arXiv subjects

Ujjwal Upadhyay

Publications and source records attributed to Ujjwal Upadhyay.

12 recordsLinked to original sources

arXiv2026

Can Distance Duality Violation Save Late-time Solutions to the Hubble Tension?

The discrepancy between early- and late-Universe determinations of the Hubble constant may point to physics beyond $Λ$CDM or to unaccounted-for systematics. Numerous late-time modifications to the expansion history have been proposed to alleviate this discrepancy, with limited success. Recent works have shown that, when the sound-horizon and supernova calibrations are held fixed, any purely late-time resolution requires a violation of the cosmic distance duality relation (CDDR). Recasting the tension in the $r_d$-$M_B$ plane, we show explicitly that distance duality, together with BAO and uncalibrated supernova data and a fixed sound-horizon calibration, determines $H_0$ independently of the late-time expansion history. We then test the viability of the required CDDR violation by separately constraining reciprocity violation and photon number non-conservation, deriving a new constraint on reciprocity-violating distortions of angular-diameter distances from BAO and cosmic-chronometer data. Combining this result with existing photon-number-conservation constraints, we find that the level of distance-duality violation needed to resolve the tension is strongly disfavoured by current data. We therefore conclude that, for fixed sound-horizon and supernova calibrations, no modification confined to the late-time expansion history -- even one violating distance duality -- can resolve the Hubble tension, pointing instead toward early-Universe physics or unresolved local systematics.

astro-ph.CO↗

arXiv2026

Time Blindness: Why Video-Language Models Can't See What Humans Can?

Recent advances in vision-language models (VLMs) have made impressive strides in understanding spatio-temporal relationships in videos. However, when spatial information is obscured, these models struggle to capture purely temporal patterns. We introduce $\textbf{SpookyBench}$, a benchmark where information is encoded solely in temporal sequences of noise-like frames, mirroring natural phenomena from biological signaling to covert communication. Interestingly, while humans can recognize shapes, text, and patterns in these sequences with over 98% accuracy, state-of-the-art VLMs achieve 0% accuracy. This performance gap highlights a critical limitation: an over-reliance on frame-level spatial features and an inability to extract meaning from temporal cues. Furthermore, when trained in data sets with low spatial signal-to-noise ratios (SNR), temporal understanding of models degrades more rapidly than human perception, especially in tasks requiring fine-grained temporal reasoning. Overcoming this limitation will require novel architectures or training paradigms that decouple spatial dependencies from temporal processing. Our systematic analysis shows that this issue persists across model scales and architectures. We release SpookyBench to catalyze research in temporal pattern recognition and bridge the gap between human and machine video understanding. Dataset and code has been made available on our project website: https://timeblindness.github.io/.

cs.CV↗

arXiv2026

A Bayesian estimator for peculiar velocity correction in cosmological inference from supernovae data

The peculiar motion of the host galaxies introduces bias in estimating cosmological parameters from supernova data. The coherent component of the peculiar motion is usually corrected for using velocity field reconstruction based on the observed galaxy distribution, while the random component is treated statistically by inflating the magnitude uncertainty in the quadrature derived using the standard error propagation. The method of velocity field reconstruction requires assuming an underlying cosmology, which can introduce its own bias in the final inference. On the other hand, the statistical treatment of the random component assumes a locally linear approximation for the magnitude-redshift relation and a Gaussian distribution for the peculiar velocities, which can have extended tails in the non-linear regime. In this work, we present a Bayesian estimator for simultaneously correcting for peculiar motion while fitting a cosmological model to the supernova data, relaxing the assumption of linearity of the model and Gaussianity of the random peculiar motion. Our approach is based on considering the problem of fitting the magnitude-redshift relation as a non-linear model with errors in both dependent and independent variables. To this end, we develop a general method for fitting such non-linear errors-in-variables models. We then specialize it to the case of fitting the magnitude-redshift relation, validating it with simulated datasets at the precision of current and upcoming surveys, and testing it on the Pantheon sample. Our method provides an alternative approach for accounting for the peculiar velocity effects, which is a complementary method for the coherent component, as it does not require independent velocity measurements, and generalizes the treatment of the random component. Moreover, our general method is applicable to various other problems in cosmology and astronomy.

astro-ph.CO↗

arXiv2026

Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations

The peculiar motions of galaxies are powerful cosmological probes that trace the growth of structures and the distribution of matter in the universe, providing a means to investigate the nature of dark energy and test gravity on cosmological scales. However, their direct observation is extremely challenging, as it requires independent and precise distance measurements to galaxies. We present a Bayesian approach to estimate the radial component of peculiar velocities of galaxies hosting Type Ia supernovae (SNe Ia), relying solely on the background cosmological model and the precision of the SNe Ia data. Unlike other peculiar velocity estimators based on Hubble residuals, our method does not assume local linearity of the magnitude-redshift relation or a fixed cosmology, making it unbiased even for large peculiar velocities and self-consistently avoiding bias due to a wrong cosmology. We validate our method using simulated supernova data with the precision of current and upcoming surveys, and further compare it with the linearized estimator to test its efficacy. We show that our estimator has lower bias than the standard estimator and remains consistent even for larger values of $v_{\rm p}/cz$. We also present a Bayesian derivation for the linearized estimator generalized to include the supernova magnitude covariance.

astro-ph.CO↗

arXiv2026

Updated Constraints on Infrared Cutoff Models and Implications for Large-Scale CMB Anomalies

The nearly scale-invariant primordial power spectrum provides the standard initial conditions for cosmological perturbations. However, the largest scales remain only weakly constrained by CMB observations, leaving room for deviations such as an infrared (IR) cut-off. This possibility is further motivated by the persistence of large-scale CMB anomalies, most notably the low quadrupole power. In this work, we revisit several broad classes of phenomenologically motivated IR cut-off scenarios using parametrised functional forms of the primordial power spectrum. We confront these models with the latest CMB, BAO, and supernova data and derive updated constraints on the cut-off scale and associated features. Our results remain consistent with earlier studies, showing that although such models suppress power at low multipoles, the improvement in fit is marginal and does not overcome the associated parameter penalties. We therefore find no statistically significant evidence favouring IR cut-off models over the standard power-law spectrum with current data. We further explore the interplay between IR cut-off features and a possible increase in the reionisation optical depth, motivated by the recent CMB-BAO tension highlighted by DESI DR2 within the $Λ$CDM framework. We show that the additional freedom introduced by large-scale suppression is generally insufficient to support a substantial increase in optical depth, owing to the weak statistical preference for suppressed large-scale temperature power. Finally, we examine the implications of IR cut-off models for large-scale CMB anomalies by analysing the corresponding anomaly statistics within a Bayesian framework.

astro-ph.CO↗

arXiv2026

Revisiting Cosmic Distance Duality with Megamasers and DESI DR2 Observations: Model Independent Constraints on Early-Late Calibration

The Cosmic Distance Duality Relation (CDDR) connects the angular diameter distance ($d_A$) and the luminosity distance ($d_L$) at a given redshift. This fundamental relation holds in any metric theory of gravity, provided that photon number is conserved and light propagates along null geodesics. A deviation from this relation could indicate new physics beyond the standard cosmological model. In this work, we test the validity of the CDDR at very low redshifts ($z < 0.04$) by combining $d_A$ from the Megamaser Cosmology Project with $d_L$ from the Pantheon+ sample of Type Ia Supernovae (SNIa). We further incorporate high-redshift Baryon Acoustic Oscillation (BAO)-based $d_A$ measurements from DESI DR2 in combination with SNIa data, highlighting the critical role of the $r_d-M_b$ (early-late) calibration in testing the CDDR using these two probes. Assuming CDDR holds, we perform a Bayesian analysis to derive model-independent constraints on the calibration parameters. Using only BAO and SNIa data, we observe a strong degeneracy between $r_d$ and $M_b$. However, the inclusion of calibration-free Megamaser measurements breaks this degeneracy, enabling independent constraints without relying on a specific cosmological model or distance-ladder techniques. Additionally, we present a forecast incorporating the expected precision from future Megamaser and SNIa observations, demonstrating their potential to significantly tighten constraints on early-late calibration parameters, under the assumption of validity of CDDR.

astro-ph.CO↗

arXiv2025

3DCoMPaT$^{++}$: An improved Large-scale 3D Vision Dataset for Compositional Recognition

In this work, we present 3DCoMPaT$^{++}$, a multimodal 2D/3D dataset with 160 million rendered views of more than 10 million stylized 3D shapes carefully annotated at the part-instance level, alongside matching RGB point clouds, 3D textured meshes, depth maps, and segmentation masks. 3DCoMPaT$^{++}$ covers 41 shape categories, 275 fine-grained part categories, and 293 fine-grained material classes that can be compositionally applied to parts of 3D objects. We render a subset of one million stylized shapes from four equally spaced views as well as four randomized views, leading to a total of 160 million renderings. Parts are segmented at the instance level, with coarse-grained and fine-grained semantic levels. We introduce a new task, called Grounded CoMPaT Recognition (GCR), to collectively recognize and ground compositions of materials on parts of 3D objects. Additionally, we report the outcomes of a data challenge organized at CVPR2023, showcasing the winning method's utilization of a modified PointNet$^{++}$ model trained on 6D inputs, and exploring alternative techniques for GCR enhancement. We hope our work will help ease future research on compositional 3D Vision.

cs.CV↗

arXiv2025

Exploring cosmological imprints of phantom crossing with dynamical dark energy in Horndeski gravity

In the current era of precision cosmology, the persistence of cosmological tensions, most notably the Hubble tension and the $S_8$ tension, challenges the standard $Λ$CDM model. To reconcile these tensions via late-time modifications to expansion history, various features such as phantom crossing in the dark energy equation of state, a negative energy density at high redshifts, etc., are favoured. However, these scenarios cannot be realized within the framework of GR without introducing ghost or gradient instabilities. In this work, we investigate a dynamical dark energy scenario within the framework of Horndeski gravity, incorporating nonminimal coupling to gravity and self-interactions. We highlight that the model can exhibit novel features like phantom crossing and negative dark energy densities at high redshifts without introducing any instabilities. For this specific Horndeski model, we perform a comprehensive analysis of the background evolution along with the effects on perturbations, examining observables like growth rate, matter and CMB power spectrum. To check the consistency of the model with the observational data, we employ MCMC analysis using BAO/$fσ_8$, Supernovae, and CMB data. While the model does not outperform the standard $Λ$CDM framework in a combined likelihood analysis, there remains a preference for non-zero values of the model parameters within the data. This suggests that dynamical dark energy scenarios, particularly those with non-minimal couplings, merit further exploration as promising alternatives to GR, offering rich phenomenology that can be tested against a broader range of current and upcoming observational datasets.

astro-ph.CO↗

arXiv2023

Identification of Hemorrhage and Infarct Lesions on Brain CT Images using Deep Learning

Head Non-contrast computed tomography (NCCT) scan remain the preferred primary imaging modality due to their widespread availability and speed. However, the current standard for manual annotations of abnormal brain tissue on head NCCT scans involves significant disadvantages like lack of cutoff standardization and degeneration identification. The recent advancement of deep learning-based computer-aided diagnostic (CAD) models in the multidisciplinary domain has created vast opportunities in neurological medical imaging. Significant literature has been published earlier in the automated identification of brain tissue on different imaging modalities. However, determining Intracranial hemorrhage (ICH) and infarct can be challenging due to image texture, volume size, and scan quality variability. This retrospective validation study evaluated a DL-based algorithm identifying ICH and infarct from head-NCCT scans. The head-NCCT scans dataset was collected consecutively from multiple diagnostic imaging centers across India. The study exhibits the potential and limitations of such DL-based software for introduction in routine workflow in extensive healthcare facilities.

eess.IV↗

arXiv2022

Deep-ASPECTS: A Segmentation-Assisted Model for Stroke Severity Measurement

A stroke occurs when an artery in the brain ruptures and bleeds or when the blood supply to the brain is cut off. Blood and oxygen cannot reach the brain's tissues due to the rupture or obstruction resulting in tissue death. The Middle cerebral artery (MCA) is the largest cerebral artery and the most commonly damaged vessel in stroke. The quick onset of a focused neurological deficit caused by interruption of blood flow in the territory supplied by the MCA is known as an MCA stroke. Alberta stroke programme early CT score (ASPECTS) is used to estimate the extent of early ischemic changes in patients with MCA stroke. This study proposes a deep learning-based method to score the CT scan for ASPECTS. Our work has three highlights. First, we propose a novel method for medical image segmentation for stroke detection. Second, we show the effectiveness of AI solution for fully-automated ASPECT scoring with reduced diagnosis time for a given non-contrast CT (NCCT) Scan. Our algorithms show a dice similarity coefficient of 0.64 for the MCA anatomy segmentation and 0.72 for the infarcts segmentation. Lastly, we show that our model's performance is inline with inter-reader variability between radiologists.

eess.IV↗

arXiv2022

Generating Out of Distribution Adversarial Attack using Latent Space Poisoning

Traditional adversarial attacks rely upon the perturbations generated by gradients from the network which are generally safeguarded by gradient guided search to provide an adversarial counterpart to the network. In this paper, we propose a novel mechanism of generating adversarial examples where the actual image is not corrupted rather its latent space representation is utilized to tamper with the inherent structure of the image while maintaining the perceptual quality intact and to act as legitimate data samples. As opposed to gradient-based attacks, the latent space poisoning exploits the inclination of classifiers to model the independent and identical distribution of the training dataset and tricks it by producing out of distribution samples. We train a disentangled variational autoencoder (beta-VAE) to model the data in latent space and then we add noise perturbations using a class-conditioned distribution function to the latent space under the constraint that it is misclassified to the target label. Our empirical results on MNIST, SVHN, and CelebA dataset validate that the generated adversarial examples can easily fool robust l_0, l_2, l_inf norm classifiers designed using provably robust defense mechanisms.

cs.CV↗

arXiv2019

Indian EmoSpeech Command Dataset: A dataset for emotion based speech recognition in the wild

Speech emotion analysis is an important task which further enables several application use cases. The non-verbal sounds within speech utterances also play a pivotal role in emotion analysis in speech. Due to the widespread use of smartphones, it becomes viable to analyze speech commands captured using microphones for emotion understanding by utilizing on-device machine learning models. The non-verbal information includes the environment background sounds describing the type of surroundings, current situation and activities being performed. In this work, we consider both verbal (speech commands) and non-verbal sounds (background noises) within an utterance for emotion analysis in real-life scenarios. We create an indigenous dataset for this task namely "Indian EmoSpeech Command Dataset". It contains keywords with diverse emotions and background sounds, presented to explore new challenges in audio analysis. We exhaustively compare with various baseline models for emotion analysis on speech commands on several performance metrics. We demonstrate that we achieve a significant average gain of 3.3% in top-one score over a subset of speech command dataset for keyword spotting.

eess.AS↗