DJDavid E. Jones

David E. Jones

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


David E. Jones has a total of 43 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Causal Estimation of User Learning in Personalized Systems

    by Evan Munro, David Jones, Jennifer Brennan, Roland Nelet, Vahab Mirrokni, Jean Pouget-Abadie

    In online platforms, the impact of a treatment on an observed outcome may change over time as 1) users learn about the intervention, and 2) the system personalization, such as individualized recommendations, change over time. We introduce a non-parametric causal model of user actions in a personalized system. We show that the Cookie-Cookie-Day (CCD) experiment, designed for the measurement of the user learning effect, is biased when there is personalization. We derive new experimental designs that intervene in the personalization system to generate the variation necessary to separately identify the causal effect mediated through user learning and personalization. Making parametric assumptions allows for the estimation of long-term causal effects based on medium-term experiments. In simulations, we show that our new designs successfully recover the dynamic causal effects of interest.


  • Open Access

    Quantifying Observed Prior Impact

    by David E. Jones, Robert N. Trangucci, Yang Chen

    When summarizing a Bayesian analysis, it is important to quantify the contribution of the prior distribution to the final posterior inference because this informs other researchers whether the prior information needs to be carefully scrutinized, and whether alternative priors are likely to substantially alter the conclusions drawn. One appealing and interpretable way to do this is to report an effective prior sample size (EPSS), which captures how many observations the information in the prior distribution corresponds to. However, typically the most important aspect of the prior distribution is its location relative to the data, and therefore traditional information measures are somewhat deficit for the purpose of quantifying EPSS, because they concentrate on the variance or spread of the prior distribution (in isolation from the data). To partially address this difficulty, Reimherr et al. (2014) introduced a class of EPSS measures based on prior-likelihood discordance. In this paper, we take this idea further by proposing a new measure of EPSS that not only incorporates the general mathematical form of the likelihood (as proposed by Reimherr et al., 2014) but also the specific data at hand. Thus, our measure considers the location of the prior relative to the current observed data, rather than relative to the average of multiple datasets from the working model, the latter being the approach taken by Reimherr et al. (2014). Consequently, our measure can be highly variable, but we demonstrate that this is because the impact of a prior on a Bayesian analysis can intrinsically be highly variable. Our measure is called the (posterior) mean Observed Prior Effective Sample Size (mOPESS), and is a Bayes estimate of a meaningful quantity. The mOPESS well communicates the extent to which inference is determined by the prior, or framed differently, the amount of sampling effort saved due to having relevant prior information. We illustrate our ideas through a number of examples including Gaussian conjugate and non-conjugate models (continuous observations), a Beta-Binomial model (discrete observations), and a linear regression model (two unknown parameters).


  • Functional PCA With Covariate-Dependent Mean and Covariance Structure

    by Fei Ding, Shiyuan He, David E. Jones, Jianhua Z. Huang

    Incorporating covariates into functional principal component analysis (PCA) can substantially improve the representation efficiency of the principal components and predictive performance. However, many existing functional PCA methods do not make use of covariates, and those that do often have high computational cost or make overly simplistic assumptions that are violated in practice. In this article, we propose a new framework, called Covariate Dependent Functional Principal Component Analysis (CD-FPCA), in which both the mean and covariance structure depend on covariates. We propose a corresponding estimation algorithm, which makes use of spline basis representations and roughness penalties, and is substantially more computationally efficient than competing approaches of adequate estimation and prediction accuracy. A key aspect of our work is our novel approach for modeling the covariance function and ensuring that it is symmetric positive semi-definite. We demonstrate the advantages of our methodology through a simulation study and an astronomical data analysis.


  • Open Access

    Toward Extremely Precise Radial Velocities: II. A Tool For Using Multivariate Gaussian Processes to Model Stellar Activity

    by Christian Gilbertson, Eric B. Ford, David E. Jones, David C. Stenning

    The radial velocity method is one of the most successful techniques for the discovery and characterization of exoplanets. Modern spectrographs promise measurement precision of ~0.2-0.5 m/s for an ideal target star. However, the intrinsic variability of stellar spectra can mimic and obscure true planet signals at these levels. Rajpaul et al. (2015) and Jones et al. (2017) proposed applying a physically motivated, multivariate Gaussian process (GP) to jointly model the apparent Doppler shift and multiple indicators of stellar activity as a function of time, so as to separate the planetary signal from various forms of stellar variability. These methods are promising, but performing the necessary calculations can be computationally intensive and algebraically tedious. In this work, we present a flexible and computationally efficient software package, GPLinearOdeMaker.jl, for modeling multivariate time series using a linear combination of univariate GPs and their derivatives. The package allows users to easily and efficiently apply various multivariate GP models and different covariance kernel functions. We demonstrate GPLinearOdeMaker.jl by applying the Jones et al. (2017) model to fit measurements of the apparent Doppler shift and activity indicators derived from simulated active solar spectra time series affected by many evolving star spots. We show how GPLinearOdeMaker.jl makes it easy to explore the effect of different choices for the GP kernel. We find that local kernels could significantly increase the sensitivity and precision of Doppler planet searches relative to the widely used quasi-periodic kernel.

  • Toward Extremely Precise Radial Velocities. II. A Tool for Using Multivariate Gaussian Processes to Model Stellar Activity

    by Christian Gilbertson, Eric B. Ford, David E. Jones, David C. Stenning

    The radial velocity method is one of the most successful techniques for the discovery and characterization of exoplanets. Modern spectrographs promise measurement precision of 0.2–0.5 m s−1 for an ideal target star. However, the intrinsic variability of stellar spectra can mimic and obscure true planet signals at these levels. Rajpaul et al. and Jones et al. proposed applying a physically motivated, multivariate Gaussian process (GP) to jointly model the apparent Doppler shift and multiple indicators of stellar activity as a function of time, so as to separate the planetary signal from various forms of stellar variability. These methods are promising, but performing the necessary calculations can be computationally intensive and algebraically tedious. In this work, we present a flexible and computationally efficient software package, GPLinearODEMaker.jl, for modeling multivariate time series using a linear combination of univariate GPs and their derivatives. The package allows users to easily and efficiently apply various multivariate GP models and different covariance kernel functions. We demonstrate GPLinearODEMaker.jl by applying the Jones et al. model to fit measurements of the apparent Doppler shift and activity indicators derived from simulated active solar spectra time series affected by many evolving starspots. We show how GPLinearODEMaker.jl makes it easy to explore the effect of different choices for the GP kernel. We find that local kernels could significantly increase the sensitivity and precision of Doppler planet searches relative to the widely used quasiperiodic kernel.


  • Warp Bridge Sampling: The Next Generation

    by Lazhi Wang, David E. Jones, Xiao-Li Meng

    Bridge sampling is an effective Monte Carlo method for estimating the ratio of normalizing constants of two probability densities, a routine computational problem in statistics, physics, chemistry, and other fields. The Monte Carlo error of the bridge sampling estimator is determined by the amount of overlap between the two densities. In the case of uni-modal densities, Warp-I, II, and III transformations (Meng and Schilling, 2002) are effective for increasing the initial overlap, but they are less so for multi-modal densities. This paper introduces Warp-U transformations that aim to transform multi-modal densities into uni-modal ones (hence ‘U’) without altering their normalizing constants. The construction of a Warp-U transformation starts with a Normal (or other convenient) mixture distribution ϕmix that has reasonable overlap with the target density p, whose normalizing constant is unknown. The stochastic transformation that maps ϕmix back to its generating distribution N(0,1) is then applied to p yielding its Warp-U version, which we denote p˜. Typically, p˜ is uni-modal and has substantially increased overlap with ϕ. Furthermore, we prove that the overlap between p˜ and N(0,1) is guaranteed to be no less than the overlap between p and ϕmix, in terms of any f-divergence. We propose a computationally efficient method to find an appropriate ϕmix, and a simple but effective approach to remove the bias which results from estimating the normalizing constants and fitting ϕmix with the same data. We illustrate our findings using 10 and 50 dimensional highly irregular multi-modal densities, and demonstrate how Warp-U sampling can be used to improve the final estimation step of the Generalized Wang-Landau algorithm (Liang, 2005), a powerful sampling and estimation approach.


  • The AMIGA sample of isolated galaxies

    by M. G. Jones, D. Espada, L. Verdes-Montenegro, W. K. Huchtmeier, U. Lisenfeld, S. Leon, J. Sulentic, J. Sabater, D. E. Jones, S. Sanchez, et al.

    Context. We present the largest catalogue of HI single dish observations of isolated galaxies to date, as part of the multi-wavelength compilation being performed by the AMIGA project (Analysis of the interstellar Medium in Isolated GAlaxies). Despite numerous studies of the HI content of galaxies, no revision focused on the HI scaling relations of the most isolated L∗ galaxies has been made since Haynes & Giovanelli (1984, AJ, 89, 758). Aims. The AMIGA sample has been demonstrated to be almost “nurture free”, therefore, by creating scaling relations for the HI content of these galaxies we will define a metric of HI normalcy in the absence of interactions. Methods. The catalogue comprises of our own HI observations with Arecibo, Effelsberg, Nançay and GBT, and spectra collected from the literature. In total we have measurements or constraints on the HI masses of 844 galaxies from the Catalogue of Isolated Galaxies (CIG). The multi-wavelength AMIGA dataset includes a revision of the B-band luminosities (LB), optical diameters (D25), morphologies, and isolation. Due to the large size of the catalogue, these revisions permit cuts to be made to ensure isolation and a high level of completeness, which was not previously possible. With this refined dataset we fit HI scaling relations based on luminosity, optical diameter and morphology. Our regression model incorporates all the data, including upper limits, and accounts for uncertainties in both variables, as well as distance uncertainties. Results. The scaling relation of HI mass with D25 is in good agreement with that of Haynes & Giovanelli (1984), but our relation with LB is considerably steeper. This disagreement is attributed to the large uncertainties in the luminosities, which introduce a bias when fitting with ordinary least squares regression (as was done in previous works), and the different morphology distributions of the samples. We find that the main effect of morphology on the D25-relation is to increase the intercept towards later types, while for the LB-relation it is to flatten the slope. These trends were not evident in previous works due to the small number of detected early-type galaxies. Applying our relations to HI detected galaxies in the Virgo cluster we find that although the typical HI-deficiency is only ~0.3 dex, the tail of the distribution extends over an order of magnitude beyond that of the AMIGA sample. These results are in general agreement with previous studies of HI-deficiency in the Virgo cluster. Conclusions. The HI scaling relations of the AMIGA sample define an up-to-date metric of the HI content of almost “nurture free” galaxies. These relations allow the expected HI mass, in the absence of interactions, of an individual galaxy to be predicted to within 0.25 dex (for typical measurement uncertainties). These relations are thus suitable for use as statistical measures of the impact of interactions on the neutral gas content of galaxies.


  • Determination of thickness of microwaveable multicompartment meals using dielectric, thermal, and physical properties

    by Jiajia Chen, Ron Lentz, Peter Pesheck, Ashu Guru, David Jones, Yiwen Li, Jeyamkondan Subbiah


  • Open Access

    DISENTANGLING OVERLAPPING ASTRONOMICAL SOURCES USING SPATIAL AND SPECTRAL INFORMATION

    by David E. Jones, Vinay L. Kashyap, David A. van Dyk

    We present a powerful new algorithm that combines both spatial information (event locations and the point-spread function) and spectral information (photon energies) to separate photons from overlapping sources. We use Bayesian statistical methods to simultaneously infer the number of overlapping sources, to probabilistically separate the photons among the sources, and to fit the parameters describing the individual sources. Using the Bayesian joint posterior distribution, we are able to coherently quantify the uncertainties associated with all these parameters. The advantages of combining spatial and spectral information are demonstrated through a simulation study. The utility of the approach is then illustrated by analysis of observations of FK Aqr and FL Aqr with the XMM-Newton Observatory and the central region of the Orion Nebula Cluster with the Chandra X-ray Observatory.