arXiv subjects

Michelle Ntampaka

Publications and source records attributed to Michelle Ntampaka.

31 records · Page 2Linked to original sources

arXiv2021

The Importance of Being Interpretable: Toward An Understandable Machine Learning Encoder for Galaxy Cluster Cosmology

We present a deep machine learning (ML) approach to constraining cosmological parameters with multi-wavelength observations of galaxy clusters. The ML approach has two components: an encoder that builds a compressed representation of each galaxy cluster and a flexible CNN to estimate the cosmological model from a cluster sample. It is trained and tested on simulated cluster catalogs built from the Magneticum simulations. From the simulated catalogs, the ML method estimates the amplitude of matter fluctuations, sigma_8, at approximately the expected theoretical limit. More importantly, the deep ML approach can be interpreted. We lay out three schemes for interpreting the ML technique: a leave-one-out method for assessing cluster importance, an average saliency for evaluating feature importance, and correlations in the terse layer for understanding whether an ML technique can be safely applied to observational data. These interpretation schemes led to the discovery of a previously unknown self-calibration mode for flux- and volume-limited cluster surveys. We describe this new mode, which uses the amplitude and peak of the cluster mass PDF as anchors for mass calibration. We introduce the term "overspecialized" to describe a common pitfall in astronomical applications of machine learning in which the ML method learns simulation-specific details, and we show how a carefully constructed architecture can be used to check for this source of systematic error.

astro-ph.IM↗

arXiv2021

Building Trustworthy Machine Learning Models for Astronomy

Astronomy is entering an era of data-driven discovery, due in part to modern machine learning (ML) techniques enabling powerful new ways to interpret observations. This shift in our scientific approach requires us to consider whether we can trust the black box. Here, we overview methods for an often-overlooked step in the development of ML models: building community trust in the algorithms. Trust is an essential ingredient not just for creating more robust data analysis techniques, but also for building confidence within the astronomy community to embrace machine learning methods and results.

astro-ph.IM↗

arXiv2021

The Role of Machine Learning in the Next Decade of Cosmology

In recent years, machine learning (ML) methods have remarkably improved how cosmologists can interpret data. The next decade will bring new opportunities for data-driven cosmological discovery, but will also present new challenges for adopting ML methodologies and understanding the results. ML could transform our field, but this transformation will require the astronomy community to both foster and promote interdisciplinary research endeavors.

astro-ph.IM↗

arXiv2021

A Machine Learning Approach for Dynamical Mass Measurements of Galaxy Clusters

We present a modern machine learning approach for cluster dynamical mass measurements that is a factor of two improvement over using a conventional scaling relation. Different methods are tested against a mock cluster catalog constructed using halos with mass >= 10^14 Msolar/h from Multidark's publicly-available N-body MDPL halo catalog. In the conventional method, we use a standard M(sigma_v) power law scaling relation to infer cluster mass, M, from line-of-sight (LOS) galaxy velocity dispersion, sigma_v. The resulting fractional mass error distribution is broad, with width=0.87 (68% scatter), and has extended high-error tails. The standard scaling relation can be simply enhanced by including higher-order moments of the LOS velocity distribution. Applying the kurtosis as a correction term to log(sigma_v) reduces the width of the error distribution to 0.74 (16% improvement). Machine learning can be used to take full advantage of all the information in the velocity distribution. We employ the Support Distribution Machines (SDMs) algorithm that learns from distributions of data to predict single values. SDMs trained and tested on the distribution of LOS velocities yield width=0.46 (47% improvement). Furthermore, the problematic tails of the mass error distribution are effectively eliminated. Decreasing cluster mass errors will improve measurements of the growth of structure and lead to tighter constraints on cosmological parameters.

astro-ph.CO↗

arXiv2020

A Robust and Efficient Deep Learning Method for Dynamical Mass Measurements of Galaxy Clusters

We demonstrate the ability of convolutional neural networks (CNNs) to mitigate systematics in the virial scaling relation and produce dynamical mass estimates of galaxy clusters with remarkably low bias and scatter. We present two models, CNN$_\mathrm{1D}$ and CNN$_\mathrm{2D}$, which leverage this deep learning tool to infer cluster masses from distributions of member galaxy dynamics. Our first model, CNN$_\text{1D}$, infers cluster mass directly from the distribution of member galaxy line-of-sight velocities. Our second model, CNN$_\text{2D}$, extends the input space of CNN$_\text{1D}$ to learn on the joint distribution of galaxy line-of-sight velocities and projected radial distances. We train each model as a regression over cluster mass using a labeled catalog of realistic mock cluster observations generated from the MultiDark simulation and UniverseMachine catalog. We then evaluate the performance of each model on an independent set of mock observations selected from the same simulated catalog. The CNN models produce cluster mass predictions with lognormal residuals of scatter as low as $0.132$ dex, greater than a factor of 2 improvement over the classical $M$-$σ$ power-law estimator. Furthermore, the CNN model reduces prediction scatter relative to similar machine learning approaches by up to $17\%$ while executing in drastically shorter training and evaluation times (by a factor of 30) and producing considerably more robust mass predictions (improving prediction stability under variations in galaxy sampling rate by $30\%$).

astro-ph.CO↗

arXiv2020

SuperRAENN: A Semi-supervised Supernova Photometric Classification Pipeline Trained on Pan-STARRS1 Medium Deep Survey Supernovae

Automated classification of supernovae (SNe) based on optical photometric light curve information is essential in the upcoming era of wide-field time domain surveys, such as the Legacy Survey of Space and Time (LSST) conducted by the Rubin Observatory. Photometric classification can enable real-time identification of interesting events for extended multi-wavelength follow-up, as well as archival population studies. Here we present the complete sample of 5,243 "SN-like" light curves (in griz) from the Pan-STARRS1 Medium-Deep Survey (PS1-MDS). The PS1-MDS is similar to the planned LSST Wide-Fast-Deep survey in terms of cadence, filters and depth, making this a useful training set for the community. Using this dataset, we train a novel semi-supervised machine learning algorithm to photometrically classify 2,315 new SN-like light curves with host galaxy spectroscopic redshifts. Our algorithm consists of a random forest supervised classification step and a novel unsupervised step in which we introduce a recurrent autoencoder neural network (RAENN). Our final pipeline, dubbed SuperRAENN, has an accuracy of 87% across five SN classes (Type Ia, Ibc, II, IIn, SLSN-I). We find the highest accuracy rates for Type Ia SNe and SLSNe and the lowest for Type Ibc SNe. Our complete spectroscopically- and photometrically-classified samples break down into: 62.0% Type Ia (1839 objects), 19.8% Type II (553 objects), 4.8% Type IIn (136 objects), 11.7% Type Ibc (291 objects), and 1.6% Type I SLSNe (54 objects). Finally, we discuss how this algorithm can be modified for online LSST data streams.

astro-ph.HE↗

arXiv2019

Algorithms and Statistical Models for Scientific Discovery in the Petabyte Era

The field of astronomy has arrived at a turning point in terms of size and complexity of both datasets and scientific collaboration. Commensurately, algorithms and statistical models have begun to adapt --- e.g., via the onset of artificial intelligence --- which itself presents new challenges and opportunities for growth. This white paper aims to offer guidance and ideas for how we can evolve our technical and collaborative frameworks to promote efficient algorithmic development and take advantage of opportunities for scientific discovery in the petabyte era. We discuss challenges for discovery in large and complex data sets; challenges and requirements for the next stage of development of statistical methodologies and algorithmic tool sets; how we might change our paradigms of collaboration and education; and the ethical implications of scientists' contributions to widely applicable algorithms and computational modeling. We start with six distinct recommendations that are supported by the commentary following them. This white paper is related to a larger corpus of effort that has taken place within and around the Petabytes to Science Workshops (https://petabytestoscience.github.io/).

astro-ph.IM↗

arXiv2019

Using X-Ray Morphological Parameters to Strengthen Galaxy Cluster Mass Estimates via Machine Learning

We present a machine learning approach for estimating galaxy cluster masses, trained using both Chandra and eROSITA mock X-ray observations of 2,041 clusters from the Magneticum simulations. We train a random forest regressor, an ensemble learning method based on decision tree regression, to predict cluster masses using an input feature set. The feature set uses core-excised X-ray luminosity and a variety of morphological parameters, including surface brightness concentration, smoothness, asymmetry, power ratios, and ellipticity. The regressor is cross-validated and calibrated on a training sample of 1,615 clusters (80% of sample), and then results are reported as applied to a test sample of 426 clusters (20% of sample). This procedure is performed for two different mock observation series in an effort to bracket the potential enhancement in mass predictions that can be made possible by including dynamical state information. The first series is computed from idealized Chandra-like mock cluster observations, with high spatial resolution, long exposure time (1 Ms), and the absence of background. The second series is computed from realistic-condition eROSITA mocks with lower spatial resolution, short exposures (2 ks), instrument effects, and background photons modeled. We report a 20% reduction in the mass estimation scatter when either series is used in our random forest model compared to a standard regression model that only employs core-excised luminosity. The morphological parameters that hold the highest feature importance are smoothness, asymmetry, and surface brightness concentration. Hence, these parameters, which encode the dynamical state of the cluster, can be used to make more accurate predictions of cluster masses in upcoming surveys, offering a crucial step forward for cosmological analyses.

astro-ph.CO↗

arXiv2019

A Hybrid Deep Learning Approach to Cosmological Constraints From Galaxy Redshift Surveys

We present a deep machine learning (ML)-based technique for accurately determining $σ_8$ and $Ω_m$ from mock 3D galaxy surveys. The mock surveys are built from the AbacusCosmos suite of $N$-body simulations, which comprises 40 cosmological volume simulations spanning a range of cosmological models, and we account for uncertainties in galaxy formation scenarios through the use of generalized halo occupation distributions (HODs). We explore a trio of ML models: a 3D convolutional neural network (CNN), a power-spectrum-based fully connected network, and a hybrid approach that merges the two to combine physically motivated summary statistics with flexible CNNs. We describe best practices for training a deep model on a suite of matched-phase simulations and we test our model on a completely independent sample that uses previously unseen initial conditions, cosmological parameters, and HOD parameters. Despite the fact that the mock observations are quite small ($\sim0.07h^{-3}\,\mathrm{Gpc}^3$) and the training data span a large parameter space (6 cosmological and 6 HOD parameters), the CNN and hybrid CNN can constrain $σ_8$ and $Ω_m$ to $\sim3\%$ and $\sim4\%$, respectively.

astro-ph.CO↗

arXiv2019

Machine Learning Applied to the Reionization History of the Universe in the 21 cm Signal

The Epoch of Reionization (EoR) features a rich interplay between the first luminous sources and the low-density gas of the intergalactic medium (IGM), where photons from these sources ionize the IGM. There are currently few observational constraints on key observables related to the EoR, such as the midpoint and duration of reionization. Although upcoming observations of the 21 cm power spectrum with next-generation radio interferometers such as the Hydrogen Epoch of Reionization Array (HERA) and the Square Kilometre Array (SKA) are expected to provide information about the midpoint of reionization readily, extracting the duration from the power spectrum alone is a more difficult proposition. As an alternative method for extracting information about reionization, we present an application of convolutional neural networks (CNNs) to images of reionization. These images are two-dimensional in the plane of the sky, and extracted at a series of redshift values to generate "image cubes" that are qualitatively similar to those of the HERA and the SKA will generate in the near future. Additionally, we include the impact that the bright foreground signal from the the Milky Way imparts on such image cubes from interferometers, but do not include the noise induced from observations. We show that we are able to recover the duration of reionization $Δ$z to within 5% using CNNs, assuming that the midpoint of reionization is already relatively well constrained. These results have exciting impacts for estimating $τ$, the optical depth to the cosmic microwave background, which can help constrain other cosmological parameters.

astro-ph.CO↗

arXiv2019

Astro2020 APC White Paper: The Early Career Perspective on the Coming Decade, Astrophysics Career Paths, and the Decadal Survey Process

In response to the need for the Astro2020 Decadal Survey to explicitly engage early career astronomers, the National Academies of Sciences, Engineering, and Medicine hosted the Early Career Astronomer and Astrophysicist Focus Session (ECFS) on October 8-9, 2018 under the auspices of Committee of Astronomy and Astrophysics. The meeting was attended by fifty six pre-tenure faculty, research scientists, postdoctoral scholars, and senior graduate students, as well as eight former decadal survey committee members, who acted as facilitators. The event was designed to educate early career astronomers about the decadal survey process, to solicit their feedback on the role that early career astronomers should play in Astro2020, and to provide a forum for the discussion of a wide range of topics regarding the astrophysics career path. This white paper presents highlights and themes that emerged during two days of discussion. In Section 1, we discuss concerns that emerged regarding the coming decade and the astrophysics career path, as well as specific recommendations from participants regarding how to address them. We have organized these concerns and suggestions into five broad themes. These include (sequentially): (1) adequately training astronomers in the statistical and computational techniques necessary in an era of "big data", (2) responses to the growth of collaborations and telescopes, (3) concerns about the adequacy of graduate and postdoctoral training, (4) the need for improvements in equity and inclusion in astronomy, and (5) smoothing and facilitating transitions between early career stages. Section 2 is focused on ideas regarding the decadal survey itself, including: incorporating early career voices, ensuring diverse input from a variety of stakeholders, and successfully and broadly disseminating the results of the survey.

astro-ph.IM↗

arXiv2019

Cluster Cosmology with the Velocity Distribution Function of the HeCS-SZ Sample

We apply the Velocity Distribution Function (VDF) to a sample of Sunyaev-Zel'dovich (SZ)-selected clusters, and we report preliminary cosmological constraints in the $σ_8$-$Ω_m$ cosmological parameter space. The VDF is a forward-modeled test statistic that can be used to constrain cosmological models directly from galaxy cluster dynamical observations. The method was introduced in Ntampaka et al. (2017) and employs line-of-sight velocity measurements to directly constrain cosmological parameters; it is less sensitive to measurement error than a standard halo mass function approach. The method is applied to the Hectospec Survey of Sunyaev-Zeldovich-Selected Clusters (HeCS-SZ) sample, which is a spectroscopic follow up of a Planck-selected sample of 83 galaxy clusters. Credible regions are calculated by comparing the VDF of the observed cluster sample to that of mock observations, yielding $\mathcal{S}_8 \equiv σ_8 \left(Ω_m/0.3\right)^{0.25} = 0.751\pm0.037$. These constraints are in tension with the Planck Cosmic Microwave Background (CMB) TT fiducial value, which lies outside of our 95% credible region, but are in agreement with some recent analyses of large scale structure that observe fewer massive clusters than are predicted by the Planck fiducial cosmological parameters.

astro-ph.CO↗

arXiv2013

A First Look at creating mock catalogs with machine learning techniques

We investigate machine learning (ML) techniques for predicting the number of galaxies (N_gal) that occupy a halo, given the halo's properties. These types of mappings are crucial for constructing the mock galaxy catalogs necessary for analyses of large-scale structure. The ML techniques proposed here distinguish themselves from traditional halo occupation distribution (HOD) modeling as they do not assume a prescribed relationship between halo properties and N_gal. In addition, our ML approaches are only dependent on parent halo properties (like HOD methods), which are advantageous over subhalo-based approaches as identifying subhalos correctly is difficult. We test 2 algorithms: support vector machines (SVM) and k-nearest-neighbour (kNN) regression. We take galaxies and halos from the Millennium simulation and predict N_gal by training our algorithms on the following 6 halo properties: number of particles, M_200, σ_v, v_max, half-mass radius and spin. For Millennium, our predicted N_gal values have a mean-squared-error (MSE) of ~0.16 for both SVM and kNN. Our predictions match the overall distribution of halos reasonably well and the galaxy correlation function at large scales to ~5-10%. In addition, we demonstrate a feature selection algorithm to isolate the halo parameters that are most predictive, a useful technique for understanding the mapping between halo properties and N_gal. Lastly, we investigate these ML-based approaches in making mock catalogs for different galaxy subpopulations (e.g. blue, red, high M_star, low M_star). Given its non-parametric nature as well as its powerful predictive and feature selection capabilities, machine learning offers an interesting alternative for creating mock catalogs.

astro-ph.CO↗