arXiv subjects

Kshitij Aggarwal

Publications and source records attributed to Kshitij Aggarwal.

At least 19 recordsLinked to original sources

arXiv2024

An unidentified Fermi source emitting radio bursts in the Galactic bulge

We report on the detection of radio bursts from the Galactic bulge using the real-time transient detection and localization system, realfast. The pulses were detected commensally on the Karl G. Jansky Very Large Array during a survey of unidentified Fermi $\gamma$-ray sources. The bursts were localized to subarcsecond precision using realfast fast-sampled imaging. Follow-up observations with the Green Bank Telescope detected additional bursts from the same source. The bursts do not exhibit periodicity in a search up to periods of 480 s, assuming a duty cycle of < 20%. The pulses are nearly 100% linearly polarized, showing circular polarization up to 12%, and exhibit variable scattering on timescales of months. The arcsecond-level realfast localization links the source confidently with the Fermi $\gamma$-ray source and places it nearby (though not coincident with) an XMM-Newton X-ray source. Based on the source's overall properties, we discuss various options for the nature of this object and propose that it could be a young pulsar, a magnetar, or a binary pulsar system.

astro-ph.HE↗

arXiv2023

The Petabyte Project

Transient radio sources, such as fast radio bursts, intermittent pulsars, and rotating radio transients, can offer a wealth of information regarding extreme emission physics as well as the intervening interstellar and/or intergalactic medium. Vital steps towards understanding these objects include characterizing their source populations and estimating their event rates across observing frequencies. However, previous efforts have been undertaken mostly by individual survey teams at disparate observing frequencies and telescopes, and with non-uniform algorithms for searching and characterization. The Petabyte Project (TPP) aims to address these issues by uniformly reprocessing data from several petabytes of radio transient surveys covering two decades of observing frequency (300 MHz-20 GHz). The TPP will provide robust event rate analyses, in-depth assessment of survey and pipeline completeness, as well as revealing discoveries from archival and ongoing radio surveys. We present an overview of TPP's processing pipeline, scope, and our potential to make new discoveries.

astro-ph.IM↗

arXiv2023

Temporal and Spectral Properties of the Persistent Radio Source Associated with FRB 20190520B with the VLA

Among more than 800 known fast radio bursts (FRBs), only two, namely FRB 20121102A and FRB 20190520B, are confirmed to be associated with a persistent radio sources (PRS). Here we report evidence of apparent temporal variability in the PRS associated with the bursting FRB 20190520B based on the Karl G. Jansky Very Large Array (VLA) observations taken in 2020 and 2021. Based on the analysis of epoch-to-epoch variability of the PRS at L, S, C, and X band in 1-12 GHz, we detected not only overall marginal variability but also a likely radio flux decrease ($\sim$ 3.2 $\sigma$) between the observations taken in 2020 and 2021 at 3 GHz. Assuming no spectral variation in the PRS during these observations, we found the evidence for an overall broadband radio flux decrease by about 20 percent between the 2020 and the 2021 observations, suggesting that the PRS probably evolves on the yearly time scale. If we attribute the marginal variability at 3 GHz as intrinsic or due to scintillation, the size of potential variable component of the PRS is constrained to be sub-parsec. On the other hand, the size of the PRS can be also constrained to be larger than about 0.22 parsec from the averaged radio spectrum and the integrated radio luminosity in the 1-12 GHz band based on equipartition and self-absorption arguments. We discuss potential origins of the PRS and suggest that an accreting compact object origin might be able to explain the PRS's temporal and spectral properties. Confirmation of variability or flux decline of the PRS would be critical to our understanding of the PRS and its relation to the bursting source.

astro-ph.HE↗

arXiv2022

Masked Image Modeling Advances 3D Medical Image Analysis

Recently, masked image modeling (MIM) has gained considerable attention due to its capacity to learn from vast amounts of unlabeled data and has been demonstrated to be effective on a wide variety of vision tasks involving natural images. Meanwhile, the potential of self-supervised learning in modeling 3D medical images is anticipated to be immense due to the high quantities of unlabeled images, and the expense and difficulty of quality labels. However, MIM's applicability to medical images remains uncertain. In this paper, we demonstrate that masked image modeling approaches can also advance 3D medical images analysis in addition to natural images. We study how masked image modeling strategies leverage performance from the viewpoints of 3D medical image segmentation as a representative downstream task: i) when compared to naive contrastive learning, masked image modeling approaches accelerate the convergence of supervised training even faster (1.40$\times$) and ultimately produce a higher dice score; ii) predicting raw voxel values with a high masking ratio and a relatively smaller patch size is non-trivial self-supervised pretext-task for medical images modeling; iii) a lightweight decoder or projection head design for reconstruction is powerful for masked image modeling on 3D medical images which speeds up training and reduce cost; iv) finally, we also investigate the effectiveness of MIM methods under different practical scenarios where different image resolutions and labeled data ratios are applied.

cs.CV↗

arXiv2022

On the radio spectra of Galactic millisecond pulsars

With recent advances in the sensitivity of radio surveys of the Galactic disk, the number of millisecond pulsars (MSPs) has increased substantially in recent years such that it is now possible to study their demographic properties in more detail than in the past. We investigate what can be learned about the radio spectra of the MSP population. Using a sample of 179 MSPs detected in eleven surveys carried out at radio frequencies in the range 0.135-6.6 GHz, we carry out detailed modeling of MSP radio spectral behaviour in this range. Employing Markov Chain Monte Carlo simulations to explore a multi-dimensional parameter space, and accurately accounting for observational selection effects, we find strong evidence in favour of the MSP population having a two-component power-law spectral model scaling with frequency, $ν$. Specifically, we find that MSP flux density spectra are approximately independent of frequency below 320 MHz, and proportional to $ν^{-1.5}$ at higher frequencies. This parameterization performs significantly better than single power-law models which over predict the number of MSPs seen in low-frequency (100-200 MHz) surveys. We compared our results with earlier work, and current understanding of the normal pulsar population, and use our model to make predictions for MSP yields in upcoming surveys. We demonstrate that the observed sample of MSPs could triple in the coming decade.

astro-ph.HE↗

arXiv2022

Magnetic field reversal in the turbulent environment around a repeating fast radio burst

Fast radio bursts (FRBs) are brief, intense flashes of radio waves from unidentified extragalactic sources. Polarized FRBs originate in highly magnetized environments. We report observations of the repeating FRB 20190520B spanning seventeen months , which show its amount of Faraday rotation is highly variable and twice changes its sign. The FRB also depolarizes below radio frequencies around 1 to 3 GHz. We interpret these properties as due to change in the parallel component of the integrated magnetic field along the line-of-sight, including reversals. This could result from propagation through a turbulent, magnetized screen of plasma located between $10^{-5}$ to 100 parsecs of the FRB source. This is consistent with the bursts passing through the stellar wind of a binary companion of the FRB source.

astro-ph.HE↗

arXiv2021

Characterizing the FRB host galaxy population and its connection to transients in the local and extragalactic Universe

We present the localization and host galaxies of one repeating and two apparently non-repeating Fast Radio Bursts. FRB20180301A was detected and localized with the Karl G. Jansky Very Large Array to a star-forming galaxy at $z=0.3304$. FRB20191228A, and FRB20200906A were detected and localized by the Australian Square Kilometre Array Pathfinder to host galaxies at $z=0.2430$ and $z=0.3688$, respectively. We combine these with 13 other well-localized FRBs in the literature, and analyze the host galaxy properties. We find no significant differences in the host properties of repeating and apparently non-repeating FRBs. FRB hosts are moderately star-forming, with masses slightly offset from the star-forming main-sequence. Star formation and low-ionization nuclear emission-line region (LINER) emission are major sources of ionization in FRB host galaxies, with the former dominant in repeating FRB hosts. FRB hosts do not track stellar mass and star formation as seen in field galaxies (more than 95% confidence). FRBs are rare in massive red galaxies, suggesting that progenitor formation channels are not solely dominated by delayed channels which lag star formation by Gigayears. The global properties of FRB hosts are indistinguishable from core-collapse supernovae (CCSNe) and short gamma-ray bursts (SGRBs) hosts, and the spatial offset (from galaxy centers) of FRBs is mostly inconsistent with that of the Galactic neutron star population (95% confidence). The spatial offsets of FRBs (normalized to the galaxy effective radius) also differ from those of globular clusters (GCs) in late- and early-type galaxies with 95% confidence.

astro-ph.HE↗

arXiv2021

Observational effects of banded repeating FRBs

Recent observations have shown that repeating Fast Radio Bursts (FRBs) exhibit band-limited emission, whose frequency-dependent amplitude can be modeled using a Gaussian function. In this analysis, we show that banded emission of FRBs can lead to incompleteness across the observing band. This biases the detected sample of bursts and can explain the various shapes of cumulative energy distributions seen for repeating FRBs. We assume a Gaussian shape of the burst spectra and used simulations to demonstrate the above bias using an FRB 121102-like example. We recovered energy distributions that showed a break in power-law and flattening of power-law at low energies, based on the fluence threshold of the observations. We provide recommendations for single-pulse searches and analysis of repeating FRBs to account for this incompleteness. Primarily, we recommend that burst spectra should be modeled to estimate the intrinsic fluence and bandwidth of the burst robustly. Also, bursts that lie mainly within the observing band should be used for analyses of energy distributions. We show that the bimodality reported in the distribution of energies of FRB 121102 by Li et al. (2021) disappears when burst bandwidth, instead of the center frequency of the observation, is used to estimate energy. Sub-banded searches will also aid in detecting band-limited bursts. All the analysis scripts used in this work are available in a Github repository.

astro-ph.HE↗

arXiv2021

Comprehensive analysis of a dense sample of FRB 121102 bursts

We present an analysis of a densely repeating sample of bursts from the first repeating fast radio burst, FRB 121102. We reanalysed the data used by Gourdji et al. (2019) and detected 93 additional bursts using our single-pulse search pipeline. In total, we detected 133 bursts in three hours of data at a center frequency of 1.4 GHz using the Arecibo telescope, and develop robust modeling strategies to constrain the spectro-temporal properties of all the bursts in the sample. Most of the burst profiles show a scattering tail, and burst spectra are well modeled by a Gaussian with a median width of 230 MHz. We find a lack of emission below 1300 MHz, consistent with previous studies of FRB 121102. We also find that the peak of the log-normal distribution of wait times decreases from 207 s to 75 s using our larger sample of bursts, as compared to that of Gourdji et al. (2019). Our observations do not favor either Poissonian or Weibull distributions for the burst rate distribution. We searched for periodicity in the bursts using multiple techniques but did not detect any significant period. The cumulative burst energy distribution exhibits a broken power-law shape, with the lower and higher-energy slopes of $-0.4\pm0.1$ and $-1.8\pm0.2$, with the break at $(2.3\pm0.2)\times 10^{37}$ ergs. We provide our burst fitting routines as a python package BURSTFIT that can be used to model the spectrogram of any complex FRB or pulsar pulse using robust fitting techniques. All the other analysis scripts and results are publicly available.

astro-ph.HE↗

arXiv2021

Multi-wavelength follow-up of FRB 180309

We report on the results of multi-wavelength follow-up observations with Gemini, VLA, and ATCA, to search for a host galaxy and any persistent radio emission associated with FRB 180309. This FRB is among the most luminous FRB detections to date, with a luminosity of $> 8.7\times 10^{32}$ erg Hz$^{-1}$ at the dispersion-based redshift upper limit of 0.32. We used the high-significance detection of FRB 180309 with the Parkes Telescope and a beam model of the Parkes Multibeam Receiver to improve the localization of the FRB to a region spanning approximately $\sim2'\times2'$. We aimed to seek bright galaxies within this region to determine the strongest candidates as the originator of this highly luminous FRB. We identified optical sources within the localization region above our r-band magnitude limit of 24.27, fourteen of which have photometric redshifts whose fitted mean is consistent with the redshift upper limit ($z < 0.32$) of our FRB. Two of these galaxies are coincident with marginally detected "persistent" radio sources of flux density 24.3$μ$Jy beam$^{-1}$ and 22.1$μ$Jy beam$^{-1}$ respectively. Our redshift-dependent limit on the luminosity of any associated persistent radio source is comparable to the luminosity limits for other localized FRBs. We analyze several properties of the candidate hosts we identified, including chance association probability, redshift, and presence of radio emission, however it remains possible that any of these galaxies could be the host of this FRB. Follow-up spectroscopy on these objects to explore their H$α$ emission and ionization contents, as well as to obtain more precisely measured redshifts, may be able to isolate a single host for this luminous FRB.

astro-ph.HE↗

arXiv2021

The host galaxy and persistent radio counterpart of FRB 20201124A

The physical properties of fast radio burst (FRB) host galaxies provide important clues towards the nature of FRB sources. The 16 FRB hosts identified thus far span three orders of magnitude in mass and specific star-formation rate, implicating a ubiquitously occurring progenitor object. FRBs localised with ~arcsecond accuracy also enable effective searches for associated multi-wavelength and multi-timescale counterparts, such as the persistent radio source associated with FRB 20121102A. Here we present a localisation of the repeating source FRB 20201124A, and its association with a host galaxy (SDSS J050803.48+260338.0, z=0.098) and persistent radio source. The galaxy is massive ($\sim3\times10^{10} M_{\odot}$), star-forming (few solar masses per year), and dusty. Very Large Array and Very Long Baseline Array observations of the persistent radio source measure a luminosity of $1.2\times10^{29}$ erg s$^{-1}$ Hz$^{-1}$, and show that is extended on scales $\gtrsim50$ mas. We associate this radio emission with the ongoing star-formation activity in SDSS J050803.48+260338.0. Deeper, more detailed observations are required to better utilise the milliarcsecond-scale localisation of FRB 20201124A reported from the European VLBI Network, and determine the origin of the large dispersion measure ($150-220$ pc cm$^{-3}$) contributed by the host. SDSS J050803.48+260338.0 is an order of magnitude more massive than any galaxy or stellar system previously associated with a repeating FRB source, but is comparable to the hosts of so far non-repeating FRBs, further building the link between the two apparent populations.

astro-ph.HE↗

arXiv2021

Robust Assessment of Clustering Methods for Fast Radio Transient Candidates

Fast radio transient search algorithms identify signals of interest by iterating and applying a threshold on a set of matched filters. These filters are defined by properties of the transient such as time and dispersion. A real transient can trigger hundreds of search trials, each of which has to be post-processed for visualization and classification tasks. In this paper, we have explored a range of unsupervised clustering algorithms to cluster these redundant candidate detections. We demonstrate this for Realfast, the commensal fast transient search system at the Very Large Array. We use four features for clustering: sky position (l, m), time and dispersion measure (DM). We develop a custom performance metric that makes sure that the candidates are clustered into a small number of pure clusters, i.e, clusters with either astrophysical or noise candidates. We then use this performance metric to compare eight different clustering algorithms. We show that using sky location along with DM/time improves clustering performance by $\sim$10% as compared to the traditional DM/time-based clustering. Therefore, positional information should be used during clustering if it can be made available. We conduct several tests to compare the performance and generalisability of clustering algorithms to other transient datasets and propose a strategy that can be used to choose an algorithm. Our performance metric and clustering strategy can be easily extended to different single-pulse search pipelines and other astronomy and non-astronomy-based applications.

astro-ph.IM↗

arXiv2021

Probabilistic Association of Transients to their Hosts (PATH)

We introduce a new method to estimate the probability that an extragalactic transient source is associated with a candidate host galaxy. This approach relies solely on simple observables: sky coordinates and their uncertainties, galaxy fluxes and angular sizes. The formalism invokes Bayes' rule to calculate the posterior probability P(O_i|x) from the galaxy prior P(O), observables x, and an assumed model for the true distribution of transients in/around their host galaxies. Using simulated transients placed in the well-studied COSMOS field, we consider several agnostic and physically motivated priors and offset distributions to explore the method sensitivity. We then apply the methodology to the set of 13~fast radio bursts (FRBs) localized with an uncertainty of several arcseconds. Our methodology finds nine of these are securely associated to a single host galaxy, P(O_i|x)>0.95. We examine the observed and intrinsic properties of these secure FRB hosts, recovering similar distributions as previous works. Furthermore, we find a strong correlation between the apparent magnitude of the securely identified host galaxies and the estimated cosmic dispersion measures of the corresponding FRBs, which results from the Macquart relation. Future work with FRBs will leverage this relation and other measures from the secure hosts as priors for future associations. The methodology is generic to transient type, localization error, and image quality. We encourage its application to other transients where host galaxy associations are critical to the science, e.g. gravitational wave events, gamma-ray bursts, and supernovae. We have encoded the technique in Python on GitHub: https://github.com/FRBs/astropath.

astro-ph.HE↗

arXiv2020

Your: Your Unified Reader

The advancement in signal processing and GPU based systems has enabled new transient detectors at various telescopes to perform much more sensitive searches than their predecessors. Typically the data output from the telescopes is in one of the two commonly used formats: psrfits and Sigproc filterbank. Software developed for transient searches often only works with one of these two formats, limiting their general applicability. Therefore, researchers have to write custom scripts to read/write the data in their format of choice before they can begin any data analysis relevant for their research. \textsc{Your} (Your Unified Reader) is a python-based library that unifies the data processing across multiple commonly used formats. \textsc{Your} implements a user-friendly interface to read and write in the data format of choice. It also generates unified metadata corresponding to the input data file for a quick understanding of observation parameters and provides utilities to perform common data analysis operations. \textsc{Your} also provides several state-of-the-art radio frequency interference mitigation (RFI) algorithms, which can now be used during any stage of data processing (reading, writing, etc.) to filter out artificial signals.

astro-ph.IM↗

arXiv2020

Host Galaxy Properties and Offset Distributions of Fast Radio Bursts: Implications for their Progenitors

We present observations and detailed characterizations of five new host galaxies of fast radio bursts (FRBs) discovered with the Australian Square Kilometre Array Pathfinder (ASKAP) and localized to $\lesssim 1''$. Combining these galaxies with FRB hosts from the literature, we introduce criteria based on the probability of chance coincidence to define a sub-sample of 10 highly-confident associations (at $z=0.03-0.52$), three of which correspond to known repeating FRBs. Overall, the FRB host galaxies exhibit a broad, continuous range of color ($M_u-M_r = 0.9 - 2.0$), stellar mass ($M_\star = 10^{8} - 6\times 10^{10}\,M_{\odot}$), and star-formation rate (${\rm SFR} = 0.05 - 10\,M_{\odot}\,{\rm yr}^{-1}$) spanning the full parameter space occupied by $z<0.5$ galaxies. However, they do not track the color-magnitude, SFR-$M_\star$, nor BPT diagrams of field galaxies surveyed at similar redshifts. There is an excess of "green valley" galaxies and an excess of emission-line ratios indicative of a harder radiation field than that generated by star-formation alone. From the observed stellar mass distribution, we rule out the hypothesis that FRBs strictly track stellar mass in galaxies ($>99\%$ c.l.). We measure a median offset of 3.3 kpc from the FRB to the estimated center of the host galaxies and compare the host-burst offset distribution and other properties with the distributions of long- and short-duration gamma-ray bursts (LGRBs and SGRBs), core-collapse supernovae (CC-SNe), and Type Ia SNe. This analysis rules out galaxies hosting LGRBs (faint, star-forming galaxies) as common hosts for FRBs ($>95\%$ c.l.). Other transient channels (SGRBs, CC- and Type Ia SNe) have host galaxy properties and offsets consistent with the FRB distributions. All of the data and derived quantities are made publicly available on a dedicated website and repository.

astro-ph.GA↗

arXiv2020

VLA/Realfast Detection of a Burst from FRB180916.J0158+65 and Tests for Periodic Activity

We report on the detection of a burst from FRB180916 by realfast/VLA and present software for interpreting fast radio bursts (FRB) periodicity. We demonstrate a range of periodicity analyses with bursts from FRB180916, FRB121102 and FRB180814. Our results for FRB180916 and FRB121102 are consistent with published results. For FRB180814, we did not detect any significant periodic episodes. The realfast-detected and other high-frequency bursts for FRB180916 tend to lie at the beginning of the activity window, indicating a possible phase-frequency relation. The python package $\texttt{frbpa}$ can be used to reproduce and expand on this analysis to test models for repeating FRBs.

astro-ph.HE↗

arXiv2020

FETCH: A deep-learning based classifier for fast transient classification

With the upcoming commensal surveys for Fast Radio Bursts (FRBs), and their high candidate rate, usage of machine learning algorithms for candidate classification is a necessity. Such algorithms will also play a pivotal role in sending real-time triggers for prompt follow-ups with other instruments. In this paper, we have used the technique of Transfer Learning to train the state-of-the-art deep neural networks for classification of FRB and Radio Frequency Interference (RFI) candidates. These are convolutional neural networks which work on radio frequency-time and dispersion measure-time images as the inputs. We trained these networks using simulated FRBs and real RFI candidates from telescopes at the Green Bank Observatory. We present 11 deep learning models, each with an accuracy and recall above 99.5% on our test dataset comprising of real RFI and pulsar candidates. As we demonstrate, these algorithms are telescope and frequency agnostic and are able to detect all FRBs with signal-to-noise ratios above 10 in ASKAP and Parkes data. We also provide an open-source python package FETCH (Fast Extragalactic Transient Candidate Hunter) for classification of candidates, using our models. Using FETCH, these models can be deployed along with any commensal search pipeline for real-time candidate classification.

astro-ph.IM↗

arXiv2020

A Demonstration of Extremely Low Latency $γ$-ray, X-Ray & UV Follow-Up of a Millisecond Radio Transient

We report results of a novel high-energy follow-up observation of a potential Fast Radio Burst. The radio burst was detected by VLA/realfast and followed-up by the Neil Gehrels Swift Observatory in very low latency utilizing new operational capabilities of Swift (arXiv:2005.01751), with pointed soft X-ray and UV observations beginning at T0+32 minutes, and hard X-ray/gamma-ray event data saved around T0. These observations are $>10$x faster than previous X-ray/UV follow-up of any radio transient to date. No emission is seen coincident with the FRB candidate at T0, with a 0.2s fluence $5σ$ upper limit of $1.35\times10^{-8}$ erg cm$^{-2}$ (14-195 keV) for a SGR 1935+2154-like flare, nor at T0+32 minutes down to $3σ$ upper limits of 22.18 AB mag in UVOT u band, and $3.33\times10^{-13}$ erg cm$^{-2}$ s$^{-1}$ from 0.3-10 keV for the 2 ks observation. The candidate FRB alone is not significant enough to be considered astrophysical, so this note serves as a technical demonstration. These new Swift operational capabilities will allow future FRB detections to be followed up with Swift at even lower latencies than demonstrated here: 15-20 minutes should be regularly achievable, and 5-10 minutes occasionally achievable. We encourage FRB detecting facilities to release alerts in low latency to enable this science.

astro-ph.HE↗