arXiv subjects

Ashwin Devaraj

Publications and source records attributed to Ashwin Devaraj.

5 recordsLinked to original sources

arXiv2025

Discovery of a new phase-transient cyclotron line in A0535+26: Constraints on the accretion geometry

In November 2020, A0535+26 underwent one of its brightest outbursts, reaching nearly 12 Crab in X-ray flux. Observed by \textit{Insight-HXMT}, \textit{NuSTAR}, \textit{NICER}, and \textit{AstroSat}, this event provided valuable insights into Be/X-ray binaries. The pulse profiles evolved significantly with luminosity, transitioning from pencil-beam to fan-beam geometries. A0535+26, known for its fundamental cyclotron line at $\sim$44 keV, became only the second source to exhibit a negative correlation between cyclotron line energy and flux at high luminosities, with a plateau phase preceding the transition from positive to negative correlation. We report the discovery of a phase-transient low-energy cyclotron line, detected in a narrow phase range ($\sim$16\%) across all seven \textit{NuSTAR} observations during the rising, peak, and declining phases of the outburst. The new line exhibited dramatic variations with pulse phase and luminosity. We explain this behavior using an accretion geometry where the accretion column sweeps across the line of sight.

astro-ph.HE↗

arXiv2023

$\textit{Insights}$ into the phase-dependent cyclotron line feature in XTE J1946+274: An $\textit{AstroSat}$ and $\textit{Insight}$-HXMT view

XTE J1946+274 is a Be/X-ray binary with a 15.8s spin period and 172 d orbital period. Using $\textit{RXTE/PCA}$ data of the 1998 outburst, a cyclotron line around 37 keV was reported. The presence of this line, its dependence on the pulse phase, and its variation with luminosity have been of some debate since. In this work, we present the reanalysis of two $\textit{AstroSat}$ observations: one made during the rising phase of the 2018 outburst and the other during the declining phase of the 2021 outburst. We also present a new analysis of the $\textit{Insight}$-HXMT observations of the source at the peak of the 2018 outburst. We find the source to be spinning up over the course of the outburst and spinning down between the two outbursts. We report the presence of a higher cyclotron line energy using the 2018 $\textit{AstroSat}$ observation ($\sim 45$ keV) and 2018 $\textit{Insight}$-HXMT observation ($\sim$ 50 keV) and a line at $\sim$ 40 keV during the declining phase of the 2021 outburst using data from $\textit{AstroSat}$. We also investigate the pulse phase dependence of the cyclotron line parameters and find that the line is significantly detected in all the phases of both $\textit{AstroSat}$ observations, along with showing variation with the pulse phase. This differs from the previous results reported using $\textit{BeppoSAX}$ and $\textit{NuSTAR}$. We explain this behaviour of the cyclotron line to be due to photon spawning and different accretion column radii at the two poles of this neutron star.

astro-ph.HE↗

arXiv2022

Evaluating Factuality in Text Simplification

Automated simplification models aim to make input texts more readable. Such methods have the potential to make complex information accessible to a wider audience, e.g., providing access to recent medical literature which might otherwise be impenetrable for a lay reader. However, such models risk introducing errors into automatically simplified texts, for instance by inserting statements unsupported by the corresponding original text, or by omitting key information. Providing more readable but inaccurate versions of texts may in many cases be worse than providing no such access at all. The problem of factual accuracy (and the lack thereof) has received heightened attention in the context of summarization models, but the factuality of automatically simplified texts has not been investigated. We introduce a taxonomy of errors that we use to analyze both references drawn from standard simplification datasets and state-of-the-art model outputs. We find that errors often appear in both that are not captured by existing evaluation metrics, motivating a need for research into ensuring the factual accuracy of automated simplification models.

cs.CL↗

arXiv2022

NuSTAR discovery of a cyclotron line in GRO J1750-27

GRO J1750-27, discovered during an outburst in 1995 with CGRO-BATSE, is one of the farthest known galactic Be-X-ray binary systems. This relatively poorly studied system recently went into an outburst in September 2021. The source was observed during the latest outburst using the NuSTAR telescope during the rising phase of the outburst. We estimate the spin period of the source to be 4.45 s using which we produced energy-resolved pulse profiles between 3 and 65 keV. We find that the profile is double-peaked at low energies (<18 keV) while evolving into a single peak at higher energies (>18 keV). The broadband spectrum of this source was fitted well with a high energy cutoff power-law model and we report the discovery of a cyclotron resonant scattering feature (CRSF) in this source at 43 keV, indicating a magnetic field strength of $3.7 \times 10^{12}$ G. Our estimate of the magnetic field strength using the cyclotron line is consistent with the estimates made earlier using the accretion torque model from measurements of spin-up rates and fluxes during the previous outbursts.

astro-ph.HE↗

arXiv2021

Paragraph-level Simplification of Medical Texts

We consider the problem of learning to simplify medical texts. This is important because most reliable, up-to-date information in biomedicine is dense with jargon and thus practically inaccessible to the lay audience. Furthermore, manual simplification does not scale to the rapidly growing body of biomedical literature, motivating the need for automated approaches. Unfortunately, there are no large-scale resources available for this task. In this work we introduce a new corpus of parallel texts in English comprising technical and lay summaries of all published evidence pertaining to different clinical topics. We then propose a new metric based on likelihood scores from a masked language model pretrained on scientific texts. We show that this automated measure better differentiates between technical and lay summaries than existing heuristics. We introduce and evaluate baseline encoder-decoder Transformer models for simplification and propose a novel augmentation to these in which we explicitly penalize the decoder for producing "jargon" terms; we find that this yields improvements over baselines in terms of readability.

cs.CL↗