arXiv subjects

Prabhav Singh

Publications and source records attributed to Prabhav Singh.

4 recordsLinked to original sources

arXiv2026

Impact of Primordial Black Holes Induced Neutrinos on the Cosmic 21-cm Brightness Temperature

We study the impact of neutrinos emitted from evaporating Primordial Black Holes (PBHs) on the global 21-cm absorption signal during the dark ages and pre-reionization epochs. PBHs emit neutrinos over a wide energy range through Hawking evaporation. We investigate the possibility that radiative scattering between these neutrinos and the Cosmic Neutrino Background (C$ν$B) generates secondary photons, leading to additional heating of the neutral hydrogen gas. This modifies the thermal history of the intergalactic medium and increases the global 21-cm brightness temperature relative to the standard cosmological prediction. Using the absorption feature at redshift $z\simeq17$, we derive new constraints on the PBH fraction for PBH masses in the range $10^{15}\mathrm{g}\lesssim m_{\rm BH,0}\lesssim10^{25}\mathrm{g}$. We further use existing PBH limits to constrain neutrino self-interaction couplings over a broad range of mediator masses. Our analysis complements previous studies that focused on direct photon injection from PBH evaporation and highlights the importance of neutrino-induced effects within a multimessenger framework for probing PBHs and beyond-standard-model neutrino interactions.

hep-ph↗

arXiv2026

KM3-230213A and IceCube Neutrino Events from Metastable Dark Matter of Primordial Black Hole Origin

We investigate a scenario in which the recently observed ultra-high-energy neutrino event KM3-230213A, with a median energy of approximately 220 PeV, as well as the high-energy neutrinos detected by IceCube Observatory, originate from the decay of superheavy dark matter (DM) particles produced through primordial black hole (PBH) evaporation. To establish this connection, we derive constraints on the PBH abundance parameter $β$ as a function of the initial PBH mass $M_{\mathrm{BH_0}}$ and DM mass $m_{\mathrm{DM}}$, by considering the bound from the observed relic DM abundance. Using these constraints, we compute the resulting neutrino flux and show that DM masses in the PeV-EeV range can yield neutrinos of comparable energies, capable of accounting for both the KM3-230213A and IceCube events while remaining consistent with the relic abundance constraint. Interestingly, the scenario remains viable over a broad region of parameter space while satisfying existing cosmological and astrophysical bounds. Overall, our results demonstrate that PBH evaporation followed by DM decay provides a consistent and natural explanation for the observed ultra-high-energy neutrino events in the absence of accompanying multimessenger signatures.

hep-ph↗

arXiv2025

Sensitivity of Stability: Theoretical & Empirical Analysis of Replicability for Adaptive Data Selection in Transfer Learning

The widespread adoption of transfer learning has revolutionized machine learning by enabling efficient adaptation of pre-trained models to new domains. However, the reliability of these adaptations remains poorly understood, particularly when using adaptive data selection strategies that dynamically prioritize training examples. We present a comprehensive theoretical and empirical analysis of replicability in transfer learning, introducing a mathematical framework that quantifies the fundamental trade-off between adaptation effectiveness and result consistency. Our key contribution is the formalization of selection sensitivity ($Δ_Q$), a measure that captures how adaptive selection strategies respond to perturbations in training data. We prove that replicability failure probability: the likelihood that two independent training runs produce models differing in performance by more than a threshold, increases quadratically with selection sensitivity while decreasing exponentially with sample size. Through extensive experiments on the MultiNLI corpus using six adaptive selection strategies - ranging from uniform sampling to gradient-based selection - we demonstrate that this theoretical relationship holds precisely in practice. Our results reveal that highly adaptive strategies like gradient-based and curriculum learning achieve superior task performance but suffer from high replicability failure rates, while less adaptive approaches maintain failure rates below 7%. Crucially, we show that source domain pretraining provides a powerful mitigation mechanism, reducing failure rates by up to 30% while preserving performance gains. These findings establish principled guidelines for practitioners to navigate the performance-replicability trade-off and highlight the need for replicability-aware design in modern transfer learning systems.

cs.LG↗

arXiv2022

SEMI-FND: Stacked Ensemble Based Multimodal Inference For Faster Fake News Detection

Fake News Detection (FND) is an essential field in natural language processing that aims to identify and check the truthfulness of major claims in a news article to decide the news veracity. FND finds its uses in preventing social, political and national damage caused due to misrepresentation of facts which may harm a certain section of society. Further, with the explosive rise in fake news dissemination over social media, including images and text, it has become imperative to identify fake news faster and more accurately. To solve this problem, this work investigates a novel multimodal stacked ensemble-based approach (SEMIFND) to fake news detection. Focus is also kept on ensuring faster performance with fewer parameters. Moreover, to improve multimodal performance, a deep unimodal analysis is done on the image modality to identify NasNet Mobile as the most appropriate model for the task. For text, an ensemble of BERT and ELECTRA is used. The approach was evaluated on two datasets: Twitter MediaEval and Weibo Corpus. The suggested framework offered accuracies of 85.80% and 86.83% on the Twitter and Weibo datasets respectively. These reported metrics are found to be superior when compared to similar recent works. Further, we also report a reduction in the number of parameters used in training when compared to recent relevant works. SEMI-FND offers an overall parameter reduction of at least 20% with unimodal parametric reduction on text being 60%. Therefore, based on the investigations presented, it is concluded that the application of a stacked ensembling significantly improves FND over other approaches while also improving speed.

cs.CL↗