Peter Nugent

Bio: Peter Nugent is an academic researcher from Lawrence Berkeley National Laboratory. The author has contributed to research in topics: Supernova & Light curve. The author has an hindex of 127, co-authored 754 publications receiving 92988 citations. Previous affiliations of Peter Nugent include Liverpool John Moores University & National Autonomous University of Mexico.

The Case for Strong Scaling in Deep Learning: Training Large 3D CNNs with Hybrid Parallelism

Abstract: We present scalable hybrid-parallel algorithms for training large-scale 3D convolutional neural networks. Deep learning-based emerging scientific workflows often require model training with large, high-dimensional samples, which can make training much more costly and even infeasible due to excessive memory usage. We solve these challenges by extensively applying hybrid parallelism throughout the end-to-end training pipeline, including both computations and I/O. Our hybrid-parallel algorithm extends the standard data parallelism with spatial parallelism, which partitions a single sample in the spatial domain, realizing strong scaling beyond the mini-batch dimension with a larger aggregated memory capacity. We evaluate our proposed training algorithms with two challenging 3D CNNs, CosmoFlow and 3D U-Net. Our comprehensive performance studies show that good weak and strong scaling can be achieved for both networks using up 2K GPUs. More importantly, we enable training of CosmoFlow with much larger samples than previously possible, realizing an order-of-magnitude improvement in prediction accuracy.

Real-Time Recovery Efficiencies and Performance of the Palomar Transient Factory's Transient Discovery Pipeline

Abstract: We present the transient source detection efficiencies of the Palomar Transient Factory (PTF), parameterizing the number of transients that PTF found, versus the number of similar transients that occurred over the same period in the survey search area but that were missed. PTF was an optical sky survey carried out with the Palomar 48-inch telescope over 2009-2012, observing more than 8000 square degrees of sky with cadences of between 1 and 5 days, locating around 50,000 non-moving transient sources, and spectroscopically confirming around 1900 supernovae. We assess the effectiveness with which PTF detected transient sources, by inserting ~7 million artificial point sources into real PTF data. We then study the efficiency with which the PTF real-time pipeline recovered these sources as a function of the source magnitude, host galaxy surface brightness, and various observing conditions (using proxies for seeing, sky brightness, and transparency). The product of this study is a multi-dimensional recovery efficiency grid appropriate for the range of observing conditions that PTF experienced, and that can then be used for studies of the rates, environments, and luminosity functions of different transient types using detailed Monte Carlo simulations. We illustrate the technique using the observationally well-understood class of type Ia supernovae.

Carnegie Supernova Project-II: The Near-infrared Spectroscopy Program

Abstract: Shifting the focus of Type Ia supernova (SN Ia) cosmology to the near-infrared (NIR) is a promising way to significantly reduce the systematic errors, as the strategy minimizes our reliance on the empirical width-luminosity relation and uncertain dust laws. Observations in the NIR are also crucial for our understanding of the origins and evolution of these events, further improving their cosmological utility. Any future experiments in the rest-frame NIR will require knowledge of the SN Ia NIR spectroscopic diversity, which is currently based on a small sample of observed spectra. Along with the accompanying paper, Phillips et al. (2018), we introduce the Carnegie Supernova Project-II (CSP-II), to follow up nearby SNe Ia in both the optical and the NIR. In particular, this paper focuses on the CSP-II NIR spectroscopy program, describing the survey strategy, instrumental setups, data reduction, sample characteristics, and future analyses on the data set. In collaboration with the Harvard-Smithsonian Center for Astrophysics (CfA) Supernova Group, we obtained 661 NIR spectra of 157 SNe Ia. Within this sample, 451 NIR spectra of 90 SNe Ia have corresponding CSP-II follow-up light curves. Such a sample will allow detailed studies of the NIR spectroscopic properties of SNe Ia, providing a different perspective on the properties of the unburned material, radioactive and stable nickel produced, progenitor magnetic fields, and searches for possible signatures of companion stars.

iPTF Survey for Cool Transients

Abstract: We performed a wide-area (2000 deg^2) g and I band experiment as part of a two month extension to the Intermediate Palomar Transient Factory. We discovered 36 extragalactic transients including iPTF17lf, a highly reddened local SN Ia, iPTF17bkj, a new member of the rare class of transitional Ibn/IIn supernovae, and iPTF17be, a candidate luminous blue variable outburst. We do not detect any luminous red novae and place an upper limit on their rate. We show that adding a slow-cadence I band component to upcoming surveys such as the Zwicky Transient Facility will improve the photometric selection of cool and dusty transients.

Searching for Planet Nine with Coadded WISE and NEOWISE-Reactivation Images

Abstract: A distant, as yet unseen ninth planet has been invoked to explain various observations of the outer solar system. While such a 'Planet Nine', if it exists, is most likely to be discovered via reflected light in the optical, it may emit much more strongly at 3$-$5$\mu$m than simple blackbody predictions would suggest, depending on its atmospheric properties (Fortney et al. 2016). As a result, Planet Nine may be detectable at 3.4$\mu$m with WISE, but single exposures are too shallow except at relatively small distances ($d_9 \lesssim 430$ AU). We develop a method to search for Planet Nine far beyond the W1 single-exposure sensitivity, to distances as large as 800 AU, using inertial coadds of W1 exposures binned into $\sim$1 day intervals. We apply our methodology to $\sim$2000 square degrees of sky identified by Holman & Payne (2016) as a potentially likely Planet Nine location, based on the Fienga et al. (2016) Cassini ranging analysis. We do not detect a plausible Planet Nine candidate, but are able to derive a detailed completeness curve, ruling out its presence within the parameter space searched at $W1 < 16.66$ (90% completeness). Our method uses all publicly available W1 imaging, spanning 2010 January to 2015 December, and will become more sensitive with future NEOWISE-Reactivation releases of additional W1 exposures. We anticipate that our method will be applicable to the entire high Galactic latitude sky, and we will extend our search to that full footprint in the near future.

Measurements of Omega and Lambda from 42 High-Redshift Supernovae

Abstract: We report measurements of the mass density, Omega_M, and cosmological-constant energy density, Omega_Lambda, of the universe based on the analysis of 42 Type Ia supernovae discovered by the Supernova Cosmology Project. The magnitude-redshift data for these SNe, at redshifts between 0.18 and 0.83, are fit jointly with a set of SNe from the Calan/Tololo Supernova Survey, at redshifts below 0.1, to yield values for the cosmological parameters. All SN peak magnitudes are standardized using a SN Ia lightcurve width-luminosity relation. The measurement yields a joint probability distribution of the cosmological parameters that is approximated by the relation 0.8 Omega_M - 0.6 Omega_Lambda ~= -0.2 +/- 0.1 in the region of interest (Omega_M <~ 1.5). For a flat (Omega_M + Omega_Lambda = 1) cosmology we find Omega_M = 0.28{+0.09,-0.08} (1 sigma statistical) {+0.05,-0.04} (identified systematics). The data are strongly inconsistent with a Lambda = 0 flat cosmology, the simplest inflationary universe model. An open, Lambda = 0 cosmology also does not fit the data well: the data indicate that the cosmological constant is non-zero and positive, with a confidence of P(Lambda > 0) = 99%, including the identified systematic uncertainties. The best-fit age of the universe relative to the Hubble time is t_0 = 14.9{+1.4,-1.1} (0.63/h) Gyr for a flat cosmology. The size of our sample allows us to perform a variety of statistical tests to check for possible systematic errors and biases. We find no significant differences in either the host reddening distribution or Malmquist bias between the low-redshift Calan/Tololo sample and our high-redshift sample. The conclusions are robust whether or not a width-luminosity relation is used to standardize the SN peak magnitudes.

Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant

Abstract: We present spectral and photometric observations of 10 Type Ia supernovae (SNe Ia) in the redshift range 0.16 " z " 0.62. The luminosity distances of these objects are determined by methods that employ relations between SN Ia luminosity and light curve shape. Combined with previous data from our High-z Supernova Search Team and recent results by Riess et al., this expanded set of 16 high-redshift supernovae and a set of 34 nearby supernovae are used to place constraints on the following cosmo- logical parameters: the Hubble constant the mass density the cosmological constant (i.e., the (H 0 ), () M ), vacuum energy density, the deceleration parameter and the dynamical age of the universe ) " ), (q 0 ), ) M \ 1) methods. We estimate the dynamical age of the universe to be 14.2 ^ 1.7 Gyr including systematic uncer- tainties in the current Cepheid distance scale. We estimate the likely e†ect of several sources of system- atic error, including progenitor and metallicity evolution, extinction, sample selection bias, local perturbations in the expansion rate, gravitational lensing, and sample contamination. Presently, none of these e†ects appear to reconcile the data with and ) " \ 0 q 0 " 0.

Machine learning

Abstract: Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Seven-year wilkinson microwave anisotropy probe (wmap *) observations: cosmological interpretation

Abstract: The combination of seven-year data from WMAP and improved astrophysical data rigorously tests the standard cosmological model and places new constraints on its basic parameters and extensions. By combining the WMAP data with the latest distance measurements from the baryon acoustic oscillations (BAO) in the distribution of galaxies and the Hubble constant (H0) measurement, we determine the parameters of the simplest six-parameter ΛCDM model. The power-law index of the primordial power spectrum is ns = 0.968 ± 0.012 (68% CL) for this data combination, a measurement that excludes the Harrison–Zel’dovich–Peebles spectrum by 99.5% CL. The other parameters, including those beyond the minimal set, are also consistent with, and improved from, the five-year results. We find no convincing deviations from the minimal model. The seven-year temperature power spectrum gives a better determination of the third acoustic peak, which results in a better determination of the redshift of the matter-radiation equality epoch. Notable examples of improved parameters are the total mass of neutrinos, � mν < 0.58 eV (95% CL), and the effective number of neutrino species, Neff = 4.34 +0.86 −0.88 (68% CL), which benefit from better determinations of the third peak and H0. The limit on a constant dark energy equation of state parameter from WMAP+BAO+H0, without high-redshift Type Ia supernovae, is w =− 1.10 ± 0.14 (68% CL). We detect the effect of primordial helium on the temperature power spectrum and provide a new test of big bang nucleosynthesis by measuring Yp = 0.326 ± 0.075 (68% CL). We detect, and show on the map for the first time, the tangential and radial polarization patterns around hot and cold spots of temperature fluctuations, an important test of physical processes at z = 1090 and the dominance of adiabatic scalar fluctuations. The seven-year polarization data have significantly improved: we now detect the temperature–E-mode polarization cross power spectrum at 21σ , compared with 13σ from the five-year data. With the seven-year temperature–B-mode cross power spectrum, the limit on a rotation of the polarization plane due to potential parity-violating effects has improved by 38% to Δα =− 1. 1 ± 1. 4(statistical) ± 1. 5(systematic) (68% CL). We report significant detections of the Sunyaev–Zel’dovich (SZ) effect at the locations of known clusters of galaxies. The measured SZ signal agrees well with the expected signal from the X-ray data on a cluster-by-cluster basis. However, it is a factor of 0.5–0.7 times the predictions from “universal profile” of Arnaud et al., analytical models, and hydrodynamical simulations. We find, for the first time in the SZ effect, a significant difference between the cooling-flow and non-cooling-flow clusters (or relaxed and non-relaxed clusters), which can explain some of the discrepancy. This lower amplitude is consistent with the lower-than-theoretically expected SZ power spectrum recently measured by the South Pole Telescope Collaboration.

Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological Constant To Appear in the Astronomical Journal

Abstract: We present spectral and photometric observations of 10 type Ia supernovae (SNe Ia) in the redshift range 0.16 � z � 0.62. The luminosity distances of these objects are determined by methods that employ relations between SN Ia luminosity and light curve shape. Combined with previous data from our High-Z Supernova Search Team (Garnavich et al. 1998; Schmidt et al. 1998) and Riess et al. (1998a), this expanded set of 16 high-redshift supernovae and a set of 34 nearby supernovae are used to place constraints on the following cosmological parameters: the Hubble constant (H0), the mass density (M), the cosmological constant (i.e., the vacuum energy density, �), the deceleration parameter (q0), and the dynamical age of the Universe (t0). The distances of the high-redshift SNe Ia are, on average, 10% to 15% farther than expected in a low mass density (M = 0.2) Universe without a cosmological constant. Different light curve fitting methods, SN Ia subsamples, and prior constraints unanimously favor eternally expanding models with positive cosmological constant (i.e., � > 0) and a current acceleration of the expansion (i.e., q0 < 0). With no prior constraint on mass density other than M � 0, the spectroscopically confirmed SNe Ia are statistically consistent with q0 < 0 at the 2.8�