of data. A conventional mixture of expert algorithms in machine learning literature cannot be used since each expert (model) uses different inputs. We also consider the computational complexity of the model by using the most expensive models only when it is necessary. We test our model with EROS-2 and MACHO data sets, and we show that we solve most of the classification challenges only by training a meta-model to learn how to integrate the previous experts.
Physical Review D, 2000
We present our first results of numerical simulations of two skyrmion dynamics using an ωmeson st... more We present our first results of numerical simulations of two skyrmion dynamics using an ωmeson stabilized effective Lagrangian. We consider skyrmion-skyrmion scattering with a fixed initial velocity of β = 0.5, for various impact parameters and groomings. The physical picture that emerges is surprisingly rich, while consistent with previous results and general conservation laws. We find meson radiation, skyrmion scattering out of the scattering plane, orbiting and capture to bound states.
arXiv (Cornell University), Dec 5, 2018
Turbulence modeling is a classical approach to address the multiscale nature of fluid turbulence.... more Turbulence modeling is a classical approach to address the multiscale nature of fluid turbulence. Instead of resolving all scales of motion, which is currently mathematically and numerically intractable, reduced models that capture the large-scale behavior are derived. One of the most popular reduced models is the Reynolds averaged Navier-Stokes (RANS) equations. The goal is to solve the RANS equations for the mean velocity and pressure field. However, the RANS equations contain a term called the Reynolds stress tensor, which is not known in terms of the mean velocity field. Many RANS turbulence models have been proposed to model the Reynolds stress tensor in terms of the mean velocity field, but are usually not suitably general for all flow fields of interest. Data-driven turbulence models have recently garnered considerable attention and have been rapidly developed. In a seminal work, Ling et al (2016) developed the tensor basis neural network (TBNN), which was used to learn a general Galilean invariant model for the Reynolds stress tensor. The TBNN was applied to a variety of flow fields with encouraging results. In the present study, the TBNN is applied to the turbulent channel flow. Its performance is compared with classical turbulence models as well as a neural network model that does not preserve Galilean invariance. A sensitivity study on the TBNN reveals that the network attempts to adjust to the dataset, but is limited by the mathematical form that guarantees Galilean invariance.
ArXiv, 2020
Solutions to differential equations are of significant scientific and engineering relevance. Rece... more Solutions to differential equations are of significant scientific and engineering relevance. Recently, there has been a growing interest in solving differential equations with neural networks. This work develops a novel method for solving differential equations with unsupervised neural networks that applies Generative Adversarial Networks (GANs) to \emph{learn the loss function} for optimizing the neural network. We present empirical results showing that our method, which we call Differential Equation GAN (DEQGAN), can obtain multiple orders of magnitude lower mean squared errors than an alternative unsupervised neural network method based on (squared) $L_2$, $L_1$, and Huber loss functions. Moreover, we show that DEQGAN achieves solution accuracy that is competitive with traditional numerical methods. Finally, we analyze the stability of our approach and find it to be sensitive to the selection of hyperparameters, which we provide in the appendix. Code available at this https URL. ...
Journal of Turbulence, 2021
The Reynolds-averaged Navier-Stokes (RANS) equations are widely used in turbulence applications. ... more The Reynolds-averaged Navier-Stokes (RANS) equations are widely used in turbulence applications. They require accurately modeling the anisotropic Reynolds stress tensor, for which traditional Reynolds stress closure models only yield reliable results in some flow configurations. In the last few years, there has been a surge of work aiming at using datadriven approaches to tackle this problem. The majority of previous work has focused on the development of fully-connected networks for modeling the anisotropic Reynolds stress tensor. In this paper, we expand upon recent work for turbulent channel flow and develop new convolutional neural network (CNN) models that are able to accurately predict the normalized anisotropic Reynolds stress tensor. We apply the new CNN model to a number of one-dimensional turbulent flows. Additionally, we present interpretability techniques that help drive the model design and provide guidance on the model behavior in relation to the underlying physics.
The Astronomical Journal, 2021
Characterizing the fundamental parameters of stars from observations is crucial for studying the ... more Characterizing the fundamental parameters of stars from observations is crucial for studying the stars themselves, their planets, and the galaxy as a whole. Stellar evolution theory predicting the properties of stars as a function of stellar age and mass enables translating observables into physical stellar parameters by fitting the observed data to synthetic isochrones. However, the complexity of overlapping evolutionary tracks often makes this task numerically challenging, and with a precision that can be highly variable, depending on the area of the parameter space the observation lies in. This work presents StelNet, a Deep Neural Network trained on stellar evolutionary tracks that quickly and accurately predicts mass and age from absolute luminosity and effective temperature for stars with close to solar metallicity. The underlying model makes no assumption on the evolutionary stage and includes the pre-main sequence phase. We use bootstrapping and train many models to quantify the uncertainty of the model. To break the model's intrinsic degeneracy resulting from overlapping evolutionary paths, we also built a hierarchical model that retrieves realistic posterior probability distributions of the stellar mass and age. We further test and train StelNet using a sample of stars with well-determined masses and ages from the literature.
Physical Review Letters, 1997
Core-particle coupling models are made viable by assuming that core properties such as matrix ele... more Core-particle coupling models are made viable by assuming that core properties such as matrix elements of multipole and pairing operators and excitation spectra are known independently. From the completeness relation, it is seen, however, that these quantities are themselves algebraic functions of the calculated core-particle amplitudes. For the deformed rare-earth nucleus 158 Gd, we find that these sum rules are well-satisfied for the ground state band, implying that we have found a self-consistent solution of the non-linear Kerman-Klein equations.
Physical Review C, 1997
The most consistently useful simple model for the study of odd deformed nuclei, the particle-roto... more The most consistently useful simple model for the study of odd deformed nuclei, the particle-rotor model (strong coupling limit of the core-particle coupling model) has nevertheless been beset by a long-standing problem: It is necessary in many cases to introduce an ad hoc parameter that reduces the size of the Coriolis interaction coupling the collective and single-particle motions. Of the numerous suggestions put forward for the origin of this supplementary interaction, none of those actually tested by calculations has been accepted as the solution of the problem. In this paper we seek a solution of the difficulty within the framework of a general formalism that starts from the spherical shell model and is capable of treating an arbitrary linear combination of multipole and pairing forces. With the restriction of the interaction to the familiar sum of a quadrupole multipole force and a monopole pairing force, we have previously studied a semi-microscopic version of the formalism whose framework is nevertheless more comprehensive than any previously applied to the problem. We obtained solutions for low-lying bands of several strongly deformed odd rare earth nuclei and found good agreement with experiment, except for an exaggerated staggering of levels for K = 1 2 bands, which can be understood as a manifestation of the Coriolis attenuation problem. We argue that within the formalism utilized, the only way to improve the physics is to add interactions to the model Hamiltonian. We verify that by adding a magnetic dipole interaction of essentially fixed strength, we can fit the K = 1 2 bands without destroying the agreement with other bands. In addition we show that our solution also fits 163 Er, a classic test case of Coriolis attenuation that we had not previously studied.
Physical Review C, 2004
The Kerman-Klein-Dönau-Frauendorf (KKDF) model is a linearized version of the nonlinear Kerman-Kl... more The Kerman-Klein-Dönau-Frauendorf (KKDF) model is a linearized version of the nonlinear Kerman-Klein (equations of motion) formulation of the nuclear many-body problem. In practice, it is a generalization of the standard core-particle coupling model that, like the latter, provides a description of the spectroscopy of odd nuclei in terms of the corresponding properties of neighboring even nuclei and of single-particle properties, which are the input parameters of the model. A divers sample of recent applications attests to the usefulness of the model. In this paper, we first present a concise general review of the fundamental equations and properties of the KKDF model. We then derive a corresponding formalism for odd-odd nuclei with proton-neutron number ͑Z , N͒ that relates their properties to those of the four neighboring even-even nuclei ͑Z +1,N +1͒, ͑Z −1,N +1͒, ͑Z +1,N −1͒, and ͑Z −1,N −1͒, all of which are required if one is to include both multipole and pairing forces. We treat these equations in two ways. In the first, we make essential use of the solutions of the neighboring odd nucleus problem, as obtained by the KKDF method. In the second, we relate the properties of the odd-odd nucleus directly to those of the even-even nuclei. For both choices, we derive equations of motion, normalization conditions, and an expression for transition amplitudes. We also resolve the problem of choosing the subspace of physical solutions that arises in an equation of motion approach that includes pairing interactions.
Physical Review C, 1996
In a previous paper a semi-microscopic core-particle coupling method that includes the convention... more In a previous paper a semi-microscopic core-particle coupling method that includes the conventional strong coupling core-particle model as a limiting case, was applied to spectra and electromagnetic properties of several well-deformed odd nuclei. This work, coupled a large single-particle space to the ground state bands of the neighboring even cores. In this paper, we generalize the theory to include excited bands of the cores, such as beta and gamma bands, and thereby show that the resulting theory can account for the location and structure of all bands up to about 1.5 MeV.
arXiv (Cornell University), Oct 21, 2021
Solving differential equations efficiently and accurately sits at the heart of progress in many a... more Solving differential equations efficiently and accurately sits at the heart of progress in many areas of scientific research, from classical dynamical systems to quantum mechanics. There is a surge of interest in using Physics-Informed Neural Networks (PINNs) to tackle such problems as they provide numerous benefits over traditional numerical approaches. Despite their potential benefits for solving differential equations, transfer learning has been under explored. In this study, we present a general framework for transfer learning PINNs that results in one-shot inference for linear systems of both ordinary and partial differential equations. This means that highly accurate solutions to many unknown differential equations can be obtained instantaneously without retraining an entire network. We demonstrate the efficacy of the proposed deep learning approach by solving several real-world problems, such as first-and second-order linear ordinary equations, the Poisson equation, and the time-dependent Schrödinger complex-value partial differential equation.
ArXiv, 2021
There is a wave of interest in using unsupervised neural networks for solving differential equati... more There is a wave of interest in using unsupervised neural networks for solving differential equations. The existing methods are based on feed-forward networks, while recurrent neural network differential equation solvers have not yet been reported. We introduce an unsupervised reservoir computing (RC), an echo-state recurrent neural network capable of discovering approximate solutions that satisfy ordinary differential equations (ODEs). We suggest an approach to calculate time derivatives of recurrent neural network outputs without using backpropagation. The internal weights of an RC are fixed, while only a linear output layer is trained, yielding efficient training. However, RC performance strongly depends on finding the optimal hyper-parameters, which is a computationally expensive process. We use Bayesian optimization to efficiently discover optimal sets in a high-dimensional hyper-parameter space and numerically show that one set is robust and can be used to solve an ODE for diff...
arXiv (Cornell University), Apr 18, 2019
Neural networks are a central technique in machine learning. Recent years have seen a wave of int... more Neural networks are a central technique in machine learning. Recent years have seen a wave of interest in applying neural networks to physical systems for which the governing dynamics are known and expressed through differential equations. Two fundamental challenges facing the development of neural networks in physics applications is their lack of interpretability and their physics-agnostic design. The focus of the present work is to embed physical constraints into the structure of the neural network to address the second fundamental challenge. By constraining tunable parameters (such as weights and biases) and adding special layers to the network, the desired constraints are guaranteed to be satisfied without the need for explicit regularization terms. This is demonstrated on supervised and unsupervised networks for two basic symmetries: even/odd symmetry of a function and energy conservation. In the supervised case, the network with embedded constraints is shown to perform well on regression problems while simultaneously obeying the desired constraints whereas a traditional network fits the data but violates the underlying constraints. Finally, a new unsupervised neural network is proposed that guarantees energy conservation through an embedded symplectic structure. The symplectic neural network is used to solve a system of energy-conserving differential equations and outperforms an unsupervised, non-symplectic neural network.
Physical Review C, 1997
We review briefly the fundamental equations of a semi-microscopic coreparticle coupling method th... more We review briefly the fundamental equations of a semi-microscopic coreparticle coupling method that makes no reference to an intrinsic system of coordinates. We then demonstrate how an intrinsic system can be introduced in the strong coupling limit so as to yield a completely equivalent formulation. It is emphasized that the conventional core-particle coupling calculation introduces a further approximation that avoids what has hitherto been the most time-consuming feature of the full theory, and that this approximation can be introduced either in the intrinsic system, the usual case, or in the laboratory system, our preference. A new algorithm is described for the full theory that largely removes the difference in complexity between the two types of calculation. Comparison of the full and approximate theories for some representative cases provides a basis for the assessment of the accuracy of the traditional approach. We find that for well-deformed nuclei, e.g. 157 Gd and 157 Tb, the core-coupling method and the full theory give similar results.
Uploads
Papers by Pavlos Protopapas