arXiv subjects

Rahul Sharma

Publications and source records attributed to Rahul Sharma.

At least 73 records · Page 4Linked to original sources

arXiv2023

Discretizing Numerical Attributes: An Analysis of Human Perceptions

Machine learning (ML) has employed various discretization methods to partition numerical attributes into intervals. However, an effective discretization technique remains elusive in many ML applications, such as association rule mining. Moreover, the existing discretization techniques do not reflect best the impact of the independent numerical factor on the dependent numerical target factor. This research aims to establish a benchmark approach for numerical attribute partitioning. We conduct an extensive analysis of human perceptions of partitioning a numerical attribute and compare these perceptions with the results obtained from our two proposed measures. We also examine the perceptions of experts in data science, statistics, and engineering by employing numerical data visualization techniques. The analysis of collected responses reveals that $68.7\%$ of human responses approximately closely align with the values generated by our proposed measures. Based on these findings, our proposed measures may be used as one of the methods for discretizing the numerical attributes.

cs.LG↗

arXiv2023

AstroSat observation of the magnetar SGR J1830$-$0645 during its first detected X-ray outburst

We present here timing and spectral analyses of SGR J1830$-$0645 based on an AstroSat observation carried out on 2020 October 16, about a week after the onset of its first detected X-ray outburst. Using data taken with the Soft X-ray Telescope (SXT) and Large Area X-ray Proportional Counter (LAXPC), we have detected 0.9$-$10 keV coherent pulsations at a period of $\sim$10.4 s. The pulse profiles were single-peaked, asymmetric and consisted of minor peaks attributable to hotspots on the neutron star surface. The pulsed fraction evolved significantly with energy, increasing to energies around 5 keV with a steep drop thereafter. The 0.9--25 keV SXT--LAXPC energy spectrum is best described with two thermal components having temperatures $\sim$0.46 and $\sim$1.1 keV (emission radii of $\sim$2.4 and $\sim$0.65 km, respectively, assuming a distance of 4 kpc) along with a power-law component having a photon index of $\sim$0.39. We report the detection of 67 X-ray bursts having an average duration of $\sim$33 ms. The brightest burst lasted for about 90 ms and had a 3--25 keV fluence of $\sim 5 \times 10^{-9}$ erg cm$^{-2}$.

astro-ph.HE↗

arXiv2023

Probing spectral and timing properties of the X-ray pulsar RX J0440.9+4431 in the giant outburst of 2022-2023

The X-ray pulsar RX J0440.9+4431 went through a giant outburst in 2022 and reached a record-high flux of 2.3 Crab, as observed by Swift/BAT. We study the evolution of different spectral and timing properties of the source using NICER observations. The pulse period is found to decrease from 208 s to 205 s, and the pulse profile evolves significantly with energy and luminosity. The hardness ratio and hardness intensity diagram (HID) show remarkable evolution during the outburst. The HID turns towards the diagonal branch from the horizontal branch above a transition (critical) luminosity, suggesting the presence of two accretion modes. Each NICER spectrum can be described using a cutoff power law with a blackbody component and a Gaussian at 6.4 keV. At higher luminosities, an additional Gaussian at 6.67 keV is used. The observed photon index shows negative and positive correlations with X-ray flux below and above the critical luminosity, respectively. The evolution of spectral and timing parameters suggests a possible change in the emission mechanism and beaming pattern of the pulsar depending on the spectral transition to sub- and super-critical accretion regimes. Based on the critical luminosity, the magnetic field of the neutron star can be estimated in the order of 10$^{12}$ or 10$^{13}$ G, assuming different theoretical models. Moreover, the observed iron emission line evolves from a narrow to a broad feature with luminosity. Two emission lines originating from neutral and highly ionized Fe atoms were evident in the spectra around 6.4 keV and 6.67 keV (higher luminosities).

astro-ph.HE↗

arXiv2023

4U 1626-67 Returns to Spin-Down: Timing features toe the line

We present a comprehensive analysis of X-ray pulsar 4U 1626-67 during its current spin-down (2SD) state, following a recent torque reversal. Since its discovery, this ultra-compact binary has experienced multiple torque states, transitioning from spin-up (1SU) during 1977-1990 to spin-down (1SD) during 1990-2008, and again spin-up (2SU) until 2023. From NuSTAR observation of May 2023, we have investigated the timing and spectral properties of this pulsar during its 2SD phase, while also comparing them with previous spin-up-down states. For energies upto 8 keV, a distinct bi-horned pulse profile was observed during the spin-up phase, while several sub-structures emerged during spin-down. Beyond 8 keV, a broad asymmetric peak was consistently observed across all torque states. The pulse fraction during the 2SD phase was higher than that during 2SU phase. A prominent ~46.8 mHz quasi-periodic oscillation has been exclusively detected during the spin-down phase. The broadband spectrum during the 2SD phase is described by empirical NPEX model, cyclotron absorption feature and its first harmonic. The spectrum during 2SU phase requires an additional blackbody component and asymmetry in the cyclotron absorption line. A significant flux drop by a factor of ~3 in the 2SD was observed.

astro-ph.HE↗

arXiv2023

StaticFixer: From Static Analysis to Static Repair

Static analysis tools are traditionally used to detect and flag programs that violate properties. We show that static analysis tools can also be used to perturb programs that satisfy a property to construct variants that violate the property. Using this insight we can construct paired data sets of unsafe-safe program pairs, and learn strategies to automatically repair property violations. We present a system called \sysname, which automatically repairs information flow vulnerabilities using this approach. Since information flow properties are non-local (both to check and repair), \sysname also introduces a novel domain specific language (DSL) and strategy learning algorithms for synthesizing non-local repairs. We use \sysname to synthesize strategies for repairing two types of information flow vulnerabilities, unvalidated dynamic calls and cross-site scripting, and show that \sysname successfully repairs several hundred vulnerabilities from open source {\sc JavaScript} repositories, outperforming neural baselines built using {\sc CodeT5} and {\sc Codex}. Our datasets can be downloaded from \url{http://aka.ms/StaticFixer}.

cs.SE↗

arXiv2023

Numerical Association Rule Mining: A Systematic Literature Review

Numerical association rule mining is a widely used variant of the association rule mining technique, and it has been extensively used in discovering patterns and relationships in numerical data. Initially, researchers and scientists integrated numerical attributes in association rule mining using various discretization approaches; however, over time, a plethora of alternative methods have emerged in this field. Unfortunately, the increase of alternative methods has resulted into a significant knowledge gap in understanding diverse techniques employed in numerical association rule mining -- this paper attempts to bridge this knowledge gap by conducting a comprehensive systematic literature review. We provide an in-depth study of diverse methods, algorithms, metrics, and datasets derived from 1,140 scholarly articles published from the inception of numerical association rule mining in the year 1996 to 2022. In compliance with the inclusion, exclusion, and quality evaluation criteria, 68 papers were chosen to be extensively evaluated. To the best of our knowledge, this systematic literature review is the first of its kind to provide an exhaustive analysis of the current literature and previous surveys on numerical association rule mining. The paper discusses important research issues, the current status, and future possibilities of numerical association rule mining. On the basis of this systematic review, the article also presents a novel discretization measure that contributes by providing a partitioning of numerical data that meets well human perception of partitions.

cs.LG↗

arXiv2023

Training with Mixed-Precision Floating-Point Assignments

When training deep neural networks, keeping all tensors in high precision (e.g., 32-bit or even 16-bit floats) is often wasteful. However, keeping all tensors in low precision (e.g., 8-bit floats) can lead to unacceptable accuracy loss. Hence, it is important to use a precision assignment -- a mapping from all tensors (arising in training) to precision levels (high or low) -- that keeps most of the tensors in low precision and leads to sufficiently accurate models. We provide a technique that explores this memory-accuracy tradeoff by generating precision assignments for convolutional neural networks that (i) use less memory and (ii) lead to more accurate convolutional networks at the same time, compared to the precision assignments considered by prior work in low-precision floating-point training. We evaluate our technique on image classification tasks by training convolutional networks on CIFAR-10, CIFAR-100, and ImageNet. Our method typically provides > 2x memory reduction over a baseline precision assignment while preserving training accuracy, and gives further reductions by trading off accuracy. Compared to other baselines which sometimes cause training to diverge, our method provides similar or better memory reduction while avoiding divergence.

cs.LG↗

arXiv2023

AstroSat and NuSTAR observations of XTE J1739-285 during the 2019-2020 outburst

We report results from a study of XTE J1739-285, a transient neutron star low mass X-ray binary observed with AstroSat and NuSTAR during its 2019-2020 outburst. We detected accretion-powered X-ray pulsations at 386 Hz during very short intervals (0.5--1 s) of X-ray flares. These flares were observed during the 2019 observation of XTE J1739-285. During this observation, we also observed a correlation between intensity and hardness ratios, suggesting an increase in hardness with the increase in intensity. Moreover, a thermonuclear X-ray burst detected in our AstroSat observation during the 2020 outburst revealed the presence of coherent burst oscillations at 383 Hz during its decay phase. The frequency drift of 3 Hz during X-ray burst can be explained with r modes. Thus, making XTE J1739-285 belong to a subset of NS-LMXBs which exhibit both nuclear- and accretion-powered pulsations. The power density spectrum created using the AstroSat-LAXPC observations in 2020 showed the presence of a quasi-periodic oscillation at ~ 0.83 Hz. Our X-ray spectroscopy revealed significant changes in the spectra during the 2019 and 2020 outburst. We found a broad iron line emission feature in the X-ray spectrum during the 2020 observation, while this feature was relatively narrow and has a lower equivalent width in 2019,~when the source was accreting at higher rates than 2020.

astro-ph.HE↗

arXiv2023

Machine Learning for Optical Motion Capture-driven Musculoskeletal Modelling from Inertial Motion Capture Data

Marker-based Optical Motion Capture (OMC) systems and associated musculoskeletal (MSK) modelling predictions offer non-invasively obtainable insights into in vivo joint and muscle loading, aiding clinical decision-making. However, an OMC system is lab-based, expensive, and requires a line of sight. Inertial Motion Capture (IMC) systems are widely-used alternatives, which are portable, user-friendly, and relatively low-cost, although with lesser accuracy. Irrespective of the choice of motion capture technique, one needs to use an MSK model to obtain the kinematic and kinetic outputs, which is a computationally expensive tool increasingly well approximated by machine learning (ML) methods. Here, we present an ML approach to map experimentally recorded IMC data to the human upper-extremity MSK model outputs computed from ('gold standard') OMC input data. Essentially, we aim to predict higher-quality MSK outputs from the much easier-to-obtain IMC data. We use OMC and IMC data simultaneously collected for the same subjects to train different ML architectures that predict OMC-driven MSK outputs from IMC measurements. In particular, we employed various neural network (NN) architectures, such as Feed-Forward Neural Networks (FFNNs) and Recurrent Neural Networks (RNNs) (vanilla, Long Short-Term Memory, and Gated Recurrent Unit) and searched for the best-fit model through an exhaustive search in the hyperparameters space in both subject-exposed (SE) & subject-naive (SN) settings. We observed a comparable performance for both FFNN & RNN models, which have a high degree of agreement (ravg, SE, FFNN = 0.90+/-0.19, ravg, SE, RNN = 0.89+/-0.17, ravg, SN, FFNN = 0.84+/-0.23, & ravg, SN, RNN = 0.78+/-0.23) with the desired OMC-driven MSK estimates for held-out test data. Mapping IMC inputs to OMC-driven MSK outputs using ML models could be instrumental in transitioning MSK modelling from 'lab to field'.

cs.LG↗

arXiv2023

Machine learning techniques for the Schizophrenia diagnosis: A comprehensive review and future research directions

Schizophrenia (SCZ) is a brain disorder where different people experience different symptoms, such as hallucination, delusion, flat-talk, disorganized thinking, etc. In the long term, this can cause severe effects and diminish life expectancy by more than ten years. Therefore, early and accurate diagnosis of SCZ is prevalent, and modalities like structural magnetic resonance imaging (sMRI), functional MRI (fMRI), diffusion tensor imaging (DTI), and electroencephalogram (EEG) assist in witnessing the brain abnormalities of the patients. Moreover, for accurate diagnosis of SCZ, researchers have used machine learning (ML) algorithms for the past decade to distinguish the brain patterns of healthy and SCZ brains using MRI and fMRI images. This paper seeks to acquaint SCZ researchers with ML and to discuss its recent applications to the field of SCZ study. This paper comprehensively reviews state-of-the-art techniques such as ML classifiers, artificial neural network (ANN), deep learning (DL) models, methodological fundamentals, and applications with previous studies. The motivation of this paper is to benefit from finding the research gaps that may lead to the development of a new model for accurate SCZ diagnosis. The paper concludes with the research finding, followed by the future scope that directly contributes to new research directions.

cs.LG↗

arXiv2022

The AstroSat observation of accreting millisecond X-ray pulsar SAX J1808.4-3658 during its 2019 outburst

We report on the analysis of the AstroSat dataset of the accreting millisecond X-ray pulsar SAX J1808.4-3658, obtained during its 2019 outburst. We found coherent pulsations at $\sim 401$ Hz and an orbital solution consistent with previous studies. The 3-20 keV pulse profile can be well fitted with three harmonically related sinusoidal components with background-corrected fractional amplitude of $\sim 3.5 \%$, $\sim 1.2 \%$ and $\sim 0.37 \%$ for fundamental, second and third harmonic, respectively. Our energy-resolved pulse profile evolution study indicate a strong energy dependence. We also observed a soft lag in fundamental and hard lag during its harmonic. The broadband spectrum of SAX J1808.4-3658 can be well described with a combination of thermal emission component with $kT \sim 1$ keV, a thermal Comptonization ($Γ\sim 1.67$) from the hot corona and broad emission lines due to Fe.

astro-ph.HE↗

arXiv2022

Broadband mHz QPOs and spectral study of LMC X$-$4 with AstroSat

We report the results of broadband timing and spectral analysis of data from an AstroSat observation of the High Mass X-ray binary LMC X$-$4. The Large Area X-ray Proportional Counter (LAXPC) and Soft X-ray Telescope (SXT) instruments on-board the AstroSat observed the source in August 2016. A complete X-ray eclipse was detected with the LAXPC. The 3$-$40 keV power density spectrum showed the presence of coherent pulsations along with a $\sim 26$ mHz quasi-periodic oscillation feature. The spectral properties of LMC X$-$4 were derived from a joint analysis of the SXT and LAXPC spectral data. The 0.5$-$25 keV persistent spectrum comprised of an absorbed high energy cutoff power law with photon index of $Γ\sim$ 0.8 and cutoff at $\sim$16 keV, a soft thermal component with kT$_{BB} \sim$ 0.14 keV and Gaussian components corresponding to Fe K$_α$, Ne \textsc{ix} and Ne \textsc{x} emission lines. Assuming a source distance of 50 kpc, we determined 0.5--25 keV luminosity to be $\sim 2 \times 10^{38}$ erg s$^{-1}$.

astro-ph.HE↗

arXiv2022

Audio-Visual Activity Guided Cross-Modal Identity Association for Active Speaker Detection

Active speaker detection in videos addresses associating a source face, visible in the video frames, with the underlying speech in the audio modality. The two primary sources of information to derive such a speech-face relationship are i) visual activity and its interaction with the speech signal and ii) co-occurrences of speakers' identities across modalities in the form of face and speech. The two approaches have their limitations: the audio-visual activity models get confused with other frequently occurring vocal activities, such as laughing and chewing, while the speakers' identity-based methods are limited to videos having enough disambiguating information to establish a speech-face association. Since the two approaches are independent, we investigate their complementary nature in this work. We propose a novel unsupervised framework to guide the speakers' cross-modal identity association with the audio-visual activity for active speaker detection. Through experiments on entertainment media videos from two benchmark datasets, the AVA active speaker (movies) and Visual Person Clustering Dataset (TV shows), we show that a simple late fusion of the two approaches enhances the active speaker detection performance.

cs.MM↗

arXiv2022

MinUn: Accurate ML Inference on Microcontrollers

Running machine learning inference on tiny devices, known as TinyML, is an emerging research area. This task requires generating inference code that uses memory frugally, a task that standard ML frameworks are ill-suited for. A deployment framework for TinyML must be a) parametric in the number representation to take advantage of the emerging representations like posits, b) carefully assign high-precision to a few tensors so that most tensors can be kept in low-precision while still maintaining model accuracy, and c) avoid memory fragmentation. We describe MinUn, the first TinyML framework that holistically addresses these issues to generate efficient code for ARM microcontrollers (e.g., Arduino Uno, Due and STM32H747) that outperforms the prior TinyML frameworks.

cs.LG↗

arXiv2022

Eclipse Timings of the LMXB XTE J1710-281 : Discovery of a third orbital period glitch

We present an updated measurement of orbital period evolution of LMXB XTE J1710-281 by using eclipse timing technique. Using data obtained with XMM-Newton, Suzaku, RXTE, Chandra and AstroSat observatories, we report 21 new measurements of X-ray mid-eclipse times. We have discovered a third orbital period glitch in XTE J1710-281 with an F-test false alarm probability of ~0.7% for occurrence of the third glitch and report detection of four distinct epochs of orbital period in this system. This work presents a more robust estimation of occurrence of the second orbital period glitch. However, the epoch of occurrence of the third glitch is poorly constrained, between MJD 55726 to 56402. We have put lower limits of 1.48 ms, 0.97 ms and 0.45 ms, on sudden changes in orbital period between the successive epochs. We discuss the implications of our findings in context of magnetic nature of the companion star and possible scattering events with circum-binary objects around this binary system.

astro-ph.HE↗

arXiv2022

Unsupervised active speaker detection in media content using cross-modal information

We present a cross-modal unsupervised framework for active speaker detection in media content such as TV shows and movies. Machine learning advances have enabled impressive performance in identifying individuals from speech and facial images. We leverage speaker identity information from speech and faces, and formulate active speaker detection as a speech-face assignment task such that the active speaker's face and the underlying speech identify the same person (character). We express the speech segments in terms of their associated speaker identity distances, from all other speech segments, to capture a relative identity structure for the video. Then we assign an active speaker's face to each speech segment from the concurrently appearing faces such that the obtained set of active speaker faces displays a similar relative identity structure. Furthermore, we propose a simple and effective approach to address speech segments where speakers are present off-screen. We evaluate the proposed system on three benchmark datasets -- Visual Person Clustering dataset, AVA-active speaker dataset, and Columbia dataset -- consisting of videos from entertainment and broadcast media, and show competitive performance to state-of-the-art fully supervised methods.

eess.IV↗

arXiv2022

Efficient ML Models for Practical Secure Inference

ML-as-a-service continues to grow, and so does the need for very strong privacy guarantees. Secure inference has emerged as a potential solution, wherein cryptographic primitives allow inference without revealing users' inputs to a model provider or model's weights to a user. For instance, the model provider could be a diagnostics company that has trained a state-of-the-art DenseNet-121 model for interpreting a chest X-ray and the user could be a patient at a hospital. While secure inference is in principle feasible for this setting, there are no existing techniques that make it practical at scale. The CrypTFlow2 framework provides a potential solution with its ability to automatically and correctly translate clear-text inference to secure inference for arbitrary models. However, the resultant secure inference from CrypTFlow2 is impractically expensive: Almost 3TB of communication is required to interpret a single X-ray on DenseNet-121. In this paper, we address this outstanding challenge of inefficiency of secure inference with three contributions. First, we show that the primary bottlenecks in secure inference are large linear layers which can be optimized with the choice of network backbone and the use of operators developed for efficient clear-text inference. This finding and emphasis deviates from many recent works which focus on optimizing non-linear activation layers when performing secure inference of smaller networks. Second, based on analysis of a bottle-necked convolution layer, we design a X-operator which is a more efficient drop-in replacement. Third, we show that the fast Winograd convolution algorithm further improves efficiency of secure inference. In combination, these three optimizations prove to be highly effective for the problem of X-ray interpretation trained on the CheXpert dataset.

cs.CR↗

arXiv2022

Change in spin-down rate and detection of emission line in HMXB 4U 2206+54 with AstroSat observation

This work presents timing and spectral analysis of 4U 2206+54 using data obtained from LAXPC instrument onboard India's AstroSat mission. This source was observed with AstroSat in September 2016 and October 2016. We report detection of $5648 (4)$ s pulsations at MJD 57669 in the latter observation of 4U 2206+54. The pulse profile is sinusoidal and the inherent shape is independent of energy up to 30 keV. The pulse fraction increases with energy from $\sim 0.5$% to $\sim 0.8$%. We report an updated spin down rate of $2.95 (14) \times 10^{-7}$ s s$^{-1}$. This is about 0.40 times smaller than the previously reported long term value. The energy spectrum is best modelled with an absorbed power-law with high energy exponential cut-off. We have detected presence of broad emission line in 4U 2206+54 at an energy of $7$ keV with equivalent width of $\sim 0.4$ keV.

astro-ph.HE↗