arXiv subjects

Amit Agarwal

Publications and source records attributed to Amit Agarwal.

At least 73 records · Page 4Linked to original sources

arXiv2025

Detecting Lifshitz Transitions Using Nonlinear Conductivity in Bilayer Graphene

The second-order nonlinear electrical response (NLER) is an intrinsic property of inversion symmetry-broken systems which can provide deep insights into the electronic band structures of atomically thin quantum materials. However, the impact of Fermi surface reconstructions, also known as Lifshitz transitions, on the NLER has remained elusive. We investigated NLER in bilayer graphene (BLG), where the low-energy bands undergo Lifshitz transitions. Here, NLER undergoes a sign change near the Lifshitz transitions even at elevated temperatures $T\gtrsim10~$K. At the band edge, NLER in BLG is modulated by both extrinsic scattering and interfacial-strain-induced intrinsic Berry curvature dipole, both of which can be finely tuned externally by varying doping and interlayer potential. Away from the band edge, BLG exhibits second-order conductivity exceeding $30~μ$mV$^{-1}Ω^{-1}$ at 3K higher than any previous report. Our work establishes NLER as a reliable tool to probe Lifshitz transitions in quantum materials.

cond-mat.mes-hall↗

arXiv2025

Spin-split magnon bands induce pure spin current in insulating altermagnets

Altermagnets offer a promising platform for dissipationless spin transport by combining zero net magnetization with spontaneous non-relativistic spin splitting. However, their magnonic transport properties remain largely unexplored. Here, we develop a quantum-kinetic theory for thermally driven magnon currents that cleanly separates Berry-curvature-driven intrinsic contributions from Drude-like scattering-dependent terms. Applying this framework to a collinear honeycomb antiferromagnet with anisotropic next-nearest-neighbor exchange and Dzyaloshinskii-Moriya interaction, we reveal spin-split magnon bands that support both intrinsic and extrinsic spin Nernst and Seebeck currents. For realistic parameters, we predict a sizable magnon spin-splitting angle (about 3.3 degrees) and a pure transverse spin current capable of exerting a strong spin-splitter torque suitable for magnetization switching.

cond-mat.mes-hall↗

arXiv2025

Farm-Level, In-Season Crop Identification for India

Accurate, timely, and farm-level crop type information is paramount for national food security, agricultural policy formulation, and economic planning, particularly in agriculturally significant nations like India. While remote sensing and machine learning have become vital tools for crop monitoring, existing approaches often grapple with challenges such as limited geographical scalability, restricted crop type coverage, the complexities of mixed-pixel and heterogeneous landscapes, and crucially, the robust in-season identification essential for proactive decision-making. We present a framework designed to address the critical data gaps for targeted data driven decision making which generates farm-level, in-season, multi-crop identification at national scale (India) using deep learning. Our methodology leverages the strengths of Sentinel-1 and Sentinel-2 satellite imagery, integrated with national-scale farm boundary data. The model successfully identifies 12 major crops (which collectively account for nearly 90% of India's total cultivated area showing an agreement with national crop census 2023-24 of 94% in winter, and 75% in monsoon season). Our approach incorporates an automated season detection algorithm, which estimates crop sowing and harvest periods. This allows for reliable crop identification as early as two months into the growing season and facilitates rigorous in-season performance evaluation. Furthermore, we have engineered a highly scalable inference pipeline, culminating in what is, to our knowledge, the first pan-India, in-season, farm-level crop type data product. The system's effectiveness and scalability are demonstrated through robust validation against national agricultural statistics, showcasing its potential to deliver actionable, data-driven insights for transformative agricultural monitoring and management across India.

cs.CV↗

arXiv2025

Hybrid AI for Responsive Multi-Turn Online Conversations with Novel Dynamic Routing and Feedback Adaptation

Retrieval-Augmented Generation (RAG) systems and large language model (LLM)-powered chatbots have significantly advanced conversational AI by combining generative capabilities with external knowledge retrieval. Despite their success, enterprise-scale deployments face critical challenges, including diverse user queries, high latency, hallucinations, and difficulty integrating frequently updated domain-specific knowledge. This paper introduces a novel hybrid framework that integrates RAG with intent-based canned responses, leveraging predefined high-confidence responses for efficiency while dynamically routing complex or ambiguous queries to the RAG pipeline. Our framework employs a dialogue context manager to ensure coherence in multi-turn interactions and incorporates a feedback loop to refine intents, dynamically adjust confidence thresholds, and expand response coverage over time. Experimental results demonstrate that the proposed framework achieves a balance of high accuracy (95\%) and low latency (180ms), outperforming RAG and intent-based systems across diverse query types, positioning it as a scalable and adaptive solution for enterprise conversational AI applications.

cs.AI↗

arXiv2025

Nonlinear bulk photocurrent probe Z2 topological phase transition

Detecting topological phase transitions in bulk is challenging due to the limitations of surface sensitive probes like ARPES. Here, we demonstrate that nonlinear bulk photocurrents, specifically shift and injection currents, serve as effective probes of Z_2 topological transitions. These photocurrents show a robust polarity reversal across the Z_2 phase transition, offering a direct optical signature that distinguishes strong topological phases from weak or trivial ones. This effect originates from a reorganization of key band geometric quantities, the Berry curvature and shift vector, on time-reversal-invariant momentum planes. Using a low energy Dirac model, we trace this behaviour to a band inversion in the time-reversal-invariant momentum plane that drives the topological transition. We validate these findings through tight-binding model for Bi_2Te_3 and first-principles calculations for ZrTe_5 and BiTeI, where the topological phase can be tuned by pressure or temperature. Our results establish nonlinear photocurrent as a sensitive and broadly applicable probe of Z_2 topological phase transitions.

cond-mat.mtrl-sci↗

arXiv2025

MVTamperBench: Evaluating Robustness of Vision-Language Models

Multimodal Large Language Models (MLLMs), are recent advancement of Vision-Language Models (VLMs) that have driven major advances in video understanding. However, their vulnerability to adversarial tampering and manipulations remains underexplored. To address this gap, we introduce \textbf{MVTamperBench}, a benchmark that systematically evaluates MLLM robustness against five prevalent tampering techniques: rotation, masking, substitution, repetition, and dropping; based on real-world visual tampering scenarios such as surveillance interference, social media content edits, and misinformation injection. MVTamperBench comprises ~3.4K original videos, expanded into over ~17K tampered clips covering 19 distinct video manipulation tasks. This benchmark challenges models to detect manipulations in spatial and temporal coherence. We evaluate 45 recent MLLMs from 15+ model families. We reveal substantial variability in resilience across tampering types and show that larger parameter counts do not necessarily guarantee robustness. MVTamperBench sets a new benchmark for developing tamper-resilient MLLM in safety-critical applications, including detecting clickbait, preventing harmful content distribution, and enforcing policies on media platforms. We release all code, data, and benchmark to foster open research in trustworthy video understanding. Code: https://amitbcp.github.io/MVTamperBench/ Data: https://huggingface.co/datasets/Srikant86/MVTamperBench

cs.CV↗

arXiv2025

Hard Negative Mining for Domain-Specific Retrieval in Enterprise Systems

Enterprise search systems often struggle to retrieve accurate, domain-specific information due to semantic mismatches and overlapping terminologies. These issues can degrade the performance of downstream applications such as knowledge management, customer support, and retrieval-augmented generation agents. To address this challenge, we propose a scalable hard-negative mining framework tailored specifically for domain-specific enterprise data. Our approach dynamically selects semantically challenging but contextually irrelevant documents to enhance deployed re-ranking models. Our method integrates diverse embedding models, performs dimensionality reduction, and uniquely selects hard negatives, ensuring computational efficiency and semantic precision. Evaluation on our proprietary enterprise corpus (cloud services domain) demonstrates substantial improvements of 15\% in MRR@3 and 19\% in MRR@10 compared to state-of-the-art baselines and other negative sampling techniques. Further validation on public domain-specific datasets (FiQA, Climate Fever, TechQA) confirms our method's generalizability and readiness for real-world applications.

cs.IR↗

arXiv2025

SweEval: Do LLMs Really Swear? A Safety Benchmark for Testing Limits for Enterprise Use

Enterprise customers are increasingly adopting Large Language Models (LLMs) for critical communication tasks, such as drafting emails, crafting sales pitches, and composing casual messages. Deploying such models across different regions requires them to understand diverse cultural and linguistic contexts and generate safe and respectful responses. For enterprise applications, it is crucial to mitigate reputational risks, maintain trust, and ensure compliance by effectively identifying and handling unsafe or offensive language. To address this, we introduce SweEval, a benchmark simulating real-world scenarios with variations in tone (positive or negative) and context (formal or informal). The prompts explicitly instruct the model to include specific swear words while completing the task. This benchmark evaluates whether LLMs comply with or resist such inappropriate instructions and assesses their alignment with ethical frameworks, cultural nuances, and language comprehension capabilities. In order to advance research in building ethically aligned AI systems for enterprise use and beyond, we release the dataset and code: https://github.com/amitbcp/multilingual_profanity.

cs.CL↗

arXiv2025

Strain-Induced Charge Density Waves with Emergent Topological States in Monolayer NbSe2

Emergence of topological states in strongly correlated systems, particularly two-dimensional (2D) transition-metal dichalcogenides, offers a platform for manipulating electronic properties in quantum materials. However, a comprehensive understanding of the intricate interplay between correlations and topology remains elusive. Here we employ first-principles modeling to reveal two distinct 2x2 charge density wave (CDW) phases in monolayer 1H-NbSe2, which become energetically favorable over the conventional 3x3 CDWs under modest biaxial tensile strain of about 1%. These strain-induced CDW phases coexist with numerous topological states characterized by Z2 topology, high mirror Chern numbers, topological nodal lines, and higher-order topological states, which we have verified rigorously by computing the topological indices and the presence of robust edge states and localized corner states. Remarkably, these topological properties emerge because of the CDW rather than a pre-existing topology in the pristine phase. These results elucidate the interplay between correlations, topology, and geometry in 2D materials and indicate that strain-induced correlation effects can be used to engineer topological states in materials with initially trivial topology. Our findings may be applied in electronics, spintronics, and other advanced quantum devices that require robust and tunable topological states.

cond-mat.mtrl-sci↗

arXiv2025

Planar Nernst effect from hidden band geometry in layered two-dimensional materials

The Nernst effect is a versatile phenomenon relevant for energy harvesting, magnetic sensing, probing band topology and charge-neutral excitations. The planar Nernst effect (PNE) generates an in-plane voltage transverse to an applied temperature gradient under an in-plane magnetic field. Conventional Berry curvature-induced PNE is absent in two-dimensional (2D) systems, as the out-of-plane Berry curvature does not couple to the in-plane electron velocity. We challenge this notion by demonstrating a distinct planar Nernst effect in quasi-2D materials (2DPNE). We show that the 2DPNE originates from previously overlooked planar components of Berry curvature and orbital magnetic moment, arising from inter-layer tunneling in multilayered 2D systems. We comprehensively analyze the band-geometric origin and crystalline symmetry constraints on 2DPNE responses. We illustrate its experimental feasibility in strained bilayer graphene. Our findings significantly expand the theoretical understanding of planar Nernst effects, providing a clear pathway for next-generation magnetic sensing and energy-harvesting applications.

cond-mat.mes-hall↗

arXiv2025

Clinical QA 2.0: Multi-Task Learning for Answer Extraction and Categorization

Clinical Question Answering (CQA) plays a crucial role in medical decision-making, enabling physicians to extract relevant information from Electronic Medical Records (EMRs). While transformer-based models such as BERT, BioBERT, and ClinicalBERT have demonstrated state-of-the-art performance in CQA, existing models lack the ability to categorize extracted answers, which is critical for structured retrieval, content filtering, and medical decision support. To address this limitation, we introduce a Multi-Task Learning (MTL) framework that jointly trains CQA models for both answer extraction and medical categorization. In addition to predicting answer spans, our model classifies responses into five standardized medical categories: Diagnosis, Medication, Symptoms, Procedure, and Lab Reports. This categorization enables more structured and interpretable outputs, making clinical QA models more useful in real-world healthcare settings. We evaluate our approach on emrQA, a large-scale dataset for medical question answering. Results show that MTL improves F1-score by 2.2% compared to standard fine-tuning, while achieving 90.7% accuracy in answer categorization. These findings suggest that MTL not only enhances CQA performance but also introduces an effective mechanism for categorization and structured medical information retrieval.

cs.CL↗

arXiv2025

Tokenization Matters: Improving Zero-Shot NER for Indic Languages

Tokenization is a critical component of Natural Language Processing (NLP), especially for low resource languages, where subword segmentation influences vocabulary structure and downstream task accuracy. Although Byte Pair Encoding (BPE) is a standard tokenization method in multilingual language models, its suitability for Named Entity Recognition (NER) in low resource Indic languages remains underexplored due to its limitations in handling morphological complexity. In this work, we systematically compare BPE, SentencePiece, and Character Level tokenization strategies using IndicBERT for NER tasks in low resource Indic languages like Assamese, Bengali, Marathi, and Odia, as well as extremely low resource Indic languages like Santali, Manipuri, and Sindhi. We assess both intrinsic linguistic properties tokenization efficiency, out of vocabulary (OOV) rates, and morphological preservation as well as extrinsic downstream performance, including fine tuning and zero shot cross lingual transfer. Our experiments show that SentencePiece is a consistently better performing approach than BPE for NER in low resource Indic Languages, particularly in zero shot cross lingual settings, as it better preserves entity consistency. While BPE provides the most compact tokenization form, it is not capable of generalization because it misclassifies or even fails to recognize entity labels when tested on unseen languages. In contrast, SentencePiece constitutes a better linguistic structural preservation model, benefiting extremely low resource and morphologically rich Indic languages, such as Santali and Manipuri, for superior entity recognition, as well as high generalization across scripts, such as Sindhi, written in Arabic. The results point to SentencePiece as the more effective tokenization strategy for NER within multilingual and low resource Indic NLP applications.

cs.CL↗

arXiv2025

Crowdsource, Crawl, or Generate? Creating SEA-VL, a Multicultural Vision-Language Dataset for Southeast Asia

Southeast Asia (SEA) is a region of extraordinary linguistic and cultural diversity, yet it remains significantly underrepresented in vision-language (VL) research. This often results in artificial intelligence (AI) models that fail to capture SEA cultural nuances. To fill this gap, we present SEA-VL, an open-source initiative dedicated to developing high-quality, culturally relevant data for SEA languages. By involving contributors from SEA countries, SEA-VL aims to ensure better cultural relevance and diversity, fostering greater inclusivity of underrepresented languages in VL research. Beyond crowdsourcing, our initiative goes one step further in the exploration of the automatic collection of culturally relevant images through crawling and image generation. First, we find that image crawling achieves approximately ~85% cultural relevance while being more cost- and time-efficient than crowdsourcing. Second, despite the substantial progress in generative vision models, synthetic images remain unreliable in accurately reflecting SEA cultures. The generated images often fail to reflect the nuanced traditions and cultural contexts of the region. Collectively, we gather 1.28M SEA culturally-relevant images, more than 50 times larger than other existing datasets. Through SEA-VL, we aim to bridge the representation gap in SEA, fostering the development of more inclusive AI systems that authentically represent diverse cultures across SEA.

cs.CV↗

arXiv2025

Quantum geometric moment encodes stacking order of moiré matter

Exploring the topological characteristics of electronic bands is essential in condensed matter physics. Moiré materials featuring flat bands provide a versatile platform for engineering band topology and correlation effects. In moiré materials that break either time-reversal symmetry or inversion symmetry or both, electronic bands exhibit Berry curvature hotspots. Different stacking orders in these materials result in varied Berry curvature distributions within the flat bands, even when the band dispersion remains similar. However, experimental studies probing the impact of stacking order on the quantum geometric quantities are lacking. 1.4$^\circ$ twisted double bilayer graphene (TDBG) facilitates two distinct stacking orders (AB-AB, AB-BA) and forms an inversion broken \moire superlattice with electrically tunable flat bands. The valley Chern numbers of the flat bands depend on the stacking order, and the nonlinear Hall (NLH) effect distinguishes the differences in Berry curvature dipole (BCD), the first moment of Berry curvature. The BCD exhibits antisymmetric behavior, flipping its sign with the polarity of the perpendicular electric field in AB-AB TDBG, while it displays a symmetric behavior, maintaining the same sign regardless of the electric field's polarity in AB-BA TDBG. This approach electronically detects stacking-induced quantum geometry, while opening a pathway to quantum geometry engineering and detection.

cond-mat.mes-hall↗

arXiv2025

Symmetry-driven Intrinsic Nonlinear Pure Spin Hall Effect

The generation of pure spin current, spin angular momentum transport without charge flow, is crucial for developing energy-efficient spintronic devices with minimal Joule heating. Here, we introduce the intrinsic nonlinear pure spin Hall effect (NPSHE), where both linear and second-order charge Hall currents vanish. We show intrinsic second-order spin angular momentum transport in metals and insulators through a detailed analysis of the quantum geometric origin of different spin current contributions. Our comprehensive symmetry analysis identifies 39 magnetic point groups that support NPSHE, providing a foundation for material design and experimental realization. We predict significant nonlinear pure spin Hall currents in Kramers-Weyl metals even at room temperature, positioning them as potential candidates for NPSHE-based spin-torque devices. Our work lays a practical pathway for realizing charge-free angular momentum transport for the development of next-generation, energy-efficient spintronic devices.

cond-mat.mes-hall↗

arXiv2025

Third-order rectification in centrosymmetric metals

Rectification, the conversion of AC fields into DC currents, is crucial for optoelectronic applications such as energy harvesting and wireless communication. However, it is conventionally absent in centrosymmetric systems due to vanishing second-order optical responses. Here, we demonstrate significant rectification and photogalvanic currents in centrosymmetric metals via third-order nonlinear optical responses, driven by finite Fermi surface and disorder-induced contributions. We unveil distinct band geometric mechanisms -- including Berry curvature quadrupole, Fermi surface injection, and shift effects -- and classify all symmetry-allowed rectification responses. Using graphene as an example, we illustrate rectification tunability via light polarization and helicity, enabling rectification engineering in centrosymmetric materials for energy-efficient photodetection and terahertz applications.

cond-mat.mes-hall↗

arXiv2024

LLM for Barcodes: Generating Diverse Synthetic Data for Identity Documents

Accurate barcode detection and decoding in Identity documents is crucial for applications like security, healthcare, and education, where reliable data extraction and verification are essential. However, building robust detection models is challenging due to the lack of diverse, realistic datasets an issue often tied to privacy concerns and the wide variety of document formats. Traditional tools like Faker rely on predefined templates, making them less effective for capturing the complexity of real-world identity documents. In this paper, we introduce a new approach to synthetic data generation that uses LLMs to create contextually rich and realistic data without relying on predefined field. Using the vast knowledge LLMs have about different documents and content, our method creates data that reflects the variety found in real identity documents. This data is then encoded into barcode and overlayed on templates for documents such as Driver's licenses, Insurance cards, Student IDs. Our approach simplifies the process of dataset creation, eliminating the need for extensive domain knowledge or predefined fields. Compared to traditional methods like Faker, data generated by LLM demonstrates greater diversity and contextual relevance, leading to improved performance in barcode detection models. This scalable, privacy-first solution is a big step forward in advancing machine learning for automated document processing and identity verification.

cs.CL↗

arXiv2024

Survey of Large Multimodal Model Datasets, Application Categories and Taxonomy

Multimodal learning, a rapidly evolving field in artificial intelligence, seeks to construct more versatile and robust systems by integrating and analyzing diverse types of data, including text, images, audio, and video. Inspired by the human ability to assimilate information through many senses, this method enables applications such as text-to-video conversion, visual question answering, and image captioning. Recent developments in datasets that support multimodal language models (MLLMs) are highlighted in this overview. Large-scale multimodal datasets are essential because they allow for thorough testing and training of these models. With an emphasis on their contributions to the discipline, the study examines a variety of datasets, including those for training, domain-specific tasks, and real-world applications. It also emphasizes how crucial benchmark datasets are for assessing models' performance in a range of scenarios, scalability, and applicability. Since multimodal learning is always changing, overcoming these obstacles will help AI research and applications reach new heights.

cs.AI↗