-
Inconsistency Masks: Removing the Uncertainty from Input-Pseudo-Label Pairs
Authors:
Michael R. H. Vorndran,
Bernhard F. Roeck
Abstract:
Efficiently generating sufficient labeled data remains a major bottleneck in deep learning, particularly for image segmentation tasks where labeling requires significant time and effort. This study tackles this issue in a resource-constrained environment, devoid of extensive datasets or pre-existing models. We introduce Inconsistency Masks (IM), a novel approach that filters uncertainty in image-p…
▽ More
Efficiently generating sufficient labeled data remains a major bottleneck in deep learning, particularly for image segmentation tasks where labeling requires significant time and effort. This study tackles this issue in a resource-constrained environment, devoid of extensive datasets or pre-existing models. We introduce Inconsistency Masks (IM), a novel approach that filters uncertainty in image-pseudo-label pairs to substantially enhance segmentation quality, surpassing traditional semi-supervised learning techniques. Employing IM, we achieve strong segmentation results with as little as 10% labeled data, across four diverse datasets and it further benefits from integration with other techniques, indicating broad applicability. Notably on the ISIC 2018 dataset, three of our hybrid approaches even outperform models trained on the fully labeled dataset. We also present a detailed comparative analysis of prevalent semi-supervised learning strategies, all under uniform starting conditions, to underline our approach's effectiveness and robustness. The full code is available at: https://github.com/MichaelVorndran/InconsistencyMasks
△ Less
Submitted 13 April, 2024; v1 submitted 25 January, 2024;
originally announced January 2024.
-
Generating the Cabibbo Angle in Flavored Gauge Mediation Models with Discrete Non-Abelian Symmetries
Authors:
Lisa L. Everett,
Todd S. Garon,
Ariel B. Rock
Abstract:
We explore the generation of fermion masses and quark mixing angles within flavored gauge mediation models of supersymmetry breaking in which the Higgs and messenger doublets are connected by a discrete non-Abelian symmetry. In this framework, the Higgs-messenger symmetry, which we take for concreteness to be the discrete group $\mathcal{S}_3$, also plays the role of a (partial) family symmetry. W…
▽ More
We explore the generation of fermion masses and quark mixing angles within flavored gauge mediation models of supersymmetry breaking in which the Higgs and messenger doublets are connected by a discrete non-Abelian symmetry. In this framework, the Higgs-messenger symmetry, which we take for concreteness to be the discrete group $\mathcal{S}_3$, also plays the role of a (partial) family symmetry. We investigate mechanisms for generating mass hierarchies for the lighter quark generations and generating the Cabibbo angle within this class of models. While perturbations of the couplings at the renormalizable level do not lead to phenomenologically viable quark mixing parameters, we show that the Cabibbo angle can be generated via couplings at the nonrenormalizable level and explore the phenomenological implications of this scenario.
△ Less
Submitted 30 December, 2019;
originally announced December 2019.
-
Predictions for the Leptonic Dirac CP-Violating Phase
Authors:
Lisa L. Everett,
Raymundo Ramos,
Ariel B. Rock,
Alexander J. Stuart
Abstract:
We explore the theoretical constraints on the observable parameters of neutrino mixing on predictions for the leptonic Dirac CP-violating phase within a well-studied class of simple theoretical models that includes a single source of CP violation due to charged lepton corrections. The approach guarantees that a physically meaningful prediction for the most likely values for the leptonic Dirac CP-v…
▽ More
We explore the theoretical constraints on the observable parameters of neutrino mixing on predictions for the leptonic Dirac CP-violating phase within a well-studied class of simple theoretical models that includes a single source of CP violation due to charged lepton corrections. The approach guarantees that a physically meaningful prediction for the most likely values for the leptonic Dirac CP-violating phase is obtained.
△ Less
Submitted 5 October, 2022; v1 submitted 20 December, 2019;
originally announced December 2019.
-
Sizable Stop Mixing in Flavored Gauge Mediation Models with Discrete Non-Abelian Symmetries
Authors:
Lisa L. Everett,
Todd S. Garon,
Ariel B. Rock
Abstract:
We analyze a minimal flavored gauge mediation model in which the electroweak Higgs and messenger doublets are embedded in multiplets of a discrete non-Abelian symmetry. In this scenario, the minimal Higgs-messenger sector that is consistent with the 125 GeV Higgs mass has two vectorlike pairs of messenger fields. This scenario is obtained in a specific limit of the superpotential interactions of t…
▽ More
We analyze a minimal flavored gauge mediation model in which the electroweak Higgs and messenger doublets are embedded in multiplets of a discrete non-Abelian symmetry. In this scenario, the minimal Higgs-messenger sector that is consistent with the 125 GeV Higgs mass has two vectorlike pairs of messenger fields. This scenario is obtained in a specific limit of the superpotential interactions of the Higgs-messenger fields and the matter fields. Due to the structure of the messenger-matter Yukawa couplings in this limit, sizable stop mixing and flavor-diagonal soft supersymmetry breaking parameters are achieved. In most of the parameter space, the masses of the colored superpartners are at most in the 5-6 TeV range.
△ Less
Submitted 27 December, 2018;
originally announced December 2018.
-
The puzzling interpretation of NIR indices: The case of NaI2.21
Authors:
B. Röck,
A. Vazdekis,
F. La Barbera,
R. F. Peletier,
J. H. Knapen,
C. Allende-Prieto,
D. S. Aguado
Abstract:
We present a detailed study of the Na I line strength index centered in the $K$-band at $22100$, Å (NaI2.21 hereafter) relying on different samples of early-type galaxies. Consistent with previous studies, we find that the observed line strength indices cannot be fit by state-of-art scaled-solar stellar population models, even using our newly developed models in the NIR. The models clearly underes…
▽ More
We present a detailed study of the Na I line strength index centered in the $K$-band at $22100$, Å (NaI2.21 hereafter) relying on different samples of early-type galaxies. Consistent with previous studies, we find that the observed line strength indices cannot be fit by state-of-art scaled-solar stellar population models, even using our newly developed models in the NIR. The models clearly underestimate the large NaI2.21 values measured for most early-type galaxies. However, we develop a Na-enhanced version of our newly developed models in the NIR, which - together with the effect of a bottom-heavy initial mass function - yield NaI2.21 indices in the range of the observations. Therefore, we suggest a scenario in which the combined effect of [Na/Fe] enhancement and a bottom-heavy initial mass function are mainly responsible for the large NaI2.21 indices observed for most early-type galaxies. To a smaller extent, also [C/Fe] enhancement might contribute to the large observed NaI2.21 values.
△ Less
Submitted 15 August, 2017;
originally announced August 2017.
-
UV-extended E-MILES stellar population models: young components in massive early-type galaxies
Authors:
A. Vazdekis,
M. Koleva,
E. Ricciardelli,
B. Röck,
J. Falcón-Barroso
Abstract:
We present UV-extended E-MILES stellar population synthesis models covering the spectral range 1680-50000A at moderately high resolution. We employ the NGSL space-based stellar library to compute spectra of single-age, single-metallicity stellar populations in the wavelength range from 1680 to 3540A. These models represent a significant improvement in resolution and age/metallicity coverage over p…
▽ More
We present UV-extended E-MILES stellar population synthesis models covering the spectral range 1680-50000A at moderately high resolution. We employ the NGSL space-based stellar library to compute spectra of single-age, single-metallicity stellar populations in the wavelength range from 1680 to 3540A. These models represent a significant improvement in resolution and age/metallicity coverage over previous studies based on earlier space-based libraries. These model spectra were joined with those we computed in the visible using MILES, and other empirical libraries for redder wavelengths. The models span the metallicity range -1.79<[M/H]<+0.26 and ages above 30 Myr, for a suite of IMF types with varying slopes. We focus on the behaviour of colours, spectra and line-strength indices in the UV range as a function of relevant stellar population parameters. Whereas some indices strengthen with increasing age and metallicity, as most metallicity indicators in the visible, other indices peak around 3 Gyr for metal-rich stellar populations, such as Mg at 2800A. Our models provide reasonably good fits to the integrated colours and most line-strengths of the stellar clusters of the Milky-Way and LMC. Our full-spectrum fits in the UV range for a representative set of ETGs of varying mass yield age and metallicity estimates in very good agreement with those obtained in the optical range. The comparison of UV colours and line-strengths of massive ETGs with our models reveals the presence of young stellar components, with ages in the range 0.1-0.5 Gyr and mass fractions 0.1-0.5%, on the top of an old stellar population.
△ Less
Submitted 4 December, 2016;
originally announced December 2016.
-
IMF and [Na/Fe] abundance ratios from optical and NIR Spectral Features in Early-type Galaxies
Authors:
F. La Barbera,
A. Vazdekis,
I. Ferreras,
A. Pasquali,
C. Allende Prieto,
B. Rock,
D. S. Aguado,
R. F. Peletier
Abstract:
We present a joint analysis of the four most prominent sodium-sensitive features (NaD, NaI8190, NaI1.14, and NaI2.21), in the optical and Near-Infrared spectral range, of two nearby, massive (sigma~300km/s), early-type galaxies (named XSG1 and XSG2). Our analysis relies on deep VLT/X-Shooter long-slit spectra, along with newly developed stellar population models, allowing for [Na/Fe] variations, u…
▽ More
We present a joint analysis of the four most prominent sodium-sensitive features (NaD, NaI8190, NaI1.14, and NaI2.21), in the optical and Near-Infrared spectral range, of two nearby, massive (sigma~300km/s), early-type galaxies (named XSG1 and XSG2). Our analysis relies on deep VLT/X-Shooter long-slit spectra, along with newly developed stellar population models, allowing for [Na/Fe] variations, up to 1.2dex, over a wide range of age, total metallicity, and IMF slope. The new models show that the response of the Na-dependent spectral indices to [Na/Fe] is stronger when the IMF is bottom heavier. For the first time, we are able to match all four Na features in the central regions of massive early-type galaxies, finding an overabundance of [Na/Fe], in the range 0.5-0.7dex, and a bottom-heavy IMF. Therefore, individual abundance variations cannot be fully responsible for the trends of gravity-sensitive indices, strengthening the case towards a non-universal IMF. Given current limitations of theoretical atmosphere models, our [Na/Fe] estimates should be taken as upper limits. For XSG1, where line strengths are measured out to 0.8Re, the radial trend of [Na/Fe] is similar to [Mg/Fe] and [C/Fe], being constant out to 0.5Re, and decreasing by 0.2-0.3dex at 0.8Re, without any clear correlation with local metallicity. Such a result seems to be in contrast with the predicted increase of Na nucleosynthetic yields from AGB stars and TypeII SNe. For XSG1, the Na-inferred IMF radial profile is consistent, within the errors, with that derived from TiO features and the Wing-Ford band, presented in a recent paper.
△ Less
Submitted 16 October, 2016; v1 submitted 12 October, 2016;
originally announced October 2016.
-
MILES extended: Stellar population synthesis models from the optical to the infrared
Authors:
B. Röck,
A. Vazdekis,
E. Ricciardelli,
R. F. Peletier,
J. H. Knapen,
J. Falcon-Barroso
Abstract:
We present the first single-burst stellar population models which covers the optical and the infrared wavelength range between 3500 and 50000 Angstrom and which are exclusively based on empirical stellar spectra. To obtain these joint models, we combined the extended MILES models in the optical with our new infrared models that are based on the IRTF (Infrared Telescope Facility) library. The latte…
▽ More
We present the first single-burst stellar population models which covers the optical and the infrared wavelength range between 3500 and 50000 Angstrom and which are exclusively based on empirical stellar spectra. To obtain these joint models, we combined the extended MILES models in the optical with our new infrared models that are based on the IRTF (Infrared Telescope Facility) library. The latter are available only for a limited range in terms of both age and metallicity. Our combined single-burst stellar population models were calculated for ages larger than 1 Gyr, for metallicities between [Fe/H] = -0.40 and 0.26, for initial mass functions of various types and slopes, and on the basis of two different sets of isochrones. They are available to the scientific community on the MILES web page. We checked the internal consistency of our models and compared their colour predictions to those of other models that are available in the literature. Optical and near infrared colours that are measured from our models are found to reproduce the colours well that were observed for various samples of early-type galaxies. Our models will enable a detailed analysis of the stellar populations of observed galaxies.
△ Less
Submitted 18 September, 2016;
originally announced September 2016.
-
Stellar population synthesis models between 2.5 and 5 μm based on the empirical IRTF stellar library
Authors:
B. Röck,
A. Vazdekis,
R. F. Peletier,
J. H. Knapen,
J. Falcón-Barroso
Abstract:
We present the first single-burst stellar population models in the infrared wavelength range between 2.5 and 5 μm which are exclusively based on empirical stellar spectra. Our models take as input 180 spectra from the stellar IRTF (Infrared Telescope Facility) library. Our final single-burst stellar population models are calculated based on two different sets of isochrones and various types of ini…
▽ More
We present the first single-burst stellar population models in the infrared wavelength range between 2.5 and 5 μm which are exclusively based on empirical stellar spectra. Our models take as input 180 spectra from the stellar IRTF (Infrared Telescope Facility) library. Our final single-burst stellar population models are calculated based on two different sets of isochrones and various types of initial mass functions of different slopes, ages larger than 1 Gyr and metallicities between [Fe/H] = -0.70 and 0.26. They are made available online to the scientific community on the MILES web page. We analyse the behaviour of the Spitzer [3.6]-[4.5] colour calculated from our single stellar population models and find only slight dependences on both metallicity and age. When comparing to the colours of observed early-type galaxies, we find a good agreement for older, more massive galaxies that resemble a single-burst population. Younger, less massive and more metal-poor galaxies show redder colours with respect to our models. This mismatch can be explained by a more extended star formation history of these galaxies which includes a metal-poor or/and young population. Moreover, the colours derived from our models agree very well with most other models available in this wavelength range. We confirm that the mass-to-light ratio determined in the Spitzer [3.6] μm band changes much less as a function of both age and metallicity than in the optical bands.
△ Less
Submitted 7 May, 2015;
originally announced May 2015.
-
The GALEX/S4G UV-IR color-color diagram: Catching spiral galaxies away from the Blue Sequence
Authors:
Alexandre Y. K. Bouquin,
Armando Gil de Paz,
Samuel Boissier,
Juan-Carlos Muñoz-Mateos,
Kartik Sheth,
Dennis Zaritsky,
Jarkko Laine,
Jesús Gallego,
Reynier F. Peletier,
Benjamin R. Röck,
Johan H. Knapen
Abstract:
We obtained GALEX FUV, NUV, and Spitzer/IRAC 3.6$μ$m photometry for > 2000 galaxies, available for 90% of the S4G sample. We find a very tight "GALEX Blue Sequence (GBS)" in the (FUV-NUV) versus (NUV-[3.6]) color-color diagram which is populated by irregular and spiral galaxies, and is mainly driven by changes in the formation timescale ($τ$) and a degeneracy between $τ$ and dust reddening. The ti…
▽ More
We obtained GALEX FUV, NUV, and Spitzer/IRAC 3.6$μ$m photometry for > 2000 galaxies, available for 90% of the S4G sample. We find a very tight "GALEX Blue Sequence (GBS)" in the (FUV-NUV) versus (NUV-[3.6]) color-color diagram which is populated by irregular and spiral galaxies, and is mainly driven by changes in the formation timescale ($τ$) and a degeneracy between $τ$ and dust reddening. The tightness of the GBS provides an unprecedented way of identifying star-forming galaxies and objects that are just evolving to (or from) what we call the "GALEX Green Valley (GGV)". At the red end of the GBS, at (NUV-[3.6]) > 5, we find a wider "GALEX Red Sequence (GRS)" mostly populated by E/S0 galaxies that has a perpendicular slope to that of the GBS and of the optical red sequence. We find no such dichotomy in terms of stellar mass (measured by $\rm{M}_{[3.6]}$), since both massive ($M_{\star} > 10^{11} M_{\odot}$) blue and red sequence galaxies are identified. The type that is proportionally more often found in the GGV are the S0-Sa's and most of these are located in high-density environments. We discuss evolutionary models of galaxies that show a rapid transition from the blue to the red sequence on timescale of $10^{8}$years.
△ Less
Submitted 26 January, 2015;
originally announced January 2015.
-
SDC13 infrared dark clouds: Longitudinally collapsing filaments?
Authors:
N. Peretto,
G. A. Fuller,
Ph. André,
D. Arzoumanian,
V. M. Rivilla,
S. Bardeau,
S. Duarte Puertas,
J. P. Guzman Fernandez,
C. Lenfestey,
G. -X. Li,
F. A. Olguin,
B. R. Röck,
H. de Villiers,
J. Williams
Abstract:
Formation of stars is now believed to be tightly linked to the dynamical evolution of interstellar filaments in which they form. In this paper we analyze the density structure and kinematics of a small network of infrared dark filaments, SDC13, observed in both dust continuum and molecular line emission with the IRAM 30m telescope. These observations reveal the presence of 18 compact sources among…
▽ More
Formation of stars is now believed to be tightly linked to the dynamical evolution of interstellar filaments in which they form. In this paper we analyze the density structure and kinematics of a small network of infrared dark filaments, SDC13, observed in both dust continuum and molecular line emission with the IRAM 30m telescope. These observations reveal the presence of 18 compact sources amongst which the two most massive, MM1 and MM2, are located at the intersection point of the parsec-long filaments. The dense gas velocity and velocity dispersion observed along these filaments show smooth, strongly correlated, gradients. We discuss the origin of the SDC13 velocity field in the context of filament longitudinal collapse. We show that the collapse timescale of the SDC13 filaments (from 1 Myr to 4 Myr depending on the model parameters) is consistent with the presence of Class I sources in them, and argue that, on top of bringing more material to the centre of the system, collapse could generate additional kinematic support against local fragmentation, hel** the formation of starless super-Jeans cores.
△ Less
Submitted 1 November, 2013;
originally announced November 2013.
-
Possible observation of phase separation near a quantum phase transition in doubly connected ultrathin superconducting cylinders of aluminum
Authors:
H. Wang,
M. M. Rosario,
N. A. Kurz,
B. Y. Rock,
M. Tian,
P. T. Carrigan,
Y. Liu
Abstract:
The kinetic energy of superconducting electrons in an ultrathin, doubly connected superconducting cylinder, determined by the applied flux, increases as the cylinder diameter decreases, leading to a destructive regime around half-flux quanta and a superconductor to normal metal quantum phase transition (QPT). Regular step-like features in resistance vs. temperature curves taken at fixed flux val…
▽ More
The kinetic energy of superconducting electrons in an ultrathin, doubly connected superconducting cylinder, determined by the applied flux, increases as the cylinder diameter decreases, leading to a destructive regime around half-flux quanta and a superconductor to normal metal quantum phase transition (QPT). Regular step-like features in resistance vs. temperature curves taken at fixed flux values were observed near the QPT in ultrathin Al cylinders. It is proposed that these features are most likely resulted from a phase separation near the QPT in which normal regions nucleate in a homogeneous superconducting cylinder.
△ Less
Submitted 5 October, 2005;
originally announced October 2005.
-
Geometrical destruction of the global phase coherence in ultrathin superconducting cylinders
Authors:
Y. Liu,
Yu. Zadorozhny,
M. M. Rosario,
B. Y. Rock,
P. T. Carrigan,
H. Wang
Abstract:
The global phase coherence in doubly-connected superconductors leads to fluxoid quantization, allowing the superfluid velocity vs to be controlled by an applied magnetic flux. In ultrasmall samples this quantization requirement leads, surprisingly, to the destruction of the phase coherence itself around half-integer flux quanta, because of the sample-size-induced growth in vs, as predicted by de…
▽ More
The global phase coherence in doubly-connected superconductors leads to fluxoid quantization, allowing the superfluid velocity vs to be controlled by an applied magnetic flux. In ultrasmall samples this quantization requirement leads, surprisingly, to the destruction of the phase coherence itself around half-integer flux quanta, because of the sample-size-induced growth in vs, as predicted by de Gennes. We report observations of the predicted phenomenon in ultrathin Al and Au0.7In0.3 cylinders, and the corresponding phase diagram for ultrathin superconducting cylinders. The new phase diagram features disconnected superconducting regions, as opposed to the single one seen in the conventional Little-Parks experiment.
△ Less
Submitted 10 October, 2001;
originally announced October 2001.
