arXiv subjects

Ashutosh Kumar

Publications and source records attributed to Ashutosh Kumar.

At least 37 records · Page 2Linked to original sources

arXiv2025

Two and three-state quantum heat engines with stochastic resetting

Quantum heat engines have undergone extensive studies over the last two decades. Simultaneously, the studies of the applications of stochastic resetting in various fields are on the rise. We explore the effect of stochastic resetting on the dynamics of a two-level and a three-level quantum heat engine. The extracted work is shown to increase with the resetting rate. The effective efficiency that takes into account the work done due to resetting remains constant. However, if the work done due to resetting is ignored, then the system can incorrectly imply a different behaviour, including the false inference that it is not working as an engine at all. The efficient power is observed to increase beyond that obtained in the absence of resetting, and is shown to be higher for a three-level engine.

cond-mat.stat-mech↗

arXiv2025

Tailored Thermal and Mechanical Performance of Biodegradable PLA-P(VDF-TrFE) Polymer Blends

The development of polymer blends has emerged as a strategic approach for designing multifunctional materials with enhanced tailored characteristics. Current work investigates and reports for the first time, the structure-property relationships in free-standing blend films of poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)) and polylactic acid (PLA), prepared to evaluate their suitability for functional applications. For this investigation, films of approximately 40 $μ$m thick were fabricated by systematically varying the P(VDF-TrFE):PLA ratio. Thermal analysis revealed a higher PLA crystallinity at 25\% P(VDF-TrFE) content, while Fourier-transform infrared spectroscopy showed the electroactive $β$-phase fraction to be highest in the 50:50 composition. These findings correlated with tensile strength measurements and morphology, demonstrating that molecular ordering and phase distribution significantly influence the mechanical performance. The 25:75 blend exhibited superior mechanical strength due to enhanced PLA crystallization and polymer chain alignment. In contrast, the 50:50 blend achieved a balance between tensile modulus and electroactive phase development, marking it a promising candidate for sensors and 3D printing applications. At higher P(VDF-TrFE) content, reduced crystallinity in PLA resulted in softer, more compliant films which would be suitable for flexible electronic applications. These results establish a pathway to tune mechanical and functional properties in semicrystalline polymer blends through facile compositional control.

cond-mat.mtrl-sci↗

arXiv2025

Hierarchical Prompting Taxonomy: A Universal Evaluation Framework for Large Language Models Aligned with Human Cognitive Principles

Assessing the effectiveness of large language models (LLMs) in performing different tasks is crucial for understanding their strengths and weaknesses. This paper presents Hierarchical Prompting Taxonomy (HPT), grounded on human cognitive principles and designed to assess LLMs by examining the cognitive demands of various tasks. The HPT utilizes the Hierarchical Prompting Framework (HPF), which structures five unique prompting strategies in a hierarchical order based on their cognitive requirement on LLMs when compared to human mental capabilities. It assesses the complexity of tasks with the Hierarchical Prompting Index (HPI), which demonstrates the cognitive competencies of LLMs across diverse datasets and offers insights into the cognitive demands that datasets place on different LLMs. This approach enables a comprehensive evaluation of an LLMs problem solving abilities and the intricacy of a dataset, offering a standardized metric for task complexity. Extensive experiments with multiple datasets and LLMs show that HPF enhances LLM performance by 2% to 63% compared to baseline performance, with GSM8k being the most cognitively complex task among reasoning and coding tasks with an average HPI of 3.20 confirming the effectiveness of HPT. To support future research and reproducibility in this domain, the implementations of HPT and HPF are available here.

cs.CL↗

arXiv2025

Structure-Property Correlations in Sb, Ge, and Ga Doped AlFe$_2$B$_2$ for Magnetocaloric Applications

This study investigates the effects of Sb, Ge, and Ga doping in AlFe$_2$B$_2$ on magnetic and magneto-caloric properties. Samples of AlFe$_2$B$_2$ and AlFe$_{1.9}$M$_{0.1}$B$_2$ (M= Ge, Ga and Sb) with 20\% excess Al were synthesized by arc melting, and the powder processed were investigated for their phase constituents, microstructure, magnetic and magneto-caloric effect. The parent compounds prepared showed the AlFe$_2$B$_2$ phase with a FeB secondary phase. However, in Sb and Ga-doped samples, an additional impurity phase, Al$_{13}$Fe$_4$, was observed apart from FeB, while in Ge-doped, only the AlB$_2$ impurity phase was present. The Curie temperature of AlFe$_2$B$_2$ is 277 K, increasing with Sb, Ge, and Ga doping to 287\,K, 297\,K, and 296\,K, respectively. The magnetization ($M$) is also higher with Ge and Ga addition in the 100-300\,K range; however, with Sb doping, the $M$ decreases significantly compared to parent AlFe$_2$B$_2$. The magnetic entropy change under 2\,T reached 2.93 JKg$^{-1}$K$^{-1}$ near 274\,K in AlFe$_2$B$_2$, which decreases to 2.53 JKg$^{-1}$K$^{-1}$ and 1.92 JKg$^{-1}$K$^{-1}$ with Ge and Ga, respectively. With Sb doping, the MC change was affected dramatically to 0.32 JKg$^{-1}$K$^{-1}$. However, the relative cooling power of Ge doped is the same as that of parent AlFe$_2$B$_2$. This research advances the understanding of the relationship between doping elements and magnetic properties in AlFe$_2$B$_2$ and opens pathways for designing magneto-caloric materials with tailored magnetic characteristics.

cond-mat.mtrl-sci↗

arXiv2025

Dynamics and Computational Principles of Echo State Networks: A Mathematical Perspective

Reservoir computing (RC) represents a class of state-space models (SSMs) characterized by a fixed state transition mechanism (the reservoir) and a flexible readout layer that maps from the state space. It is a paradigm of computational dynamical systems that harnesses the transient dynamics of high-dimensional state spaces for efficient processing of temporal data. Rooted in concepts from recurrent neural networks, RC achieves exceptional computational power by decoupling the training of the dynamic reservoir from the linear readout layer, thereby circumventing the complexities of gradient-based optimization. This work presents a systematic exploration of RC, addressing its foundational properties such as the echo state property, fading memory, and reservoir capacity through the lens of dynamical systems theory. We formalize the interplay between input signals and reservoir states, demonstrating the conditions under which reservoirs exhibit stability and expressive power. Further, we delve into the computational trade-offs and robustness characteristics of RC architectures, extending the discussion to their applications in signal processing, time-series prediction, and control systems. The analysis is complemented by theoretical insights into optimization, training methodologies, and scalability, highlighting open challenges and potential directions for advancing the theoretical underpinnings of RC.

cs.LG↗

arXiv2025

From Fog to Failure: The Unintended Consequences of Dehazing on Object Detection in Clear Images

This study explores the challenges of integrating human visual cue-based dehazing into object detection, given the selective nature of human perception. While human vision adapts dynamically to environmental conditions, computational dehazing does not always enhance detection uniformly. We propose a multi-stage framework where a lightweight detector identifies regions of interest (RoIs), which are then improved via spatial attention-based dehazing before final detection by a heavier model. Though effective in foggy conditions, this approach unexpectedly degrades the performance on clear images. We analyze this phenomenon, investigate possible causes, and offer insights for designing hybrid pipelines that balance enhancement and detection. Our findings highlight the need for selective preprocessing and challenge assumptions about universal benefits from cascading transformations.

cs.CV↗

arXiv2025

Perceptual Piercing: Human Visual Cue-based Object Detection in Low Visibility Conditions

This study proposes a novel deep learning framework inspired by atmospheric scattering and human visual cortex mechanisms to enhance object detection under poor visibility scenarios such as fog, smoke, and haze. These conditions pose significant challenges for object recognition, impacting various sectors, including autonomous driving, aviation management, and security systems. The objective is to enhance the precision and reliability of detection systems under adverse environmental conditions. The research investigates the integration of human-like visual cues, particularly focusing on selective attention and environmental adaptability, to ascertain their impact on object detection's computational efficiency and accuracy. This paper proposes a multi-tiered strategy that integrates an initial quick detection process, followed by targeted region-specific dehazing, and concludes with an in-depth detection phase. The approach is validated using the Foggy Cityscapes, RESIDE-beta (OTS and RTTS) datasets and is anticipated to set new performance standards in detection accuracy while significantly optimizing computational efficiency. The findings offer a viable solution for enhancing object detection in poor visibility and contribute to the broader understanding of integrating human visual principles into deep learning algorithms for intricate visual recognition challenges.

cs.CV↗

arXiv2025

Moonshine: Distilling Game Content Generators into Steerable Generative Models

Procedural Content Generation via Machine Learning (PCGML) has enhanced game content creation, yet challenges in controllability and limited training data persist. This study addresses these issues by distilling a constructive PCG algorithm into a controllable PCGML model. We first generate a large amount of content with a constructive algorithm and label it using a Large Language Model (LLM). We use these synthetic labels to condition two PCGML models for content-specific generation, a diffusion model and the five-dollar model. This neural network distillation process ensures that the generation aligns with the original algorithm while introducing controllability through plain text. We define this text-conditioned PCGML as a Text-to-game-Map (T2M) task, offering an alternative to prevalent text-to-image multi-modal tasks. We compare our distilled models with the baseline constructive algorithm. Our analysis of the variety, accuracy, and quality of our generation demonstrates the efficacy of distilling constructive methods into controllable text-conditioned PCGML models.

cs.AI↗

arXiv2025

IndicMMLU-Pro: Benchmarking Indic Large Language Models on Multi-Task Language Understanding

Known by more than 1.5 billion people in the Indian subcontinent, Indic languages present unique challenges and opportunities for natural language processing (NLP) research due to their rich cultural heritage, linguistic diversity, and complex structures. IndicMMLU-Pro is a comprehensive benchmark designed to evaluate Large Language Models (LLMs) across Indic languages, building upon the MMLU Pro (Massive Multitask Language Understanding) framework. Covering major languages such as Hindi, Bengali, Gujarati, Marathi, Kannada, Punjabi, Tamil, Telugu, and Urdu, our benchmark addresses the unique challenges and opportunities presented by the linguistic diversity of the Indian subcontinent. This benchmark encompasses a wide range of tasks in language comprehension, reasoning, and generation, meticulously crafted to capture the intricacies of Indian languages. IndicMMLU-Pro provides a standardized evaluation framework to push the research boundaries in Indic language AI, facilitating the development of more accurate, efficient, and culturally sensitive models. This paper outlines the benchmarks' design principles, task taxonomy, and data collection methodology, and presents baseline results from state-of-the-art multilingual models.

cs.CL↗

arXiv2024

Structural and Thermoelectric Properties of Al$_{1+x}$Fe$_{2-y}$M$_y$B$_2$ (M=Ag, Ni, Sb, Ga, and Ge) Intermetallic Borides

AlFe$_2$B$_2$, a ternary transition metal boride doped with Ag, Ni, Sb, Ga, and Ge was investigated for the phase constituents, microstructure, and thermoelectric properties. The parent compound with 20\% excess Al (Al$_{1.2}$Fe$_2$B$_2$) prepared by vacuum arc melting contains orthorhombic AlFe$_2$B$_2$ and FeB upon doping form additional phases such as Ag$_5$Al, AlNi$_3$, Al$_3$Ni$_2$, AlSb, and AlB$_2$. The structural parameters of the AlFe$_2$B$_2$ phase in Ag, Ni, and Sb-doped samples are nearly the same as the un-doped one. The lattice parameters and cell volume increase noticeably in Ga and Ge-doped compounds. The Ga-doped sample shows only FeB as a secondary phase; however, the AlB$_2$ phase is observed in Ge-doped AlFe$_2$B$_2$. The microstructure investigated in FEG-SEM with EDS shows the AlFe$_2$B$_2$ matrix is chemically homogenous in all samples, except the Ga-doped one, with uniformly distributed single or multiple secondary phases. The DSC-TG results show that all the dopants decrease the AlFe$_2$B$_2$ phase peritectic decomposition reaction temperature, indicating the structure destabilizes. The thermoelectric properties measured show that AlFe$_2$B$_2$ is an n-type compound with electrical conductivity in the range of 0.35-0.46$\times$10$^6$ S/m. The addition of Ag, Ni, Ga, and Ge marginally alters the Seebeck and electrical conductivity. The noticeable improvement in Seebeck coefficient and with negligible change in electrical conductivity resulted in a highest power factor of 0.4 mW/mK$^2$ for Sb-doped sample. The thermal conductivity of AlFe$_2$B$_2$, which is in the range of 6.4-9.1 W/m-K between 300 to 773\,K, decreases with Sb doping to 5.2-8.5 W/m-K resulting in a maximum zT of 0.04 at 773K.

cond-mat.mtrl-sci↗

arXiv2024

MoonMetaSync: Lunar Image Registration Analysis

This paper compares scale-invariant (SIFT) and scale-variant (ORB) feature detection methods, alongside our novel feature detector, IntFeat, specifically applied to lunar imagery. We evaluate these methods using low (128x128) and high-resolution (1024x1024) lunar image patches, providing insights into their performance across scales in challenging extraterrestrial environments. IntFeat combines high-level features from SIFT and low-level features from ORB into a single vector space for robust lunar image registration. We introduce SyncVision, a Python package that compares lunar images using various registration methods, including SIFT, ORB, and IntFeat. Our analysis includes upscaling low-resolution lunar images using bi-linear and bi-cubic interpolation, offering a unique perspective on registration effectiveness across scales and feature detectors in lunar landscapes. This research contributes to computer vision and planetary science by comparing feature detection methods for lunar imagery and introducing a versatile tool for lunar image registration and evaluation, with implications for multi-resolution image analysis in space exploration applications.

cs.CV↗

arXiv2024

WTS: A Pedestrian-Centric Traffic Video Dataset for Fine-grained Spatial-Temporal Understanding

In this paper, we address the challenge of fine-grained video event understanding in traffic scenarios, vital for autonomous driving and safety. Traditional datasets focus on driver or vehicle behavior, often neglecting pedestrian perspectives. To fill this gap, we introduce the WTS dataset, highlighting detailed behaviors of both vehicles and pedestrians across over 1.2k video events in hundreds of traffic scenarios. WTS integrates diverse perspectives from vehicle ego and fixed overhead cameras in a vehicle-infrastructure cooperative environment, enriched with comprehensive textual descriptions and unique 3D Gaze data for a synchronized 2D/3D view, focusing on pedestrian analysis. We also pro-vide annotations for 5k publicly sourced pedestrian-related traffic videos. Additionally, we introduce LLMScorer, an LLM-based evaluation metric to align inference captions with ground truth. Using WTS, we establish a benchmark for dense video-to-text tasks, exploring state-of-the-art Vision-Language Models with an instance-aware VideoLLM method as a baseline. WTS aims to advance fine-grained video event understanding, enhancing traffic safety and autonomous driving development.

cs.CV↗

arXiv2024

The Ethics of Interaction: Mitigating Security Threats in LLMs

This paper comprehensively explores the ethical challenges arising from security threats to Large Language Models (LLMs). These intricate digital repositories are increasingly integrated into our daily lives, making them prime targets for attacks that can compromise their training data and the confidentiality of their data sources. The paper delves into the nuanced ethical repercussions of such security threats on society and individual privacy. We scrutinize five major threats--prompt injection, jailbreaking, Personal Identifiable Information (PII) exposure, sexually explicit content, and hate-based content--going beyond mere identification to assess their critical ethical consequences and the urgency they create for robust defensive strategies. The escalating reliance on LLMs underscores the crucial need for ensuring these systems operate within the bounds of ethical norms, particularly as their misuse can lead to significant societal and individual harm. We propose conceptualizing and developing an evaluative tool tailored for LLMs, which would serve a dual purpose: guiding developers and designers in preemptive fortification of backend systems and scrutinizing the ethical dimensions of LLM chatbot responses during the testing phase. By comparing LLM responses with those expected from humans in a moral context, we aim to discern the degree to which AI behaviors align with the ethical values held by a broader society. Ultimately, this paper not only underscores the ethical troubles presented by LLMs; it also highlights a path toward cultivating trust in these systems.

cs.CR↗

arXiv2024

A Methodology-Oriented Study of Catastrophic Forgetting in Incremental Deep Neural Networks

Human being and different species of animals having the skills to gather, transferring knowledge, processing, fine-tune and generating information throughout their lifetime. The ability of learning throughout their lifespan is referred as continuous learning which is using neurocognition mechanism. Consequently, in real world computational system of incremental learning autonomous agents also needs such continuous learning mechanism which provide retrieval of information and long-term memory consolidation. However, the main challenge in artificial intelligence is that the incremental learning of the autonomous agent when new data confronted. In such scenarios, the main concern is catastrophic forgetting(CF), i.e., while learning the sequentially, neural network underfits the old data when it confronted with new data. To tackle this CF problem many numerous studied have been proposed, however it is very difficult to compare their performance due to dissimilarity in their evaluation mechanism. Here we focus on the comparison of all algorithms which are having similar type of evaluation mechanism. Here we are comparing three types of incremental learning methods: (1) Exemplar based methods, (2) Memory based methods, and (3) Network based method. In this survey paper, methodology oriented study for catastrophic forgetting in incremental deep neural network is addressed. Furthermore, it contains the mathematical overview of impact-full methods which can be help researchers to deal with CF.

cs.LG↗

arXiv2023

Thermoelectric properties of high-entropy wolframite oxide: (CoCuNiFeZn)$_{1-x}$Ga$_x$WO$_4$

In this report, the synthesis of high-entropy wolframite oxide (CoCuNiFeZn)$_{1-x}$Ga$_x$WO$_4$ through standard solid-state route followed by spark plasma sintering (SPS) and their structural, microstructural, and thermoelectric properties are investigated. X-ray diffraction pattern followed by pattern matching refinement shows monoclinic structure with volume of the unit cell decreasing with increasing Ga content. The optical band gap for these oxides shows a cocktail effect in high entropy configuration. The Seebeck coefficient indicates electrons as dominating charge carriers with a non-degenerate behavior. The electrical resistivity decreases with increasing temperature depicting a semiconducting nature. Thermal conductivity in high-entropy samples ($κ\sim$2.1 W/mK @ 300\,K) is significantly lower as compared to MgWO$_4$ ($κ\sim$11.5 W/mK @ 300\,K), which can be explained by the strong phonon scattering due to large lattice disorder in high entropy configuration. The thermoelectric figure of merit zT increases with Ga doping via modifying all three thermoelectric parameters positively.

cond-mat.mtrl-sci↗

arXiv2023

Thermodynamics of one and two-qubit quantum refrigerators interacting with squeezed baths: a comparative study

We investigate the nonequilibrium refrigeration of one and two-qubit systems in a squeezed thermal bath. We characterize the performance of one and two-qubit refrigerators in the presence of squeezed heat baths, in terms of their coefficients of performance, cooling rates, and figures of merit. Our results show that the performance of the refrigerators is strongly influenced by the squeezing parameter and the number of qubits. The performance of the two-qubit refrigerator is found to be better than that of the one-qubit refrigerator under the same operating conditions. Our findings suggest that a squeezed thermal bath can be a promising resource for the design of efficient quantum refrigerators in the non-equilibrium regime.

cond-mat.stat-mech↗

arXiv2023

Novel entropy-stabilized fluorite oxides with multifunctional properties

Development of new high-entropy oxides having configurational entropy dominating the phase stability has become a hot topic since the discovery of rock salt structure entropy-stabilized (ES)(MgCoNiCuZn)O in 2015. Herein, we report a set of novel entropy-stabilized fluorite oxides: Zr0.2Hf0.2Ce0.2Sn0.2Mn0.2O2-δ, Zr0.2Hf0.2Ti0.2Mn0.2Ce0.2O2-δ, Zr0.225Hf0.225Ti0.225Mn0.225Ce0.1O2-δ, and Zr0.2Hf0.2Ti0.2Mn0.2Ce0.1Ta0.05Fe0.05O2-δ synthesized using standard solid-state reaction. These compounds have been investigated using X-ray diffraction, scanning electron microscopy, and X-ray photoelectron spectroscopy techniques to discern their structural, microstructural, and chemical properties. The configurational-entropy dominated phase stability and hence the entropy stabilization of the compounds is confirmed by cyclic heat treatments. The mismatch in the ionic radii and oxidation state of the cations are the key factors in achieving a single-phase fluorite structure. Further, screening of physical properties including thermal conductivity, optical band gap, magnetic properties, and impedance spectroscopy is discussed. Thermal conductivity of 1.4-1.7 Wm-1K-1 is observed at 300 K and remains mostly invariant across a wide temperature range (300K-1073K), favorable for thermal barrier coating applications. These ES samples have an optical band gap of 1.6-1.8 eV, enabling light absorption across the visible spectrum and hence could be promising for photocatalytic applications. The impedance spectroscopy data of the entropy-stabilized samples reveal the presence of electronic contributions with small activation energy (0.3-0.4 eV) across a temperature range of 298K-423K. These observations in ES fluorite systems show potential for their multifunctional applications via further optimization and confirm the great chemical versatility of entropy-stabilized oxides.

cond-mat.mtrl-sci↗

arXiv2023

Quantum self-consistent equation-of-motion method for computing molecular excitation energies, ionization potentials, and electron affinities on a quantum computer

Near-term quantum computers are expected to facilitate material and chemical research through accurate molecular simulations. Several developments have already shown that accurate ground-state energies for small molecules can be evaluated on present-day quantum devices. Although electronically excited states play a vital role in chemical processes and applications, the search for a reliable and practical approach for routine excited-state calculations on near-term quantum devices is ongoing. Inspired by excited-state methods developed for the unitary coupled-cluster theory in quantum chemistry, we present an equation-of-motion-based method to compute excitation energies following the variational quantum eigensolver algorithm for ground-state calculations on a quantum computer. We perform numerical simulations on H$_2$, H$_4$, H$_2$O, and LiH molecules to test our quantum self-consistent equation-of-motion (q-sc-EOM) method and compare it to other current state-of-the-art methods. q-sc-EOM makes use of self-consistent operators to satisfy the vacuum annihilation condition, a critical property for accurate calculations. It provides real and size-intensive energy differences corresponding to vertical excitation energies, ionization potentials and electron affinities. We also find that q-sc-EOM is more suitable for implementation on NISQ devices as it is expected to be more resilient to noise compared with the currently available methods.

quant-ph↗