Cecilia R. Aragon

Optimization by simulated annealing: an experimental evaluation. Part I, graph partitioning

In this and two companion papers, we report on an extended empirical study of the simulated annealing approach to combinatorial optimization proposed by S. Kirkpatrick et al. That study investigated how best to adapt simulated annealing to particular problems and compared its performance to that of more traditional algorithms. This paper (Part I) discusses annealing and our parameterized generic implementation of it, describes how we adapted this generic algorithm to the graph partitioning problem, and reports how well it compared to standard algorithms like the Kernighan-Lin algorithm. (For sparse random graphs, it tended to outperform Kernighan-Lin as the number of vertices become large, even when its much greater running time was taken into account. It did not perform nearly so well, however, on graphs generated with a built-in geometric structure.) We also discuss how we went about optimizing our implementation, and describe the effects of changing the various annealing parameters or varying the basic annealing algorithm itself.

Optimization by simulated annealing: an experimental evaluation; part II, graph coloring and number partitioning

This is the second in a series of three papers that empirically examine the competitiveness of simulated annealing in certain well-studied domains of combinatorial optimization. Simulated annealing is a randomized technique proposed by S. Kirkpatrick, C. D. Gelatt and M. P. Vecchi for improving local optimization algorithms. Here we report on experiments at adapting simulated annealing to graph coloring and number partitioning, two problems for which local optimization had not previously been thought suitable. For graph coloring, we report on three simulated annealing schemes, all of which can dominate traditional techniques for certain types of graphs, at least when large amounts of computing time are available. For number partitioning, simulated annealing is not competitive with the differencing algorithm of N. Karmarkar and R. M. Karp, except on relatively small instances. Moreover, if running time is taken into account, natural annealing schemes cannot even outperform multiple random runs of the local optimization algorithms on which they are based, in sharp contrast to the observed performance of annealing on other problems.

Randomized search trees

A randomized strategy for maintaining balance in dynamically changing search trees that has optimal expected behavior is presented. In particular, in the expected case an update takes logarithmic time and requires fewer than two rotations. Moreover, the update time remains logarithmic, even if the cost of a rotation is taken to be proportional to the size of the rotated subtree. The approach generalizes naturally to weighted trees, where the expected time bounds for accesses and updates again match the worst case time bounds of the best deterministic methods. The balancing strategy and algorithms are exceedingly simple and should be fast in practice.<<ETX>>

The reddening law of type Ia supernovae: separating intrinsic variability from dust using equivalent widths

We employ 76 type Ia supernovae (SNe Ia) with optical spectrophotometry within 2.5 days of B-band maximum light obtained by the Nearby Supernova Factory to derive the impact of Si and Ca features on the supernovae intrinsic luminosity and determine a dust reddening law. We use the equivalent width of Si II λ4131 in place of the light curve stretch to account for first-order intrinsic luminosity variability. The resulting empirical spectral reddening law exhibits strong features that are associated with Ca II and Si II λ6355. After applying a correction based on the Ca II H&K equivalent width we find a reddening law consistent with a Cardelli extinction law. Using the same input data, we compare this result to synthetic rest-frame UBVRI-like photometry to mimic literature observations. After corrections for signatures correlated with Si II λ4131 and Ca II H&K equivalent widths and introducing an empirical correlation between colors, we determine the dust component in each band. We find a value of the total-to-selective extinction ratio, R v = 2.8 ± 0.3. This agrees with the Milky Way value, in contrast to the low R v values found in most previous analyses. This result suggests that the long-standing controversy in interpreting SN Ia colors and their compatibility with a classical extinction law, which is critical to their use as cosmological probes, can be explained by the treatment of the dispersion in colors, and by the variability of features apparent in SN Ia spectra.

HOST GALAXY PROPERTIES AND HUBBLE RESIDUALS OF TYPE Ia SUPERNOVAE FROM THE NEARBY SUPERNOVA FACTORY

We examine the relationship between Type Ia supernova (SN Ia) Hubble residuals and the properties of their host galaxies using a sample of 115 SNe Ia from the Nearby Supernova Factory. We use host galaxy stellar masses and specific star formation rates fitted from photometry for all hosts, as well as gas-phase metallicities for a subset of 69 star-forming (non-active galactic nucleus) hosts, to show that the SN Ia Hubble residuals correlate with each of these host properties. With these data we find new evidence for a correlation between SN Ia intrinsic color and host metallicity. When we combine our data with those of other published SN Ia surveys, we find the difference between mean SN Ia brightnesses in low- and high-mass hosts is 0.077 ? 0.014?mag. When viewed in narrow (0.2?dex) bins of host stellar mass, the data reveal apparent plateaus of Hubble residuals at high and low host masses with a rapid transition over a short mass range (9.8 ? log (M */M ?) ? 10.4). Although metallicity has been a favored interpretation for the origin of the Hubble residual trend with host mass, we illustrate how dust in star-forming galaxies and mean SN Ia progenitor age both evolve along the galaxy mass sequence, thereby presenting equally viable explanations for some or all of the observed SN Ia host bias.

A tale of two online communities: fostering collaboration and creativity in scientists and children

There has been much recent interest in the development of tools to foster remote collaboration and shared creative work. An open question is: what are the guidelines for this process? What are the key socio-technical preconditions required for a geographically distributed group to collaborate effectively on creative work, and are they different from the conditions of a decade or two ago? In an attempt to answer these questions, we conducted empirical studies of two seemingly very different online communities, both requiring effective collaboration and creative work: an international collaboration of astrophysicists studying supernovae to learn more about the expansion rate of the universe, and a group of children, ages 8-15, from different parts of the world, creating and sharing animated stories and video games on the Scratch online community developed at MIT. Both groups produced creative technical work jointly and were considered successful in their communities. Data included the analysis of thousands of lines from chat and comment logs over a period of several months, and interviews with community members. We discovered some surprising commonalities and some intriguing possibilities, and suggest guidelines for successful creative collaborations. Specifically, systems that support social creativity must facilitate sharing and play, and their design must consider the effects of repurposing, augmentation and behavior adaptation.

Evidence of environmental dependencies of Type Ia supernovae from the Nearby Supernova Factory indicated by local H alpha

Context. Use of Type Ia supernovae (SNe Ia) as distance indicators has proven to be a powerful technique for measuring the darkenergy equation of state. However, recent studies have highlighted potential biases correlated with the global properties of their host galaxies, large enough to induce systematic errors into such cosmological measurements if not properly treated. Aims. We study the host galaxy regions in close proximity to SNe Ia in order to analyze relations between the properties of SN Ia events and environments where their progenitors most likely formed. In this paper we focus on local H emission as an indicator of young progenitor environments. Methods. The Nearby Supernova Factory has obtained flux-calibrated spectral timeseries for SNe Ia using integral field spectroscopy. These observations enabled the simultaneous measurement of the SN and its immediate vicinity. For 89 SNe Ia we measured or set limits on H emission, used as a tracer of ongoing star formation, within a 1 kpc radius around each SN. This constitutes the first direct study of the local environment for a large sample of SNe Ia with accurate luminosity, color, and stretch measurements. Results. Our local star formation measurements provide several critical new insights. We find that SNe Ia with local H emission are redder by 0:036 0:017 mag, and that the previously noted correlation between stretch and host mass is driven entirely by the SNe Ia coming from locally passive environments, in particular at the low-stretch end. There is no such trend for SNe Ia in locally star-forming environments. Our most important finding is that the mean standardized brightness for SNe Ia with local H emission is 0:094 0:031 mag fainter on average than for those without. This o set arises from a bimodal structure in the Hubble residuals, with one mode being shared by SNe Ia in all environments and the other one exclusive to SNe Ia in locally passive environments. This structure also explains the previously known host-mass bias. We combine the star formation dependence of this bimodality with the cosmic star formation rate to predict changes with redshift in the mean SN Ia brightness and the host-mass bias. The strong change predicted is confirmed using high-redshift SNe Ia from the literature. Conclusions. The environmental dependences in SN Ia Hubble residuals and color found here point to remaining systematic errors in the standardization of SNe Ia. In particular, the observed brightness o set associated with local H emission is predicted to cause a significant bias in current measurements of the dark energy equation of state. Recognition of these e ects o ers new opportunities to improve SNe Ia as cosmological probes. For instance, we note that the SNe Ia associated with local H emission are more homogeneous, resulting in a brightness dispersion of only 0:105 0:012 mag.

Biometric authentication via oculomotor plant characteristics

A novel biometrics approach that performs authentication via the internal non-visible anatomical structure of an individual human eye is proposed and evaluated. To provide authentication, the proposed method estimates the anatomical characteristics of the oculomotor plant (comprising the eye globe, its muscles and the brains control signals). The estimation of the oculomotor plant characteristics (OPC) is achieved by analyzing the recorded eye movement trajectories via a 2D linear homeomorphic mathematical representation of the oculomotor plant. The derived OPC allow authentication via various statistical methods and information fusion techniques. The proposed authentication method yielded Half Total Error Rate of 19% for a pool of 59 recorded subjects in the best case. The OPC biometric authentication has high counterfeit resistance potential, because it includes both behavioral and physiological human attributes that are hard to reproduce.

Type Ia supernova bolometric light curves and ejected mass estimates from the Nearby Supernova Factory

We present a sample of normal type Ia supernovae from the Nearby Supernova Factory dataset with spectrophotometry at sufficiently late phases to estim ate the ejected mass using the bolometric light curve. We measure 56 Ni masses from the peak bolometric luminosity, then compare the luminosity in the 56 Co-decay tail to the expected rate of radioactive energy release from ejecta of a given mass. We infer the ejected mass in a Bayesian context using a semi-analytic model of the ejecta, incorporating constra ints from contemporary numerical models as priors on the density structure and distribution o f 56 Ni throughout the ejecta. We find a strong correlation between ejected mass and light curv e decline rate, and consequently 56 Ni mass, with ejected masses in our data ranging from 0.9‐1.4 M⊙. Most fast-declining (SALT2x1 < 1) normal SNe Ia have significantly sub-Chandrasekhar ejecte d masses in our fiducial analysis.

Atmospheric extinction properties above Mauna Kea from the Nearby SuperNova Factory spectro-photometric data set

We present a new atmospheric extinction curve for Mauna Kea spanning 3200--9700 \AA. It is the most comprehensive to date, being based on some 4285 standard star spectra obtained on 478 nights spread over a period of 7 years obtained by the Nearby SuperNova Factory using the SuperNova Integral Field Spectrograph. This mean curve and its dispersion can be used as an aid in calibrating spectroscopic or imaging data from Mauna Kea, and in estimating the calibration uncertainty associated with the use of a mean extinction curve. Our method for decomposing the extinction curve into physical components, and the ability to determine the chromatic portion of the extinction even on cloudy nights, is described and verified over the wide range of conditions sampled by our large dataset. We demonstrate good agreement with atmospheric science data obtain at nearby Mauna Loa Observatory, and with previously published measurements of the extinction above Mauna Kea.