arXiv subjects

Pawan Kumar

Publications and source records attributed to Pawan Kumar.

At least 37 records · Page 2Linked to original sources

arXiv2025

Sports Business Administration and New Age Technology: Role of AI

This chapter explores the complexities of sports governance, taxation, dispute resolution, and the impact of digital transformation within the sports sector. This study identifies a critical research gap regarding the integration of innovative technologies to enhance governance and talent identification in sports law. The objective is to evaluate how data-driven approaches and AI can optimize recruitment processes; also ensuring compliance with existing regulations. A comprehensive analysis of current governance structures and taxation policies,(ie Income Tax Act and GST Act), reveals preliminary results indicating that reform is necessary to support sustainable growth in the sports economy. Key findings demonstrate that AI enhances player evaluation by minimizing biases and expanding access to diverse talent pools. While the Court of Arbitration for Sport provides an efficient mechanism for dispute resolution. The implications emphasize the need for regulatory reforms that align taxation policies with international best practices, promoting transparency and accountability in sports organizations. This research contributes valuable insights into the evolving dynamics of sports management, aiming to foster innovation and integrity in the industry.

cs.CY↗

arXiv2025

Hierarchical Sparse Plus Low Rank Compression of LLM

Modern large language models (LLMs) place extraordinary pressure on memory and compute budgets, making principled compression indispensable for both deployment and continued training. We present Hierarchical Sparse Plus Low-Rank (HSS) compression, a two-stage scheme that (i) removes the largest-magnitude weights into a sparse matrix S and (ii) applies a recursive Hierarchically Sparse Separable (HSS) low-rank factorisation to the dense residual matrix. A recursive rank-reducing strategy and a reverse Cuthill-Mckee (RCM) permutation are introduced to align high weights towards the diagonal with the block-diagonal hierarchy, maximising off-diagonal compressibility (because they are touched only once). HSS is hardware-friendly: its matrix-vector multiply reduces to one sparse and a sequence of thin-matrix multiplications and can be trained end-to-end with standard optimisers. Experiments on LLaMA-7B show that targeting only the self-attention projections (1.6 B parameters of Q, K, and V matrices out of a total 7B parameters) suffices to yield large memory savings while retaining comparable state-of-the-art perplexity scores on test samples of the WikiText dataset. For example, with a 30\% sparsity budget and an outer rank of 512, sHSS-RCM achieves a perplexity of 1.64, outperforming dense baselines and classical sparse-plus-SVD variants, while also achieving significant memory savings.

cs.LG↗

arXiv2025

Polarization Angle Orthogonal Jumps in Fast Radio Bursts

Recently, polarization angle (PA) orthogonal jumps over millisecond timescales were discovered from three bursts of a repeating fast radio burst source FRB 20201124A by the FAST telescope. In general, PA jumps can arise from the coherent or incoherent superposition of two electromagnetic waves, with total polarization fraction remains constant in the former and not in the latter. The observations seem to be more consistent with incoherent superposition. The amplitudes of the two orthogonal modes are required to be comparable when jumps occur. We provide general constraints on FRB emission and propagation mechanisms based on the data. Physically, it is difficult to produce PA jumps through switching the dominance of the two orthogonal modes within millisecond timescales, and a geometric effect due to the source rotation is more plausible. This requires that the emission region be within the magnetosphere of a spinning central engine, likely a magnetar. The two orthogonal modes in different directions can arise when the source rotation brings two independent emission regions with different dominant modes successively into the line-of-sight, either due to intrinsic radiation mechanisms or the O-mode undergoing a delayed transparency because of the Alfvén-O-mode conversion. Splitting of emission directions for the two modes due to plasma birefringence is not easy to achieve when the plasma is moving relativistically. For intrinsic radiation mechanisms, curvature radiation always predicts $|E_{\rm X}/E_{\rm O}|\gtrsim1$, and is difficult to produce jumps; whereas inverse Compton scattering can achieve the conversion amplitude ratio $|E_{\rm X}/E_{\rm O}|=1$ to allow jumps to occur under special geometric configurations.

astro-ph.HE↗

arXiv2025

Simulating droplet adhesion on superhydrophobic surfaces

A numerical model is proposed to simulate the adhesion, compression, and subsequent detachment of a micro-liter droplet from a superhydrophobic surface composed of chemically homogeneous pillars arranged in a periodic fashion, replicating a typical force probe microscopy experiment. We observe that as the droplet is pulled away from the surface, the net vertical force varies in a typical sawtooth manner with peculiar peaks and troughs, characteristic of the surface. The force first reaches a maximum before the droplet detaches from the surface with a comparatively lower force. The force variation predicted by the numerical model is in good agreement with the experimental results of Kumar et al. [1]. We also studied the effect of evaporation on the variation in the adhesion force by simulating an evaporating droplet on a superhydrophobic surface. For an evaporating droplet, the numerically predicted maximum and detachment force magnitudes are in good agreement with those obtained experimentally when we take into account the change in the droplet weight as it evaporates. The proposed method will be useful for the quantitative analysis and design of a variety of superhydrophobic surfaces and will pave the way for more accurate surface characterization based on droplet adhesion force measurements.

physics.flu-dyn↗

arXiv2025

Probing superhydrophobic surface topography using droplet adhesion

Understanding contact line dynamics on superhydrophobic surfaces with microscopic structures is essential for designing materials with reduced drag, anti-icing, self-cleaning, and anti-fouling properties. Using numerical simulations, we demonstrate that forces on droplets receding over structured surfaces are governed by microscale deformations near the contact line. We present and experimentally validate an expression demonstrating that adhesion force increases logarithmically with pillar area fraction at constant droplet volume and pillar surface chemistry. Furthermore, we establish that the average tensile force measured in direct force measurements provides a more reliable indicator of surface structure than the commonly used maximum force. This newfound insight enables precise quantification of superhydrophobic surface structure using a droplet probe.

physics.flu-dyn↗

arXiv2025

A Fast and Efficient Modern BERT based Text-Conditioned Diffusion Model for Medical Image Segmentation

In recent times, denoising diffusion probabilistic models (DPMs) have proven effective for medical image generation and denoising, and as representation learners for downstream segmentation. However, segmentation performance is limited by the need for dense pixel-wise labels, which are expensive, time-consuming, and require expert knowledge. We propose FastTextDiff, a label-efficient diffusion-based segmentation model that integrates medical text annotations to enhance semantic representations. Our approach uses ModernBERT, a transformer capable of processing long clinical notes, to tightly link textual annotations with semantic content in medical images. Trained on MIMIC-III and MIMIC-IV, ModernBERT encodes clinical knowledge that guides cross-modal attention between visual and textual features. This study validates ModernBERT as a fast, scalable alternative to Clinical BioBERT in diffusion-based segmentation pipelines and highlights the promise of multi-modal techniques for medical image analysis. By replacing Clinical BioBERT with ModernBERT, FastTextDiff benefits from FlashAttention 2, an alternating attention mechanism, and a 2-trillion-token corpus, improving both segmentation accuracy and training efficiency over traditional diffusion-based models.

cs.CV↗

arXiv2025

Quasi-Biennial Oscillations and Rieger-type Periodicities in a Babcock-Leighton Solar Dynamo

The Sun's magnetic field shows the 11-year solar cycle and shorter periodicities, popularly known as the quasi-biennial oscillations (QBOs) and Rieger-type periods, or ``season of the Sun." Although several theories have been proposed to explain the origin of QBOs and Rieger-type periods, no single theory has widespread acceptance. We explore whether the \bl\ dynamo can produce Rieger-type periodicity and QBOs and investigate their underlying physical mechanisms. We use the observationally guided three-dimensional kinematic \bl\ dynamo model, which has emerged as a successful model for reproducing many characteristic features of the solar cycle. We use Morlet wavelet and global wavelet power spectrum techniques to analyze the data obtained from the model. In our model, we report QBOs and Rieger-type periods for the first time. Further, we investigate the individual \bl\ parameters (fluctuations in flux, latitude, time delay and tilt scatter) role in the occurrence of QBOs and Rieger-type periods. We find that while fluctuations in the individual parameters of the \bl\ process can produce QBOs and Rieger-type periodicity, their occurrence probability is enhanced when we consider combined fluctuations of all parameters in the \bl\ process. Finally, we find that with the increase of dynamo supercriticality, the model tends to suppress the generation of Rieger-type periodicity. Thus, this result supports earlier studies that suggest the solar dynamo is not highly supercritical.

astro-ph.SR↗

arXiv2025

Asking Fast Radio Bursts for More than Reionization History

We propose different estimators to probe the intergalactic medium (IGM) during epoch of reionization (EoR) using the dispersion measure (${\rm DM}$) of the fast radio bursts. We consider three different reionization histories, which we can distinguish with a total of $\lesssim 1000\,{\rm DM}$ measurements during EoR if their redshifts are known. We note that the redshift derivatives of ${\rm DM}$ are also directly sensitive to the reionization history. The major point of this work is to explore the variance in the ${\rm DM}$ measurements and the information encoded in them. We find that the all-sky average $\overline{\rm DM}(z)$ gets biased from the line-of-sight (LoS) fluctuations in the ${\rm DM}$ measurements introduced by the ionization of IGM during EoR. We find that the ratio $σ_{\rm DM}/\overline{\rm DM}$ depends directly on the ionization bubble sizes as well as the reionization history. On the other hand, we also find that angular variance (coined as $\textit{structure function}$) of ${\rm DM}$ encodes the information about the duration of reionization and the typical bubble sizes as well. We establish the usefulness of variances in ${\rm DM}$ using toy models of reionization and later verify it with the realistic reionization simulations.

astro-ph.CO↗

arXiv2025

Classical to Quantum Diffusive Transport in Atomically Thin Semiconductors Capped with High-k Dielectric

The dielectric environment surrounding semiconductors plays a crucial role in determining device performance, a role that becomes especially pronounced in atomically thin semiconductors where charge carriers are confined within a few atomic layers and strongly interact with their surroundings. High-k dielectrics, such as hafnium oxide (HfO2), have been shown to enhance the performance of two-dimensional (2D) materials by suppressing scattering from charged impurities and phonons, but most studies to date have focused on room-temperature transistor operation. Their influence on quantum transport properties at low temperatures remains largely unexplored. In this work, we investigate how capping monolayer molybdenum disulfide (MoS2) with HfO2 modifies its electronic behavior in the quantum regime. By comparing devices with and without HfO2 capping, we find that uncapped devices exhibit transport dominated by classical diffusive scattering, whereas capped devices show clear Fabry-Perot interference patterns, providing direct evidence of phase-coherent quantum transport enabled by dielectric screening. To gain further insight, we develop a tight-binding interferometer model that captures the effect of dielectric screening on conductive modes and reproduces the experimental trends. Our findings demonstrate that dielectric engineering provides a powerful route to control transport regimes in TMD devices.

cond-mat.mes-hall↗

arXiv2025

Can repeating and non-repeating FRBs be drawn from the same population?

Do all Fast Radio Burst (FRB) sources repeat? We present evidence that FRB sources follow a Zipf-like distribution, in which the number density of sources is approximately inversely proportional to their burst rate above a fixed energy threshold-even though both the burst rate and number density span many orders of magnitude individually. We introduce a model-independent framework that predicts the distribution of observed fluences and distances, and repetition rates of an FRB population based on an assumed burst rate distribution per source. Using parameters derived directly from observations, this framework simultaneously explains several key features of the FRB population: (i) The observed ratio of repeaters to apparent non-repeaters; (ii) The much lower ratio of apparent non-repeaters to the total number of Soft Gamma Repeater (SGR) sources within the observable Universe; And (iii) the slightly smaller average distances of known repeaters compared to non-repeaters. We further explore how survey parameters, such as radio sensitivity and observation time, influence these statistics. Notably, we find that the fraction of repeaters rises only mildly with improved sensitivity or longer exposure. This weak dependence could be misinterpreted as evidence that not all FRBs repeat. Overall, our results support the idea that a single population-likely magnetars-can account for the full observed diversity of FRB activity, from very inactive FRB sources like SGR 1935+2154 to the most active repeaters.

astro-ph.HE↗

arXiv2025

Engineering 2D Van der Waals Electrode via MBE Grown Weyl Semimetal 1T-WTe2 for Enhanced Photodetection in InSe

Achieving low contact resistance in advanced quantum electronic devices remains a critical challenge. With the growing demand for faster and energy-efficient devices, 2D contact engineering offers a promising solution. Beyond graphene, 1T-WTe2 has attracted attention for its excellent electrical transport, quantum phenomena, and Weyl semimetallic properties. Here, we demonstrate the direct wafer-scale growth of 1T-WTe2 via molecular beam epitaxy (MBE) and its use as a 2D contact for layered materials such as InSe. The 1T WTe2/InSe interface exhibits a barrier height nearly half that of conventional metal contacts, and its contact resistance is reduced by a factor of 21, effectively suppressing Fermi level pinning and enabling efficient electron injection. InSe/1T WTe2 photodetectors show broad photoresponsivity (0.14 to 217.58 A/W) under NIR to DUV illumination with fast rise/fall times of 42/126 ms, compared to lower responsivity (0.000865 A/W to 3.64 A/W) and slower response (150/144 ms) for InSe/Ti Au devices. The 1T WTe2/InSe devices thus exhibit approximately 60 times higher responsivity and 4 times faster response than conventional metal contacts. These results establish MBE-grown 1T-WTe2 as an effective 2D electrode, enhancing photodetection performance while simplifying device architecture, making it a strong candidate for next generation nanoelectronic and optoelectronic devices.

physics.optics↗

arXiv2025

Reinforcement Learning Based Sensor Optimization for Bio-markers

Radio frequency (RF) biosensors, in particular those based on inter-digitated capacitors (IDCs), are pivotal in areas like biomedical diagnosis, remote sensing, and wireless communication. Despite their advantages of low cost and easy fabrication, their sensitivity can be hindered by design imperfections, environmental factors, and circuit noise. This paper investigates enhancing the sensitivity of IDC-based RF sensors using novel reinforcement learning based Binary Particle Swarm Optimization (RLBPSO), and it is compared to Ant Colony Optimization (ACO), and other state-of-the-art methods. By focusing on optimizing design parameters like electrode design and finger width, the proposed study found notable improvements in sensor sensitivity. The proposed RLBPSO method shows best optimized design for various frequency ranges when compared to current state-of-the-art methods.

cs.LG↗

arXiv2025

Experimental demonstration of corrugated nanolaminate films as reflective light sails

Achieving laser-driven, reflective, relativistic light sails would represent a tremendous breakthrough for humankind, allowing us to advance our understanding of the solar system and deep space far beyond what we know from space probes, telescopes, and objects passing near Earth. Numerous sail film designs have been proposed, but none have been demonstrated that satisfy all of the stringent optical, mechanical, and mass budget constraints. Here we overcome this challenge by experimentally demonstrating a novel class of optically-optimized nanolaminate sails with strong and flexible hexagonally-corrugated microstructures. Our prototypes, fabricated from alumina and molybdenum disulfide using scalable semiconductor processing techniques, feature ultra-low areal densities of <1 g/m^2 and achieve experimentally-measured reflectivities of >50% and absorptivities of <4% within the Doppler-shifted laser wavelength range corresponding to accelerating to a fifth the speed of light. Moreover, we analyze reflectivity, strength, and mass constraints to show that our sails have the potential to achieve greater maximum velocities than other sail designs in the literature. Broadly, our films mark a significant leap forward toward plausible relativistic interstellar propulsion for intragalactic exploration

physics.optics↗

arXiv2025

On connected graphs with finite spectral redundancy index and Pythagorean triplets

This article investigates spectral redundancy, a concept initially introduced by Alberto Seeger. Spectral redundancy arises when different connected induced subgraphs of a graph share the same spectral radius in their adjacency spectrum. Let $b(G)$ denote the total number of non-isomorphic induced subgraphs of $G$, and $c(G)$ represents the cardinality of the set of spectral radius of all connected induced subgraphs of $G$. The spectral redundancy of a graph $ G $ is defined as the ratio $ \frac{b(G)}{c(G)} $. The supremum of this ratio across all graphs in a family is called the spectral redundancy index of that family. We focus on a family of graphs that exhibit spectral redundancy and we find out the spectral redundancy index of this family. Furthermore, we investigate the connection between the spectral redundancy of these graphs and the presence of Pythagorean triplets.

math.CO↗

arXiv2025

Real-Time Agile Software Management for Edge and Fog Computing Based Smart City Infrastructure

The evolution of smart cities demands scalable, secure, and energy-efficient architectures for real-time data processing. With the number of IoT devices expected to exceed 40 billion by 2030, traditional cloud-based systems are increasingly constrained by bandwidth, latency, and energy limitations. This paper leverages the ROOF (Real-time Onsite Operations Facilitation) framework with decentralized computing at intermediary fog and peripheral edge network layers to reduce latency by processing data near its point of origin. ROOF features fog caching to avoid redundancy, ultra-low-power wireless transmission for energy savings, and AI-driven resource allocation for efficiency. Security is enhanced through TLS encryption, blockchain-based authentication, and edge-level access control. Case studies from Bhubaneswar, Barcelona and Copenhagen validate the use of ROOF in traffic systems and environmental monitoring. The paper concludes by outlining key challenges and prospects of AI-driven analytics in smart urban infrastructure.

cs.SE↗

arXiv2025

The Cosmic Evolution of Fast Radio Bursts Inferred from the CHIME/FRB Baseband Catalog 1

Redshift and luminosity distributions are essential for understanding the cosmic evolution of extragalactic objects and phenomena, such as galaxies, gamma-ray bursts, and fast radio bursts (FRBs). For FRBs, these distributions are primarily estimated using the fluence and the Dispersion Measure (DM). Calibrating their joint distribution has been challenging due to a lack of accurate fluences in the intensity data of the CHIME/FRB survey. Using the baseband update of CHIME/FRB Catalog 1, we calibrate the 2D fluence-DM distribution for the first time. We find the energy distribution is described well by a Schechter function with power-law slope of $-1.94^{+0.14}_{-0.12}$. Testing two types of redshift evolution models suggests a likely combination of young and old formation channels. $31^{+31}_{-21}$% of FRB sources may track star formation, or correspondingly, FRB sources may have delay times of $1.94^{+1.54}_{-1.31}$ Gyr. A pure star formation tracking population is excluded by only one model at $> 2σ$ confidence. An updated cosmic star formation rate density evolution up to redshift 14 is constrained by compiling results from several JWST studies. The furthest FRB detection with planned radio facilities is expected to be at $z \approx 5$. A radio telescope operating at 200 MHz with a system-equivalent flux density of $\leq 0.07$ Jy (equivalent to a detection threshold of 1 mJy ms) and instantaneous sky coverage of $\gtrsim 400$ square degrees should be able to detect $630^{+730}_{-485}$ FRBs year$^{-1}$ at $z \gtrsim 6$ and $53^{+83}_{-43}$ FRBs year$^{-1}$ at $z\gtrsim 8$, which is sufficient to differentiate between reionization histories.

astro-ph.HE↗

arXiv2025

Direct Joule-Heated Non-Equilibrium Synthesis Enables High Performing Thermoelectrics

High-throughput synthesis of bulk inorganic materials is crucial for accelerating functional materials discovery but is hindered by slow, energy-intensive solid-state methods. We introduce Direct Joule-Heated Synthesis (DJS), a rapid, single-step and scalable solid-state synthesis technique achieving a $10^5$-fold speedup and 20,000x energy efficiency improvement over conventional synthesis. DJS enables the synthesis of dense, bulk chalcogenides ($\mathrm{Bi_{0.5}Sb_{1.5}Te_3}$, $\mathrm{AgSbTe_2}$), achieving a zT of 2.3 at 573 K in optimally Cd/Se co-doped $\mathrm{AgSbTe_2}$, one of the highest for polycrystalline materials at this temperature. DJS enables optimal co-doping and rapid, non-equilibrium solidification, producing lamellar microstructures, interfacial regions, and cation-ordered nanodomains that scatter all-scale phonons, achieving ultralow lattice thermal conductivity (~0.2 $W m^{-1} K^{-1}$ at 573 K). DJS establishes a new benchmark for scalable and fast synthesis, accelerating functional material discovery.

cond-mat.mtrl-sci↗

arXiv2025

Review of Extreme Multilabel Classification

Extreme multi-label classification or XMLC, is an active area of interest in machine learning. Compared to traditional multi-label classification, here the number of labels is extremely large, hence, the name extreme multi-label classification. Using classical one-versus-all classification does not scale in this case due to large number of labels; the same is true for any other classifier. Embedding labels and features into a lower-dimensional space is a common first step in many XMLC methods. Moreover, other issues include existence of head and tail labels, where tail labels are those that occur in a relatively small number of samples. The existence of tail labels creates issues during embedding. This area has invited application of wide range of approaches ranging from bit compression motivated from compressed sensing, tree based embeddings, deep learning based latent space embedding including using attention weights, linear algebra based embeddings such as SVD, clustering, hashing, to name a few. The community has come up with a useful set of metrics to identify correctly the prediction for head or tail labels.

cs.LG↗