BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Yogesh Wadadekar

Yogesh Wadadekar contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
astro-ph.GAastro-ph.IMMachine LearningArtificial Intelligenceastro-ph.CO

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
8works
0followers
5topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

8 published item(s)

preprint2026arXiv

Photometric Redshift Estimation Using Scaled Ensemble Learning

The development of the state-of-the-art telescopic systems capable of performing expansive sky surveys such as the Sloan Digital Sky Survey, Euclid, and the Rubin Observatory's Legacy Survey of Space and Time (LSST) has significantly advanced efforts to refine cosmological models. These advances offer deeper insight into persistent challenges in astrophysics and our understanding of the Universe's evolution. A critical component of this progress is the reliable estimation of photometric redshifts (Pz). To improve the precision and efficiency of such estimations, the application of machine learning (ML) techniques to large-scale astronomical datasets has become essential. This study presents a new ensemble-based ML framework aimed at predicting Pz for faint galaxies and higher redshift ranges, relying solely on optical (grizy) photometric data. The proposed architecture integrates several learning algorithms, including gradient boosting machine, extreme gradient boosting, k-nearest neighbors, and artificial neural networks, within a scaled ensemble structure. By using bagged input data, the ensemble approach delivers improved predictive performance compared to stand-alone models. The framework demonstrates consistent accuracy in estimating redshifts, maintaining strong performance up to z ~ 4. The model is validated using publicly available data from the Hyper Suprime-Cam Strategic Survey Program by the Subaru Telescope. Our results show marked improvements in the precision and reliability of Pz estimation. Furthermore, this approach closely adheres to-and in certain instances exceeds-the benchmarks specified in the LSST Science Requirements Document. Evaluation metrics include catastrophic outlier, bias, and rms.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Star-forming S0 Galaxies in SDSS-MaNGA: fading spirals or rejuvenated S0s?

We investigate the origin of rare star-formation in an otherwise red-and-dead population of S0 galaxies using spatially resolved spectroscopy. Our sample consists of $120$ low redshift ($z<0.1$) star-forming S0 (SF-S0) galaxies from the SDSS-IV MaNGA DR15. We have selected this sample after a visual inspection of deep images from the DESI Legacy Imaging Surveys DR9 and the Subaru/HSC-SSP survey PDR3, to remove contamination from spiral galaxies. We also construct two control samples of star-forming spirals (SF-Sps) and quenched S0s (Q-S0s) to explore their evolutionary link with the star-forming S0s. To study star-formation at resolved scales, we use dust-corrected $H_α$ luminosity and stellar density ($Σ_\star$) maps to construct radial profiles of star-formation rate (SFR) surface density ($Σ_{SFR}$) and specific SFR (sSFR). Examining these radial profiles, we find that star-formation in SF-S0s is centrally dominated as opposed to disc dominated star-formation in spirals. We also compared various global (size-mass relation, bulge-to-total luminosity ratio) and local (central stellar velocity dispersion) properties of SF-S0s to those of the control sample galaxies. We find that SF-S0s are structurally similar to the quenched S0s and are different from star-forming spirals. We infer that SF-S0s are unlikely to be fading spirals. Inspecting stellar and gas velocity maps, we find that more than $50\%$ of the SF-S0 sample shows signs of recent galaxy interactions such as kinematic misalignment, counter-rotation, and unsettled kinematics. Based on these results, we conclude that in our sample of SF-S0s, star-formation has been rejuvenated, with minor mergers likely to be a major driver.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Machine Learning for Scientific Discovery

Machine Learning algorithms are good tools for both classification and prediction purposes. These algorithms can further be used for scientific discoveries from the enormous data being collected in our era. We present ways of discovering and understanding astronomical phenomena by applying machine learning algorithms to data collected with radio telescopes. We discuss the use of supervised machine learning algorithms to predict the free parameters of star formation histories and also better understand the relations between the different input and output parameters. We made use of Deep Learning to capture the non-linearity in the parameters. Our models are able to predict with low error rates and give the advantage of predicting in real time once the model has been trained. The other class of machine learning algorithms viz. unsupervised learning can prove to be very useful in finding patterns in the data. We explore how we use such unsupervised techniques on solar radio data to identify patterns and variations, and also link such findings to theories, which help to better understand the nature of the system being studied. We highlight the challenges faced in terms of data size, availability, features, processing ability and importantly, the interpretability of results. As our ability to capture and store data increases, increased use of machine learning to understand the underlying physics in the information captured seems inevitable.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Remnant radio galaxy candidates of small angular sizes

Remnant radio galaxies (RRGs), characterized by the cessation of AGN activity, represent a short-lived last phase of radio galaxy&#39;s life-cycle. Hitherto, searches for RRGs, mainly based on the morphological criteria, have identified large angular size sources resulting into a bias towards the remnants of powerful FR-II radio galaxies. In this study we make the first attempt to perform a systematic search for RRGs of small angular sizes ($<$30$^{\prime\prime}$) in the XMM-LSS field. By using spectral curvature criterion we discover 48 remnant candidates exhibiting strong spectral curvature ${\it i.e.}$; $α_{\rm 150~MHz}^{\rm 325~MHz}$ - $α_{\rm 325~MHz}^{\rm 1.4~GHz}$ $\geq$0.5. Spectral characteristics at higher frequency regime ($>$1.4 GHz) indicate that some of our remnant candidates can depict recurrent AGN activity with an active core. We place an upper limit on the remnant fraction ($f_{\rm rem}$) to be 3.9$\%$, which increases to 5.4$\%$ if flux cutoff limit of S$_{\rm 150~MHz}$ $\geq$10 mJy is considered. Our study unveils, hitherto unexplored, a new population of small-size ($<$200 kpc) remnant candidates that are often found to reside in less dense environments and at higher redshifts ($z$) $>$1.0. We speculate that a relatively shorter active phase and/or low jet power can be plausible reasons for the small size of remnant candidates.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

An upper limit calculator (UL-CALC) for undetected extended sources with radio interferometers: radio halo upper limits

Radio halos are diffuse, extended sources of radio emission detected primarily in massive, merging galaxy clusters. In smaller and/or relaxed clusters, where no halos are detected, one can instead place upper limits to a possible radio emission. Detections and upper limits are both crucial to constrain theoretical models for the generation of radio halos. The upper limits are model dependent for radio interferometers and thus the process of obtaining these is tedious to perform manually. In this paper, we present a Python based tool to automate this process of estimating the upper limits. The tool allows users to create radio halos with defined parameters like physical size, redshift and brightness model. A family of radio halo models with a range of flux densities, decided based on the rms noise of the image, is then injected into the parent visibility file and imaged. The halo injected image and the original image are then compared to check for the radio halo detection using a threshold on the detected excess flux density. Injections separated by finer differences in the flux densities are carried out once the coarse range where the upper limit is likely to be located has been identified. The code recommends an upper limit and provides a range of images for manual inspection. The user may then decide on the upper limit. We discuss the advantages and limitations of this tool. A wider usage of this tool in the context of the ongoing and upcoming all sky surveys with the LOFAR and SKA is proposed with the aim of constraining the physics of radio halo formation. The tool is publicly available at https://github.com/lijotgeorge/UL-CALC.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Characterizing EoR foregrounds: A study of the Lockman Hole Region at 325 MHz

One of the key science goals for the most sensitive telescopes, both current and upcoming, is the detection of the redshifted 21-cm signal from the Cosmic Dawn and Epoch of Reionization. The success of detection relies on accurate foreground modeling for their removal from data sets. This paper presents the characterization of astrophysical sources in the Lockman Hole region. Using 325 MHz data obtained from the GMRT, a $6^\circ \times 6^\circ$ mosaiced map is produced with an RMS reaching 50 $μ$Jy $\mathrm{beam}^{-1}$. A source catalog containing 6186 sources is created, and the Euclidean normalized differential source counts have been derived from it, consistent with previous observations as well as simulations. A detailed comparison of the source catalog is also made with previous findings - at both lower and higher frequencies. The angular power spectrum (APS) of the diffuse Galactic synchrotron emission is determined for three different galactic latitudes using the Tapered Gridded Estimator. The values of the APS lie between $\sim$1 mK$^2$ to $\sim$100 mK$^2$. Fitting a power law of the form $A\ell^{-β}$ gives values of $A$ and $β$ varying across the latitudes considered. This paper demonstrates, for the first time, the variation of the power-law index for diffuse emission at very high galactic locations. It follows the same trend that is seen at locations near the galactic plane, thus emphasizing the need for low-frequency observations for developing better models of the diffuse emission.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Predicting star formation properties of galaxies using deep learning

Understanding the star-formation properties of galaxies as a function of cosmic epoch is a critical exercise in studies of galaxy evolution. Traditionally, stellar population synthesis models have been used to obtain best fit parameters that characterise star formation in galaxies. As multiband flux measurements become available for thousands of galaxies, an alternative approach to characterising star formation using machine learning becomes feasible. In this work, we present the use of deep learning techniques to predict three important star formation properties -- stellar mass, star formation rate and dust luminosity. We characterise the performance of our deep learning models through comparisons with outputs from a standard stellar population synthesis code.

0 citations0 reviewsNo code linkedOpen access
preprint2018arXiv

Outlying H$α$ blobs in SDSS IV MaNGA

We have discovered a population of 29 outlying H$α$ emitters which appear like unresolved blobs in the DR14 data release of the SDSS IV MaNGA integral field unit survey. They do not have any underlying optical continuum emission in deep imaging from the DECam Legacy Survey or Beijing-Arizona Sky Survey. These blobs either lie away from the disc of the host galaxy in the MaNGA IFU and/or have velocities which are different from the velocity map of the host galaxy. Interestingly, all of them show photoionisation due to star formation. These galaxies have very high specific star formation rates compared to the known population of dwarf galaxies. However, their metallicities are consistent with or even lower than those of the local volume dwarfs. Thus, we can possibly rule out tidal dwarf galaxies. They could represent a new population of low mass and starbursting dwarf galaxies.

0 citations0 reviewsNo code linkedOpen access