arXiv subjects

Sanjay Kumar

Publications and source records attributed to Sanjay Kumar.

At least 19 recordsLinked to original sources

arXiv2026

Smart Railway Obstruction Detection System using IoT and Computer Vision

Railway track intrusions pose a critical safety challenge for Indian Railways, encompassing wildlife incursions and deliberate malicious obstructions. The December 2025 collision in Assam, in which seven elephants were killed by the Rajdhani Express, underscores the urgency of effective real-time detection. Existing solutions such as the optical fiber-based Gajraj system suffer from prohibitive costs (\$1000/km) and high false alarm rates, limiting deployment to only 20 of India's 101 elephant corridors. This paper proposes NETRA, a cost-effective, internet-independent intrusion detection system deployed on Raspberry Pi Zero W and Raspberry Pi 4 edge platforms. NETRA employs probabilistic sensor fusion integrating a PIR motion sensor and an HC-SR04 ultrasonic distance sensor with a tunable threshold (tau_c = 0.65), enabling event-driven camera activation that reduces unnecessary visual processing by 52%. Upon confirmed intrusion, edge-AI classification using MobileNet-SSD (Pi Zero) or YOLOv5 ONNX (Pi 4) identifies threats including humans, large animals, and track obstructions. Confirmed threats are transmitted via LoRa (868 MHz) to alert the locomotive driver within 2.4 seconds end-to-end. Experimental evaluation across 113 motion events demonstrated 95% detection accuracy with zero false alarms through probabilistic fusion, compared to 85% for binary methods. Raspberry Pi 4 with YOLOv5 achieved 83.5% elephant F1-score, a 5.6x improvement over Pi Zero's heuristic approach (14.8%). LoRa communication achieved 100% packet delivery across 1-2 km in field trials. NETRA reduces deployment cost by 75% (\$247/km vs \$1000/km for Gajraj) while providing unified detection of both wildlife and obstruction threats.

cs.CV↗

arXiv2026

Network exploration by random walks: A large deviation perspective

We study exploration properties of a random walk on a network. For a fully connected network we find that the problem can be mapped to the well known coupon collector problem, thus allowing us to estimate form of $P(S,t)$: the distribution of number of distinct nodes $S$ visited by the random walk upto time $t$. From a practical point of view, however, both the fully connected network and hops taking place after fixed intervals are an idealization. We solve this problem by introducing the formalism of continuous time random walks wherein the random walk spends a random amount of time a node before hopping to its neighboring node. The formalism allows us to study the large deviation limit of $P(S,t)$ under very mild conditions that the distribution of waiting times $\psi(\tau)$ exhibits analyticity at small times. Furthermore, we find that at small times, the properties of $P(S,t)$ are largely independent of the network topology, and are governed solely by the waiting time characteristics.

physics.soc-ph↗

arXiv2026

Magnetic Reconnection at Hyperbolic Flux Tube associated with a Confined Flare in NOAA Active Region 12268

In this paper, we identify the magnetic reconnections at the hyperbolic flux tube (HFT), aided by slipping reconnection at quasi-separatrix layers (QSLs), which are pivotal to the occurrence of a confined M2.1 class flare in NOAA active region 12268. The magnetic field topology before the flare's onset is obtained through a non-force-free-field extrapolation scheme that accommodates a non-zero Lorentz force. A key aspect is the presence of an HFT in the computational domain above the flaring region, along with two QSLs at the lower boundary. To simulate the dynamics of the active region, we conduct a data-constrained magnetohydrodynamics (MHD) simulation initiated by the extrapolated field. The dynamics captured in the simulation document the formation of a current sheet within the HFT configuration, leading to magnetic reconnection at the HFT. Additionally, we observe the slipping motion of the footpoints of the magnetic field lines in the QSLs at the bottom boundary, which indicates the occurrence of slipping reconnection in the QSLs. Importantly, the magnetic reconnection at the HFT is suggested to be the primary driver in the development of the intricate flare brightenings and the flare ribbons.

astro-ph.SR↗

arXiv2026

FLUX: Data Worth Training On

Modern large language model training is no longer limited by data availability, but by the inability of existing preprocessing pipelines to simultaneously achieve massive scale and high data quality. Current approaches are forced to sacrifice one for the other: either aggressively filtering to improve quality at the cost of severe token loss, or retaining large volumes of data while introducing substantial noise. In this work, we introduce FLUX, a preprocessing pipeline specifically designed to break this long-standing trade-off by maximizing token retention while enforcing rigorous quality control. Models trained on FLUX-curated data consistently outperform prior methods. A 3B-parameter model trained on 60B tokens with FLUX achieves 32.14% MMLU accuracy, surpassing the previous state-of-the-art pipeline DCLM (31.98%) and significantly outperforming FineWeb (29.88%). FLUX achieves the same aggregate score as a model trained on DCLM data using only 39B tokens, resulting in a 34.4% reduction in training compute. At the data level, FLUX extracts 50B usable tokens from a single dump (CC-MAIN-2025-51), compared to 40B from DCLM (+25% retention). FLUX-Base yields 192B tokens, exceeding FineWeb's 170B while still maintaining superior quality. Overall, FLUX establishes a new state of the art in web-scale data preprocessing by demonstrating that high retention, strong quality control, and computational efficiency can be achieved simultaneously, redefining the limits of scalable dataset construction for modern language models.

cs.CL↗

arXiv2026

Implementation and Performance Evaluation of CMOS-integrated Memristor-driven Flip-flop Circuits

In this work, we report implementation and performance evaluation of memristor-driven fundamental logic gates, including NOT, AND, NAND, OR, NOR, and XOR, and novel and optimized design of the sequential logic circuits, such as D flip-flop, T-flip-flop, JK-flip-flop, and SR-flip-flop. The design, implementation, and optimization of these logic circuits were performed in SPECTRE in Cadence Virtuoso and integrated with 90 nm CMOS technology node. Additionally, we discuss an optimized design of memristor-driven logic gates and sequential logic circuits, and draw a comparative analysis with the other reported state-of-the-art work on sequential circuits. Moreover, the utilized memristor framework was experimentally pre-validated with the experimental data of Y2O3-based memristive devices, which shows significantly low values of variability during switching in both device-to-device (D2D) and cycle-to-cycle (C2C) operation. The performance metrics were calculated in terms of area, power, and delay of these sequential circuits and were found to be reduced by more than ~24%, 60%, and 58%, respectively, as compared to the other state-of-the-art work on sequential circuits. Therefore, the implemented memristor-based design significantly improves the performance of various logic designs, which makes it more area and power-efficient and shows the potential of memristor in designing various low-power, low-cost, ultrafast, and compact circuits.

cs.AR↗

arXiv2025

Blu-WERP (Web Extraction and Refinement Pipeline): A Scalable Pipeline for Preprocessing Large Language Model Datasets

High-quality training data is fundamental to large language model (LLM) performance, yet existing preprocessing pipelines often struggle to effectively remove noise and unstructured content from web-scale corpora. This paper presents Blu-WERP, a novel data preprocessing pipeline designed to optimize the quality of Common Crawl WARC files for LLM training. We demonstrate that Blu-WERP significantly outperforms established baselines including DCLM across multiple model scales and evaluation benchmarks. Our pipeline processes CC WARC dumps, implementing advanced filtering and quality assessment mechanisms. We conducted comprehensive evaluations using models with 150M, 400M, 530M, 750M, and 1B parameters, testing against nine standard benchmarks categorized as World Knowledge & Reasoning, Language Understanding, and Commonsense Reasoning. Results show Blu-WERP consistently achieved superior performance across all model scales. At the 1B parameter scale, Relatively Blu-WERP demonstrates a 4.0% and 9.5% aggregate improvement over DCLM and Fineweb respectively, while achieving quality-per-token efficiency gain. Categorical analysis reveals 2.4% improvement in World Knowledge & Reasoning, 6.2% improvement in Language Understanding, and 4.2% improvement in Commonsense Reasoning. These results establish Blu-WERP as a state-of-the-art preprocessing pipeline that substantially improves LLM training data quality and downstream model performance with reduced computational cost. Our findings contribute to the growing body of research on data-centric AI, demonstrating that preprocessing pipeline design significantly impacts LLM capabilities. The Blu-WERP pipeline represents a practical advancement in data quality optimization, offering researchers and practitioners an effective solution for improving LLM training efficiency and model performance.

cs.CL↗

arXiv2025

Evaluating the Impact of Weather-Induced Sensor Occlusion on BEVFusion for 3D Object Detection

Accurate 3D object detection is essential for automated vehicles to navigate safely in complex real-world environments. Bird's Eye View (BEV) representations, which project multi-sensor data into a top-down spatial format, have emerged as a powerful approach for robust perception. Although BEV-based fusion architectures have demonstrated strong performance through multimodal integration, the effects of sensor occlusions, caused by environmental conditions such as fog, haze, or physical obstructions, on 3D detection accuracy remain underexplored. In this work, we investigate the impact of occlusions on both camera and Light Detection and Ranging (LiDAR) outputs using the BEVFusion architecture, evaluated on the nuScenes dataset. Detection performance is measured using mean Average Precision (mAP) and the nuScenes Detection Score (NDS). Our results show that moderate camera occlusions lead to a 41.3% drop in mAP (from 35.6% to 20.9%) when detection is based only on the camera. On the other hand, LiDAR sharply drops in performance only under heavy occlusion, with mAP falling by 47.3% (from 64.7% to 34.1%), with a severe impact on long-range detection. In fused settings, the effect depends on which sensor is occluded: occluding the camera leads to a minor 4.1% drop (from 68.5% to 65.7%), while occluding LiDAR results in a larger 26.8% drop (to 50.1%), revealing the model's stronger reliance on LiDAR for the task of 3D object detection. Our results highlight the need for future research into occlusion-aware evaluation methods and improved sensor fusion techniques that can maintain detection accuracy in the presence of partial sensor failure or degradation due to adverse environmental conditions.

cs.CV↗

arXiv2025

Occluded nuScenes: A Multi-Sensor Dataset for Evaluating Perception Robustness in Automated Driving

Robust perception in automated driving requires reliable performance under adverse conditions, where sensors may be affected by partial failures or environmental occlusions. Although existing autonomous driving datasets inherently contain sensor noise and environmental variability, very few enable controlled, parameterised, and reproducible degradations across multiple sensing modalities. This gap limits the ability to systematically evaluate how perception and fusion architectures perform under well-defined adverse conditions. To address this limitation, we introduce the Occluded nuScenes Dataset, a novel extension of the widely used nuScenes benchmark. For the camera modality, we release both the full and mini versions with four types of occlusions, two adapted from public implementations and two newly designed. For radar and LiDAR, we provide parameterised occlusion scripts that implement three types of degradations each, enabling flexible and repeatable generation of occluded data. This resource supports consistent, reproducible evaluation of perception models under partial sensor failures and environmental interference. By releasing the first multi-sensor occlusion dataset with controlled and reproducible degradations, we aim to advance research on robust sensor fusion, resilience analysis, and safety-critical perception in automated driving.

cs.CV↗

arXiv2025

Bunching of extreme events on complex network

Extreme events such as earthquakes, floods, and power blackouts often display burst phenomena where multiple extreme events occur in quick succession or in bunches. This study examines bunching of extreme events on a complex network using a random walk transport model. We find that in a modular network, a small cluster sparsely connected with the rest, shows bunching and correlations among extreme events. The bunching and correlations emerge naturally in our system. We use several characterization techniques, namely the recurrence time distribution, autocorrelation function, bursty trains, burstiness parameter and memory coefficient to quantify the bunching and correlations of extreme events. Our study shows that the network structure plays a significant role in the bunching of extreme events.

physics.soc-ph↗

arXiv2025

On the Magnetic Reconnection and its Properties during a Flare Using a Magnetohydrodynamics Simulation

We study the magnetic reconnection during a flare by investigating flare ribbon dynamics using observations and data-constrained magnetohydrodynamics (MHD) simulation. In particular, we estimate the reconnection flux and the reconnection flux rates using flare ribbons of an M1.1 flare hosted by the active region 12184 utilizing the technique developed by Qiu et al. (2002, ApJ, 565, 1335). The reconnection flux and corresponding flux rates are found to be 10^20 Mx and 10^18 Mx/s respectively. To understand the flare onset and the origin of flare ribbons, we perform an MHD simulation initiated by the non-force-free-field extrapolation. Importantly, the extrapolated configuration identifies a three-dimensional (3D) magnetic neutral point and a flux rope in the flaring region, which is crucial to the flaring activity. The reconnection initiates at the null point and, subsequently the flux rope rises and appears to reconnect there, which is favorable for the eruption of the filament. The surrounding field lines also seem to take part in the null point reconnection. In later stage, a current sheet is formed below the null point ensuing a secondary reconnection near an X-type topology, further contributing to the energy release process in the flare. We trace the footpoint evolution of the field lines lying over the flare ribbons and find a significant similarity between the observed flare ribbons and the evolution of footpoints computed from the MHD simulation. We estimated induced electric field during the flare and found it to be $\approx$ .52 V/cm, a slight less value, as per many past literatures. Additional findings are the enhancement of vertical current density near the flaring ribbons, a signature of successive reconnections near the null point. Overall, the present work contributes to the understanding of the ribbon formation in a flaring process and the involved magnetic reconnection.

astro-ph.SR↗

arXiv2025

GBSVR: Granular Ball Support Vector Regression

Support Vector Regression (SVR) and its variants are widely used to handle regression tasks, however, since their solution involves solving an expensive quadratic programming problem, it limits its application, especially when dealing with large datasets. Additionally, SVR uses an epsilon-insensitive loss function which is sensitive to outliers and therefore can adversely affect its performance. We propose Granular Ball Support Vector Regression (GBSVR) to tackle problem of regression by using granular ball concept. These balls are useful in simplifying complex data spaces for machine learning tasks, however, to the best of our knowledge, they have not been sufficiently explored for regression problems. Granular balls group the data points into balls based on their proximity and reduce the computational cost in SVR by replacing the large number of data points with far fewer granular balls. This work also suggests a discretization method for continuous-valued attributes to facilitate the construction of granular balls. The effectiveness of the proposed approach is evaluated on several benchmark datasets and it outperforms existing state-of-the-art approaches

cs.LG↗

arXiv2025

Data-Constrained Magnetohydrodynamics Simulation of a Confined X-class Flare in NOAA Active Region 11166

In this paper, we present a magnetohydrodynamics simulation of NOAA active region 11166 to understand the origin of a confined X-class flare that peaked at 23:23 UT on 2011 March 9. The simulation is initiated with a magnetic field extrapolated from the corresponding photospheric magnetogram using a non-force-free-field extrapolation technique. Importantly, the initial magnetic configuration identifies three-dimensional (3D) magnetic nulls and quasi-separatrix layers (QSLs), which nearly agree with the bright structures appeared in multi-wavelength observations. The Lorentz force associated with the extrapolated field self-consistently generates the dynamics that leads to the magnetic reconnections at the 3D nulls and the QSLs. These reconnections are found to contribute to the pre-flare activities and, ultimately, lead to the development of the flare ribbons. Notably, the anchored spine of the 3D null and the complete absence of flux rope in the flaring region are congruent with the confined nature of the flare. Furthermore, the simulation also suggests the role of reconnections at the 3D null with an open spine in the onset of a jet away from the flaring site.

astro-ph.SR↗

arXiv2025

Minimizing Occlusion Effect on Multi-View Camera Perception in BEV with Multi-Sensor Fusion

Autonomous driving technology is rapidly evolving, offering the potential for safer and more efficient transportation. However, the performance of these systems can be significantly compromised by the occlusion on sensors due to environmental factors like dirt, dust, rain, and fog. These occlusions severely affect vision-based tasks such as object detection, vehicle segmentation, and lane recognition. In this paper, we investigate the impact of various kinds of occlusions on camera sensor by projecting their effects from multi-view camera images of the nuScenes dataset into the Bird's-Eye View (BEV) domain. This approach allows us to analyze how occlusions spatially distribute and influence vehicle segmentation accuracy within the BEV domain. Despite significant advances in sensor technology and multi-sensor fusion, a gap remains in the existing literature regarding the specific effects of camera occlusions on BEV-based perception systems. To address this gap, we use a multi-sensor fusion technique that integrates LiDAR and radar sensor data to mitigate the performance degradation caused by occluded cameras. Our findings demonstrate that this approach significantly enhances the accuracy and robustness of vehicle segmentation tasks, leading to more reliable autonomous driving systems.

cs.CV↗

arXiv2024

Maximum Solar Energy Tracking Leverage High-DoF Robotics System with Deep Reinforcement Learning

Solar trajectory monitoring is a pivotal challenge in solar energy systems, underpinning applications such as autonomous energy harvesting and environmental sensing. A prevalent failure mode in sustained solar tracking arises when the predictive algorithm erroneously diverges from the solar locus, erroneously anchoring to extraneous celestial or terrestrial features. This phenomenon is attributable to an inadequate assimilation of solar-specific objectness attributes within the tracking paradigm. To mitigate this deficiency inherent in extant methodologies, we introduce an innovative objectness regularization framework that compels tracking points to remain confined within the delineated boundaries of the solar entity. By encapsulating solar objectness indicators during the training phase, our approach obviates the necessity for explicit solar mask computation during operational deployment. Furthermore, we leverage the high-DoF robot arm to integrate our method to improve its robustness and flexibility in different outdoor environments.

cs.RO↗

arXiv2024

A determinant formula of the Jones polynomial for a family of braids

In 2012, Cohen, Dasbach, and Russell presented an algorithm to construct a weighted adjacency matrix for a given knot diagram. In the case of pretzel knots, it is shown that after evaluation, the determinant of the matrix recovers the Jones polynomial. Although the Jones polynomial is known to be #P-hard by Jaeger, Vertigan, and Welsh, this presents a class of knots for which the Jones polynomial can be computed in polynomial time by using the determinant. In this paper, we extend these results by recovering the Jones polynomial as the determinant of a weighted adjacency matrix for certain subfamilies of the braid group. Lastly, we compute the Kauffman polynomial of (2,q) torus knots in polynomial time using the balanced overlaid Tait graphs. This is the first known example of generalizing the methodology of Cohen to a class of quantum invariants which cannot be derived from the HOMFLYPT polynomial.

math.GT↗

arXiv2024

Exploring the generation and annihilation of three dimensional nulls through MHD simulations in initially chaotic magnetic field devoid of nulls

Three-dimensional (3D) magnetic nulls are abundant in the solar atmosphere, as been firmly established through contemporary observations. They are established to be important magnetic structures in, for example, jets and circular ribbon flares. While simulations and extrapolations support this, the mechanisms behind 3D null generation remain an open question. Recent magnetohydrodynamics (MHD) simulations propose that magnetic reconnection is responsible for both generating and annihilating 3D nulls, a novel concept. However, these simulations began with initial magnetic fields already supporting pre-existing nulls, raising the question of whether magnetic reconnection can create nulls in fields initially devoid of them. Previously, this question was briefly explored in a simulation with an initial chaotic magnetic field. However, the study failed to precisely identify locations, topological degrees, and natures (spiral or radial) of nulls, and it approximated magnetic reconnection without fully tracking field line in time. In this paper these findings are revisited in light of recent advancements and tools used to locate and trace nulls, along with the tracing of field lines, through which the concept of generation/annihilation of 3D nulls from chaotic fields is established in a precise manner.

astro-ph.SR↗

arXiv2024

Generative Artificial Intelligence: A Systematic Review and Applications

In recent years, the study of artificial intelligence (AI) has undergone a paradigm shift. This has been propelled by the groundbreaking capabilities of generative models both in supervised and unsupervised learning scenarios. Generative AI has shown state-of-the-art performance in solving perplexing real-world conundrums in fields such as image translation, medical diagnostics, textual imagery fusion, natural language processing, and beyond. This paper documents the systematic review and analysis of recent advancements and techniques in Generative AI with a detailed discussion of their applications including application-specific models. Indeed, the major impact that generative AI has made to date, has been in language generation with the development of large language models, in the field of image translation and several other interdisciplinary applications of generative AI. Moreover, the primary contribution of this paper lies in its coherent synthesis of the latest advancements in these areas, seamlessly weaving together contemporary breakthroughs in the field. Particularly, how it shares an exploration of the future trajectory for generative AI. In conclusion, the paper ends with a discussion of Responsible AI principles, and the necessary ethical considerations for the sustainability and growth of these generative models.

cs.LG↗

arXiv2024

Polymer translocation: Effects of confinement

We investigate the influence of varying confinement on the dynamics of polymer translocation through a cone shaped channel. For this, a linear polymer chain is modeled by a self voiding walks (SAWs) on a square lattice. The cis-side of a cone-shaped channel has a finite volume, whereas the trans-side has a semi-infinite space. The confining environment is varied either by changing the position of the back wall while keeping the apex angle fixed or altering the apex angle while keeping the position of the back wall fixed. In both cases, the effective space {\phi}, which represents the number of monomers in a chain relative to the total number of accessible sites within the cone, is reduced due to the imposed confinement. Consequently, the translocation dynamics are affected. We analyze the entropy of the confined system as a function {\phi}, which exhibits non-monotonic behaviour. We also calculate the free energy associated with the confinement as a function of a virtual coordinate for different positions of the back wall (base of the cone) along the conical axis for various apex angles.Employing the Fokker-Planck equation, we calculate the translocation time as a function of {\phi} for different solvent conditions across the pore. Our findings indicate that the translocation time decreases as {\phi} increases, but it eventually reaches a saturation point at a certain value of {\phi}. Moreover, we highlight the possibility to control the translocation dynamics by manipulating the solvent quality across the pore. Furthermore, our investigation delves into the intricacies of polymer translocation through a cone-shaped channel, considering both repulsive and neutral interactions with the pore wall. This exploration unveils nuanced dynamics and sheds light on the factors that significantly impact translocation within confined channels.

cond-mat.soft↗