Browsing by Author "Andrianomena, Sambatra"

Now showing 1 - 8 of 8
  • Thumbnail Image
    Item
    Classifying galaxies according to their H I content
    (Oxford University Press, 2020) Andrianomena, Sambatra; Rafieferantsoa, Mika; Dave, Romeel
    We use machine learning to classify galaxies according to their H I content, based on both their optical photometry and environmental properties. The data used for our analyses are the outputs in the range z = 0–1 from MUFASA cosmological hydrodynamic simulation. In our previous paper, where we predicted the galaxy H I content using the same input features, H I-rich galaxies were only selected for the training. In order for the predictions on real observation data to be more accurate, the classifiers built in this study will first establish if a galaxy is H I rich (⁠log(MHI/M∗)>−2⁠) before estimating its neutral hydrogen content using the regressors developed in the first paper.
  • Thumbnail Image
    Item
    Constraining the astrophysics and cosmology from 21 cm tomography using deep learning with the SKA
    (Oxford University Press, 2020) Hassan, Sultan; Andrianomena, Sambatra; Doughty, Caitlin
    Future Square Kilometre Array (SKA) surveys are expected to generate huge data sets of 21 cm maps on cosmological scales from the Epoch of Reionization. We assess the viability of exploiting machine learning techniques, namely, convolutional neural networks (CNNs), to simultaneously estimate the astrophysical and cosmological parameters from 21 cm maps from seminumerical simulations. We further convert the simulated 21 cm maps into SKA-like mock maps using the detailed SKA antennae distribution, thermal noise, and a recipe for foreground cleaning. We successfully design two CNN architectures (VGGNet-like and ResNet-like) that are both efficiently able to extract simultaneously three astrophysical parameters, namely the photon escape fraction (fesc), the ionizing emissivity power dependence on halo mass (Cion), and the ionizing emissivity redshift evolution index (Dion), and three cosmological parameters, namely the matter density parameter (Ωm), the dimensionless Hubble constant (h), and the matter fluctuation amplitude (σ8), from 21 cm maps at several redshifts.
  • Thumbnail Image
    Item
    Cosmological multifield emulator
    (Springer Science and Business Media B.V., 2025) Andrianomena, Sambatra; Hassan, Sultan; Villaescusa-Navarro, Francisco
    We present the application of deep networks to learn the distribution of multiple large-scale fields, conditioned exclusively on cosmology while marginalizing over astrophysics. Our approach develops a generalized multifield emulator based purely on theoretical predictions from the state-of-the-art hydrodynamic simulations of the CAMELS project, without incorporating instrumental effects which limit the analysis to specifics of a particular large-scale survey design. To this end, we train a generative adversarial network to generate images composed of three different channels that represent gas density (Mgas), neutral hydrogen density (HI), and magnetic field amplitudes (B). We consider an unconstrained model and another scenario where the model is conditioned on the matter density Ωm and the amplitude of density fluctuations σ8. We find that the generated images exhibit great quality which is on a par with that of data, visually. Quantitatively, we find that our model generates maps whose statistical properties, quantified by probability distribution function (PDF) of pixel values and auto-power spectra, agree reasonably well up to the second moment with those of the real maps. The relative deviation between the PDFs is about 25% in both moments with larger deviations at the tails. The error between the two auto-power spectra is approximately less than 20% on scales larger than k=10h/Mpc, but becomes larger on smaller scales. Moreover, the mean and standard deviation of the cross-correlations between fields in all maps produced by the emulator are in good agreement with those of the real images, which indicates that our model generates instances whose maps in all three channels describe the same physical region. Furthermore, a CNN regressor, which has been trained to extract Ωm and σ8 from CAMELS multifield dataset, recovers the cosmology from the maps generated by our conditional model, achieving coefficient of determination values R2=0.96 and 0.83 corresponding to Ωm and σ8 respectively. This further demonstrates the great capability of the model to mimic CAMELS data. Our model can be useful for generating data, 1000 multiple images in ∼3 seconds as opposed to a simulation which takes days for one realization, that are required to analyze the information from upcoming multi-wavelength cosmological surveys.
  • Thumbnail Image
    Item
    Predicting the neutral hydrogen content of galaxies from optical data using machine learning
    (Oxford University Press, 2018) Rafieferantsoa, Mika; Andrianomena, Sambatra; Dave, Romeel
    We develop a machine learning-based framework to predict the Hi content of galaxies using more straightforwardly observable quantities such as optical photometry and environmental parameters. We train the algorithm on z = 0 - 2 outputs from the Mufasa cosmological hydrodynamic simulation, which includes star formation, feedback, and a heuristic model to quench massive galaxies that yields a reasonable match to a range of survey data including Hi. We employ a variety of machine learning methods (regressors), and quantify their performance using the root mean square error (rmse) and the Pearson correlation coefficient (r). Considering SDSS photometry, 3rd nearest neighbor environment and line of sight peculiar velocities as features, we obtain r > 0:8 accuracy of the Hi-richness prediction, corresponding to rmse< 0:3. Adding near-IR photometry to the features yields some improvement to the prediction. Compared to all the regressors, random forest shows the best performance, with r > 0:9 at z = 0, followed by a Deep Neural Network with r > 0:85. All regressors exhibit a declining performance with increasing redshift, which limits the utility of this approach to z ~<1, and they tend to somewhat over-predict the Hi content of low-Hi galaxies which might be due to Eddington bias in the training sample.We test our approach on the RESOLVE survey data. Training on a subset of RESOLVE, we find that our machine learning method can reasonably well predict the Hi-richness of the remaining RESOLVE data, with rmse~ 0:28. Whenwe train on mock data fromMufasa and test onRESOLVE, this increases to rmse~ 0:45. Our method will be useful for making galaxy-by-galaxy survey predictions and incompleteness corrections for upcoming Hi 21cm surveys such as the LADUMA and MIGHTEE surveys on MeerKAT, over regions where photometry is already available.
  • Thumbnail Image
    Item
    Probabilistic cosmological inference on HI tomographic data
    (Springer Science and Business Media B.V., 2025) Andrianomena, Sambatra
    We explore the possibility of retrieving cosmological information along with its inherent uncertainty from 21-cm tomographic data at intermediate redshift. The first step in our approach consists of training an encoder, composed of several three dimensional convolutional layers, to cast the neutral hydrogen 3D data into a lower dimension latent space. Once pre-trained, the featurizer is able to generate 3D grid representations which, in turn, will be mapped onto cosmology (Ωm, σ8) via likelihood-free inference. For the latter, which is framed as a density estimation problem, we consider a Bayesian approximation method which exploits the capacity of Masked Autoregressive Flow to estimate the posterior. It is found that the representations learned by the deep encoder are separable in latent space. Results show that the neural density estimator, trained on the latent codes, is able to constrain cosmology with a precision of R2≥0.91 on all parameters and that most of the ground truth of the instances in the test set fall within 1σ uncertainty. It is established that the posterior uncertainty from the density estimator is reasonably calibrated. We also investigate the robustness of the feature extractor by using it to compress out-of-distribution dataset, that is either from a different simulation or from the same simulation but at different redshift. We find that, while trained on the latent codes corresponding to different types of out-of-distribution dataset, the probabilistic model is still reasonably capable of constraining cosmology, with R2≥0.80 in general. This highlights both the predictive power of the density estimator considered in this work and the meaningfulness of the latent codes retrieved by the encoder. We believe that the approach prescribed in this proof of concept will be of great use when analyzing 21-cm data from various surveys in the near future.
  • No Thumbnail Available
    Item
    Probabilistic learning for pulsar classification
    (IOP Publishing, 2022) Andrianomena, Sambatra
    In this work, we explore the possibility of using probabilistic learning to identify pulsar candidates. We make use of Deep Gaussian Process (DGP) and Deep Kernel Learning (DKL). Trained on a balanced training set in order to avoid the effect of class imbalance, the performance of the models, achieving relatively high probability of differentiating the positive class from the negative one (roc-auc ∼ 0.98), is very promising overall. We estimate the predictive entropy of each model predictions and find that DKL is more confident than DGP in its predictions and provides better uncertainty calibration. Upon investigating the effect of training with imbalanced dataset on the models, results show that each model performance decreases with an increasing number of the majority class in the training set. Interestingly, with a number of negative class 10× that of positive class, the models still provide reasonably well calibrated uncertainty, i.e. an expected Uncertainty Calibration Error (UCE) less than 6%.
  • Thumbnail Image
    Item
    Timing and noise analysis of five millisecond pulsars observed with MeerKAT
    (Oxford University Press, 2025) Andrianomena, Sambatra; Geyer, Marisa; Enwelum, U
    Millisecond pulsars (MSPs) in binary systems are precise laboratories for tests of gravity and the physics of dense matter. Their orbits can show relativistic effects that provide a measurement of the neutron star mass and the pulsars are included in timing array experiments that search for gravitational waves. Neutron star mass measurements are key to eventually solving the neutron star equation of state and these can be obtained by a measure of the Shapiro delay if the orbit is viewed near edge-on. Here, we report on the timing and noise analysis of five MSPs observed with the MeerKAT radio telescope: PSRs J0900–3144, J0921–5202, J1216–6410, J1327–0755, and J1543–5149. We searched for the Shapiro delay in all of the pulsars and obtain weak detections for PSRs J0900–3144, J1216–6410, and J1327–0755. We report a higher significance detection of the Shapiro delay for PSR J1543–5149, giving a precise pulsar mass of Mp = 1.349+0.043-0.061M☉ and companion white-dwarf mass Mc = 0.223+0.005-0.007M☉. This is an atypically low-mass measurement for a recycled MSP. In addition to these Shapiro delays, we also obtain timing model parameters including proper motions and parallax constraints for most of the pulsars.
  • Thumbnail Image
    Item
    Towards cosmological inference on unlabeled out-of-distribution hi observational data
    (Springer Science and Business Media B.V., 2025) Andrianomena, Sambatra; Hassan, Sultan
    We present an approach that can be utilized in order to account for the covariate shift between two datasets of the same observable with different distributions. This helps improve the generalizability of a neural network model trained on in-distribution samples (IDs) when inferring cosmology at the field level on out-of-distribution samples (OODs) of unknown labels. We make use of HI maps from the two simulation suites in CAMELS, IllustrisTNG and SIMBA. We consider two different techniques, namely adversarial approach and optimal transport, to adapt a target network whose initial weights are those of a source network pre-trained on a labeled dataset. Results show that after adaptation, salient features that are extracted by source and target encoders are well aligned in the embedding space. This indicates that the target encoder has learned the representations of the target domain via the adversarial training and optimal transport. Furthermore, in all scenarios considered in our analyses, the target encoder, which does not have access to any labels (Ωm) during adaptation phase, is able to retrieve the underlying Ωm from out-of-distribution maps to a great accuracy of R2 score ≥ 0.9, comparable to the performance of the source encoder trained in a supervised learning setup. We further test the viability of the techniques when only a few out-of-distribution instances are available for training and find that the target encoder still reasonably recovers the matter density. Our approach is critical in extracting information from upcoming large scale surveys.