Showing 1–11 of 11 results for author: Graham, D R

Utilitarian Algorithm Configuration

Authors: Devon R. Graham, Kevin Leyton-Brown, Tim Roughgarden

Abstract: We present the first nontrivial procedure for configuring heuristic algorithms to maximize the utility provided to their end users while also offering theoretical guarantees about performance. Existing procedures seek configurations that minimize expected runtime. However, very recent theoretical work argues that expected runtime minimization fails to capture algorithm designers' preferences. Here… ▽ More We present the first nontrivial procedure for configuring heuristic algorithms to maximize the utility provided to their end users while also offering theoretical guarantees about performance. Existing procedures seek configurations that minimize expected runtime. However, very recent theoretical work argues that expected runtime minimization fails to capture algorithm designers' preferences. Here we show that the utilitarian objective also confers significant algorithmic benefits. Intuitively, this is because mean runtime is dominated by extremely long runs even when they are incredibly rare; indeed, even when an algorithm never gives rise to such long runs, configuration procedures that provably minimize mean runtime must perform a huge number of experiments to demonstrate this fact. In contrast, utility is bounded and monotonically decreasing in runtime, allowing for meaningful empirical bounds on a configuration's performance. This paper builds on this idea to describe effective and theoretically sound configuration procedures. We prove upper bounds on the runtime of these procedures that are similar to theoretical lower bounds, while also demonstrating their performance empirically. △ Less

Submitted 31 October, 2023; originally announced October 2023.

Formalizing Preferences Over Runtime Distributions

Authors: Devon R. Graham, Kevin Leyton-Brown, Tim Roughgarden

Abstract: When trying to solve a computational problem, we are often faced with a choice between algorithms that are guaranteed to return the right answer but differ in their runtime distributions (e.g., SAT solvers, sorting algorithms). This paper aims to lay theoretical foundations for such choices by formalizing preferences over runtime distributions. It might seem that we should simply prefer the algori… ▽ More When trying to solve a computational problem, we are often faced with a choice between algorithms that are guaranteed to return the right answer but differ in their runtime distributions (e.g., SAT solvers, sorting algorithms). This paper aims to lay theoretical foundations for such choices by formalizing preferences over runtime distributions. It might seem that we should simply prefer the algorithm that minimizes expected runtime. However, such preferences would be driven by exactly how slow our algorithm is on bad inputs, whereas in practice we are typically willing to cut off occasional, sufficiently long runs before they finish. We propose a principled alternative, taking a utility-theoretic approach to characterize the scoring functions that describe preferences over algorithms. These functions depend on the way our value for solving our problem decreases with time and on the distribution from which captimes are drawn. We describe examples of realistic utility functions and show how to leverage a maximum-entropy approach for modeling underspecified captime distributions. Finally, we show how to efficiently estimate an algorithm's expected utility from runtime samples. △ Less

Submitted 2 June, 2023; v1 submitted 25 May, 2022; originally announced May 2022.

High-resolution observations of the solar photosphere, chromosphere and transition region. A database of coordinated IRIS and SST observations

Authors: L. H. M. Rouppe van der Voort, B. De Pontieu, M. Carlsson, J. de la Cruz Rodriguez, S. Bose, G. Chintzoglou, A. Drews, C. Froment, M. Gosic, D. R. Graham, V. H. Hansteen, V. M. J. Henriques, S. Jafarzadeh, J. Joshi, L. Kleint, P. Kohutova, T. Leifsen, J. Martinez-Sykora, D. Nobrega-Siverio, A. Ortiz, T. M. D. Pereira, A. Popovas, C. Quintero Noda, A. Sainz Dalda, G. B. Scharmer , et al. (8 additional authors not shown)

Abstract: NASA's Interface Region Imaging Spectrograph (IRIS) provides high resolution observations of the solar atmosphere through UV spectroscopy and imaging. Since the launch of IRIS in June 2013, we have conducted systematic observation campaigns in coordination with the Swedish 1-m Solar Telescope (SST) on La Palma. The SST provides complementary high-resolution observations of the photosphere and chro… ▽ More NASA's Interface Region Imaging Spectrograph (IRIS) provides high resolution observations of the solar atmosphere through UV spectroscopy and imaging. Since the launch of IRIS in June 2013, we have conducted systematic observation campaigns in coordination with the Swedish 1-m Solar Telescope (SST) on La Palma. The SST provides complementary high-resolution observations of the photosphere and chromosphere. The SST observations include spectro-polarimetric imaging in photospheric Fe I lines and spectrally-resolved imaging in the chromospheric Ca II 8542 A, H-alpha, and Ca II K lines. We present a database of co-aligned IRIS and SST datasets that is open for analysis to the scientific community. The database covers a variety of targets including active regions, sunspots, plage, quiet Sun, and coronal holes. △ Less

Submitted 29 July, 2020; v1 submitted 28 May, 2020; originally announced May 2020.

Comments: Accepted for publication in Astronomy & Astrophysics (July 21, 2020)

Journal ref: A&A 641, A146 (2020)

Spectral signatures of chromospheric condensation in a major solar flare

Authors: David R. Graham, Gianna Cauzzi, Luca Zangrilli, Adam Kowalski, Paulo Simões, Joel Allred

Abstract: We study the evolution of chromospheric line and continuum emission during the impulsive phase of the X-class SOL2014-09-10T17:45 solar flare. We extend previous analyses of this flare to multiple chromospheric lines of Fe I, Fe II, Mg II, C I, and Si II, observed with IRIS, combined with radiative-hydrodynamical (RHD) modeling. For multiple flaring kernels, the lines all show a rapidly evolving d… ▽ More We study the evolution of chromospheric line and continuum emission during the impulsive phase of the X-class SOL2014-09-10T17:45 solar flare. We extend previous analyses of this flare to multiple chromospheric lines of Fe I, Fe II, Mg II, C I, and Si II, observed with IRIS, combined with radiative-hydrodynamical (RHD) modeling. For multiple flaring kernels, the lines all show a rapidly evolving double-component structure: an enhanced, emission component at rest, and a broad, highly red-shifted component of comparable intensity. The red-shifted components migrate from 25-50 km s$^{-1}$ towards the rest wavelength within $\sim$30 seconds. Using Fermi hard X-ray observations, we derive the parameters of an accelerated electron beam impacting the dense chromosphere, using them to drive a RHD simulation with the RADYN code. As in Kowalski et al. 2017a, our simulations show that the most energetic electrons penetrate into the deep chromosphere, heating it to T$\sim$10,000 K, while the bulk of the electrons dissipate their energy higher, driving an explosive evaporation, and its counterpart condensation -- a very dense (n$_e \sim 2 \times 10^{14}$ cm$^{-3}$), thin layer (30--40 km thickness), heated to 8--12,000 K, moving towards the stationary chromosphere at up to 50 km s$^{-1}$. The synthetic Fe II 2814.45A profiles closely resemble the observational data, including a continuum enhancement, and both a stationary and a highly red-shifted component, rapidly moving towards the rest wavelength. Importantly, the absolute continuum intensity, ratio of component intensities, relative time of appearance, and red-shift amplitude, are sensitive to the model input parameters, showing great potential as diagnostics. △ Less

Submitted 10 April, 2020; originally announced April 2020.

Comments: Accepted for publication by the Astrophysical Journal - 9th April 2020

Impulsive coronal heating from large-scale magnetic rearrangements: from IRIS to SDO/AIA

Authors: Fabio Reale, Paola Testa, Antonino Petralia, David R. Graham

Abstract: The Interface Region Imaging Spectrograph (IRIS) has observed bright spots at the transition region footpoints associated with heating in the overlying loops, as observed by coronal imagers. Some of these brightenings show significant blueshifts in the Si iv line at 1402.77 A (logT[K] = 4.9). Such blueshifts cannot be reproduced by coronal loop models assuming heating by thermal conduction only, b… ▽ More The Interface Region Imaging Spectrograph (IRIS) has observed bright spots at the transition region footpoints associated with heating in the overlying loops, as observed by coronal imagers. Some of these brightenings show significant blueshifts in the Si iv line at 1402.77 A (logT[K] = 4.9). Such blueshifts cannot be reproduced by coronal loop models assuming heating by thermal conduction only, but are consistent with electron beam heating, highlighting for the first time the possible importance of non-thermal electrons in the heating of non-flaring active regions. Here we report on the coronal counterparts of these brightenings observed in the hot channels of the Atmospheric Imaging Assembly (AIA) on board the Solar Dynamics Observatory. We show that the IRIS bright spots are the footpoints of very hot and transient coronal loops which clearly experience strong magnetic interactions and rearrangements, thus confirming the impulsive nature of the heating and providing important constraints for their physical interpretation. △ Less

Submitted 4 July, 2019; originally announced July 2019.

Comments: 13 pages, 9 figures, accepted by ApJ, movie available upon request to F. Reale

Deep Models of Interactions Across Sets

Authors: Jason Hartford, Devon R Graham, Kevin Leyton-Brown, Siamak Ravanbakhsh

Abstract: We use deep learning to model interactions across two or more sets of objects, such as user-movie ratings, protein-drug bindings, or ternary user-item-tag interactions. The canonical representation of such interactions is a matrix (or a higher-dimensional tensor) with an exchangeability property: the encoding's meaning is not changed by permuting rows or columns. We argue that models should hence… ▽ More We use deep learning to model interactions across two or more sets of objects, such as user-movie ratings, protein-drug bindings, or ternary user-item-tag interactions. The canonical representation of such interactions is a matrix (or a higher-dimensional tensor) with an exchangeability property: the encoding's meaning is not changed by permuting rows or columns. We argue that models should hence be Permutation Equivariant (PE): constrained to make the same predictions across such permutations. We present a parameter-sharing scheme and prove that it could not be made any more expressive without violating PE. This scheme yields three benefits. First, we demonstrate state-of-the-art performance on multiple matrix completion benchmarks. Second, our models require a number of parameters independent of the numbers of objects, and thus scale well to large datasets. Third, models can be queried about new objects that were not available at training time, but for which interactions have since been observed. In experiments, our models achieved surprisingly good generalization performance on this matrix extrapolation task, both within domains (e.g., new users and new movies drawn from the same distribution used for training) and even across domains (e.g., predicting music ratings after training on movies). △ Less

Submitted 8 June, 2018; v1 submitted 7 March, 2018; originally announced March 2018.

Temporal evolution of multiple evaporating ribbon sources in a solar flare

Authors: D. R. Graham, G. Cauzzi

Abstract: We present new results from the Interface Region Imaging Spectrograph showing the dynamic evolution of chromospheric evaporation and condensation in a flare ribbon, with the highest temporal and spatial resolution to date. IRIS observed the entire impulsive phase of the X-class flare SOL2014-09-10T17:45 using a 9.4 second cadence `sit-and-stare' mode. As the ribbon brightened successively at new p… ▽ More We present new results from the Interface Region Imaging Spectrograph showing the dynamic evolution of chromospheric evaporation and condensation in a flare ribbon, with the highest temporal and spatial resolution to date. IRIS observed the entire impulsive phase of the X-class flare SOL2014-09-10T17:45 using a 9.4 second cadence `sit-and-stare' mode. As the ribbon brightened successively at new positions along the slit, a unique impulsive phase evolution was observed for many tens of individual pixels in both coronal and chromospheric lines. Each activation of a new footpoint displays the same initial coronal up-flows of up to ~300 km/s, and chromospheric downflows up to 40 km/s. Although the coronal flows can be delayed by over 1 minute with respect to those in the chromosphere, the temporal evolution of flows is strikingly similar between all pixels, and consistent with predictions from hydrodynamic flare models. Given the large sample of independent footpoints, we conclude that each flaring pixel can be considered a prototypical, `elementary' flare kernel. △ Less

Submitted 10 June, 2015; originally announced June 2015.

Comments: Accepted for publication - ApJ letters

Impulsive Heating of Solar Flare Ribbons Above 10 MK

Authors: Paulo J. A. Simões, David R. Graham, Lyndsay Fletcher

Abstract: The chromospheric response to the input of flare energy is marked by extended extreme ultraviolet (EUV) ribbons and hard X-ray (HXR) footpoints. These are usually explained as the result of heating and bremsstrahlung emission from accelerated electrons colliding in the dense chromospheric plasma. We present evidence of impulsive heating of flare ribbons above 10 MK in a two-ribbon flare. We analys… ▽ More The chromospheric response to the input of flare energy is marked by extended extreme ultraviolet (EUV) ribbons and hard X-ray (HXR) footpoints. These are usually explained as the result of heating and bremsstrahlung emission from accelerated electrons colliding in the dense chromospheric plasma. We present evidence of impulsive heating of flare ribbons above 10 MK in a two-ribbon flare. We analyse the impulsive phase of SOL2013-11-09T06:38, a C2.6 class event using data from Atmospheric Imaging Assembly (AIA) on board of Solar Dynamics Observatory (SDO) and the Reuven Ramaty High Energy Solar Spectroscopic Imager (RHESSI) to derive the temperature, emission measure and differential emission measure of the flaring regions and investigate the evolution of the plasma in the flaring ribbons. The ribbons were visible at all SDO/AIA EUV/UV wavelengths, in particular, at 94 and 131 Å filters, sensitive to temperatures of 8 MK and 12 MK. Time evolution of the emission measure of the plasma above 10 MK at the ribbons has a peak near the HXR peak time. The presence of hot plasma in the lower atmosphere is further confirmed by RHESSI imaging spectroscopy analysis, which shows resolved sources at 11-13 MK associated with at least one ribbon. We found that collisional beam heating can only marginally explain the necessary power to heat the 10 MK plasma at the ribbons. △ Less

Submitted 13 May, 2015; originally announced May 2015.

Comments: 21 pages, 15 figures

Direct observation of the energy release site in a solar flare by SDO/AIA, Hinode/EIS and RHESSI

Authors: Paulo J. A. Simões, David R. Graham, Lyndsay Fletcher

Abstract: We present direct evidence for the detection of the main energy release site in a non-eruptive solar flare, SOL2013-11-09T06:38UT. This GOES C2.7 event was characterised by two flaring ribbons and a compact, bright coronal source located between them, which is the focus of our study. We use imaging from SDO/AIA, and imaging spectroscopy from RHESSI to characterise the thermal and non-thermal emiss… ▽ More We present direct evidence for the detection of the main energy release site in a non-eruptive solar flare, SOL2013-11-09T06:38UT. This GOES C2.7 event was characterised by two flaring ribbons and a compact, bright coronal source located between them, which is the focus of our study. We use imaging from SDO/AIA, and imaging spectroscopy from RHESSI to characterise the thermal and non-thermal emission from the coronal source, and EUV spectroscopy from the Hinode/EIS, which scanned the coronal source during the impulsive peak, to analyse Doppler shifts in Fe XII and Fe XXIV emission lines, and determine the source density. The coronal source exhibited an impulsive emission lightcurve in all AIA filters during the impulsive phase. RHESSI hard X-ray images indicate both thermal and non-thermal emission at the coronal source, and its plasma temperature derived from RHESSI imaging spectroscopy shows an impulsive rise, reaching a maximum at 12-13 MK about 10 seconds prior to the hard X-ray peak. High redshifts associated with this bright source indicate downflows of 40-250 km/s at a broad range of temperatures, interpreted as loop shrinkage and/or outflows along the magnetic field. Outflows from the coronal source towards each ribbon are also observed by AIA images at 171, 193, 211, 304 and 1600 A. The electron density of the source obtained from a Fe XIV line pair is $10^{11.50}$ which is collisionally thick to electrons with energy up to 45-65 keV, responsible for the source's non-thermal X-ray emission. We conclude that the bright coronal source is the location of the main release of magnetic energy in this flare, with a geometry consistent with component reconnection between crossing, current-carrying loops. We argue that the energy that can be released via reconnection, based on observational estimates, can plausibly account for the non-thermal energetics of the flare. △ Less

Submitted 4 March, 2015; originally announced March 2015.

Comments: 10 pages, 7 figures

Journal ref: A&A 577, A68 (2015)

Determining Energy Balance in the Flaring Chromosphere from Oxygen V Line Ratios

Authors: David R. Graham, Lyndsay Fletcher, Nicolas Labrosse

Abstract: The impulsive phase of solar flares is a time of rapid energy deposition and heating in the lower solar atmosphere, leading to changes in the temperature and density structure of the region. We use an O V density diagnostic formed of the 192 to 248 line ratio, provided by Hinode EIS, to determine the density of flare footpoint plasma, at O V formation temperatures of 250,000 K, giving a constraint… ▽ More The impulsive phase of solar flares is a time of rapid energy deposition and heating in the lower solar atmosphere, leading to changes in the temperature and density structure of the region. We use an O V density diagnostic formed of the 192 to 248 line ratio, provided by Hinode EIS, to determine the density of flare footpoint plasma, at O V formation temperatures of 250,000 K, giving a constraint on the properties of the heated transition region. Hinode EIS rasters from 2 small flare events in December 2007 were used. Raster images were co-aligned to identify and establish the footpoint pixels, multiple-component Gaussian line fitting of the spectra was carried out to isolate the diagnostic pair, and the density was calculated for several footpoint areas. The assumptions of equilibrium ionization and optically thin radiation for the O V lines were found to be acceptable. Properties of the electron distribution, for one event, were deduced from earlier RHESSI hard X-ray observations and used to calculate the plasma heating rate, delivered by an electron beam adopting collisional thick-target assumptions, for 2 model atmospheres. Electron number densities of at least log n = 12.3 cm-3 were measured during the flare impulsive phase, far higher than previously expected. For one footpoint, the radiative loss rate for this plasma was found to exceed that which can be delivered by an electron beam implied by the RHESSI data. However, when assuming a completely ionised target atmosphere the heating rate exceeded the losses. A chromospheric thickness of 70-700 km was found to be required to balance a conductive input to the O V-emitting region with radiative losses. The analysis shows that for heating by collisional electrons, it is difficult, or impossible to raise the temperature of the chromosphere to explain the observed densities without assuming a completely ionised atmosphere. △ Less

Submitted 14 September, 2015; v1 submitted 17 November, 2014; originally announced November 2014.

Comments: Accepted to A&A 14th September 2015

Journal ref: A&A 584, A6 (2015)

The emission measure distribution of impulsive phase flare footpoints

Authors: D. R. Graham, I. G. Hannah, L. Fletcher, R. O. Milligan

Abstract: The temperature distribution of the emitting plasma is a crucial constraint when studying the heating of solar flare footpoints. However, determining this for impulsive phase footpoints has been difficult in the past due to insufficient spatial resolution to resolve the footpoints from the loop structures, and a lack of spectral and temporal coverage. We use the capabilities of Hinode/EIS to obtai… ▽ More The temperature distribution of the emitting plasma is a crucial constraint when studying the heating of solar flare footpoints. However, determining this for impulsive phase footpoints has been difficult in the past due to insufficient spatial resolution to resolve the footpoints from the loop structures, and a lack of spectral and temporal coverage. We use the capabilities of Hinode/EIS to obtain the first emission measure distributions (EMDs) from impulsive phase footpoints in six flares. Observations with good spectral coverage were analysed using a regularized inversion method to recover the EMDs. We find that the EMDs all share a peak temperature of around 8 MK, with lines formed around this temperature having emission measures peaking between 10^28 and 10^29 cm^-5, indicating a substantial presence of plasma at very high temperatures within the footpoints. An EMD gradient of EM(T) ~ T is found in all events. Previous theoretical work on emission measure gradients shows this to be consistent with a scenario in which the deposited flare energy directly heats only the top layer of the flare chromosphere, while deeper layers are heated by conduction. △ Less

Submitted 11 February, 2013; originally announced February 2013.