arXiv subjects

Shivam Sharma

Publications and source records attributed to Shivam Sharma.

At least 19 recordsLinked to original sources

arXiv2026

Do Thought Streams Matter? Evaluating Reasoning in Gemini Vision-Language Models for Video Scene Understanding

We benchmark how internal reasoning traces, which we call thought streams, affect video scene understanding in vision-language models. Using four configurations of Google's Gemini 2.5 Flash and Flash Lite across scenes extracted from 100 hours of video, we ask three questions: does more thinking lead to better outputs, where do the gains stop, and what do these models actually think about? We introduce three evaluation metrics. Contentfulness measures how much of the thought stream is useful scene content versus meta-commentary. Thought-Final Coverage measures how faithfully the thought stream translates into the final output. Dominant Entity Analysis identifies which subjects, actions, and settings the model focuses on. GPT-5 serves as an independent judge. We find that quality gains from additional thinking plateau quickly, with most improvement occurring in the first few hundred tokens. Flash Lite offers the best balance between quality and token usage. Tight reasoning budgets cause the model to add content in the final output that it never reasoned about, a form of compression-step hallucination. Despite being different model tiers, Flash and Flash Lite produce similar thought streams, though they differ in style: Flash discusses its reasoning process, while Lite focuses on describing the scene.

cs.CV↗

arXiv2025

PustakAI: Curriculum-Aligned and Interactive Textbooks Using Large Language Models

Large Language Models (LLMs) have demonstrated remarkable capabilities in understanding and generating human-like content. This has revolutionized various sectors such as healthcare, software development, and education. In education, LLMs offer potential for personalized and interactive learning experiences, especially in regions with limited teaching resources. However, adapting these models effectively to curriculum-specific content, such as the National Council of Educational Research and Training (NCERT) syllabus in India, presents unique challenges in terms of accuracy, alignment, and pedagogical relevance. In this paper, we present the framework "PustakAI"\footnote{Pustak means `book' in many Indian languages.} for the design and evaluation of a novel question-answering dataset "NCERT-QA" aligned with the NCERT curriculum for English and Science subjects of grades 6 to 8. We classify the curated QA pairs as Factoid, Inferential, and Others (evaluative and reasoning). We evaluate the dataset with various prompting techniques, such as meta-prompt, few-shot, and CoT-style prompting, using diverse evaluation metrics to understand which approach aligns more efficiently with the structure and demands of the curriculum. Along with the usability of the dataset, we analyze the strengths and limitations of current open-source LLMs (Gemma3:1b, Llama3.2:3b, and Nemotron-mini:4b) and high-end LLMs (Llama-4-Scout-17B and Deepseek-r1-70B) as AI-based learning tools in formal education systems.

cs.CL↗

arXiv2025

Toward Quantum Utility in Finance: A Robust Data-Driven Algorithm for Asset Clustering

Clustering financial assets based on return correlations is a fundamental task in portfolio optimization and statistical arbitrage. However, classical clustering methods often fall short when dealing with signed correlation structures, typically requiring lossy transformations and heuristic assumptions such as a fixed number of clusters. In this work, we apply the Graph-based Coalition Structure Generation algorithm (GCS-Q) to directly cluster signed, weighted graphs without relying on such transformations. GCS-Q formulates each partitioning step as a QUBO problem, enabling it to leverage quantum annealing for efficient exploration of exponentially large solution spaces. We validate our approach on both synthetic and real-world financial data, benchmarking against state-of-the-art classical algorithms such as SPONGE and k-Medoids. Our experiments demonstrate that GCS-Q consistently achieves higher clustering quality, as measured by Adjusted Rand Index and structural balance penalties, while dynamically determining the number of clusters. These results highlight the practical utility of near-term quantum computing for graph-based unsupervised learning in financial applications.

quant-ph↗

arXiv2025

Decoding Memes: Benchmarking Narrative Role Classification across Multilingual and Multimodal Models

This work investigates the challenging task of identifying narrative roles - Hero, Villain, Victim, and Other - in Internet memes, across three diverse test sets spanning English and code-mixed (English-Hindi) languages. Building on an annotated dataset originally skewed toward the 'Other' class, we explore a more balanced and linguistically diverse extension, originally introduced as part of the CLEF 2024 shared task. Comprehensive lexical and structural analyses highlight the nuanced, culture-specific, and context-rich language used in real memes, in contrast to synthetically curated hateful content, which exhibits explicit and repetitive lexical markers. To benchmark the role detection task, we evaluate a wide spectrum of models, including fine-tuned multilingual transformers, sentiment and abuse-aware classifiers, instruction-tuned LLMs, and multimodal vision-language models. Performance is assessed under zero-shot settings using precision, recall, and F1 metrics. While larger models like DeBERTa-v3 and Qwen2.5-VL demonstrate notable gains, results reveal consistent challenges in reliably identifying the 'Victim' class and generalising across cultural and code-mixed content. We also explore prompt design strategies to guide multimodal models and find that hybrid prompts incorporating structured instructions and role definitions offer marginal yet consistent improvements. Our findings underscore the importance of cultural grounding, prompt engineering, and multimodal reasoning in modelling subtle narrative framings in visual-textual content.

cs.CL↗

arXiv2025

Entity Framing and Role Portrayal in the News

We introduce a novel multilingual hierarchical corpus annotated for entity framing and role portrayal in news articles. The dataset uses a unique taxonomy inspired by storytelling elements, comprising 22 fine-grained roles, or archetypes, nested within three main categories: protagonist, antagonist, and innocent. Each archetype is carefully defined, capturing nuanced portrayals of entities such as guardian, martyr, and underdog for protagonists; tyrant, deceiver, and bigot for antagonists; and victim, scapegoat, and exploited for innocents. The dataset includes 1,378 recent news articles in five languages (Bulgarian, English, Hindi, European Portuguese, and Russian) focusing on two critical domains of global significance: the Ukraine-Russia War and Climate Change. Over 5,800 entity mentions have been annotated with role labels. This dataset serves as a valuable resource for research into role portrayal and has broader implications for news analysis. We describe the characteristics of the dataset and the annotation process, and we report evaluation results on fine-tuned state-of-the-art multilingual transformers and hierarchical zero-shot learning using LLMs at the level of a document, a paragraph, and a sentence.

cs.CL↗

arXiv2025

Nonlinear valley thermal physics in two dimensional materials

This study delves into the intrinsic nonlinear valley thermal effects, driven by the Quantum Metric of the system. Our findings elucidate that valley indices in the nonlinear effect are distinguishable by the thermoelectric correction to the orbital magnetization, which adopts opposite signs across valleys mirroring the role of orbital angular momentum in the linear valley Hall effect. Through a prototypical two-band model on an anisotropic tilted Dirac semimetal, we investigate how intrinsic material parameters modulate this nonlinear valley thermal response. Extending to realistic PT symmetric anisotropic semiconductors, our findings enrich the understanding of valley-based phenomena, with implications for advanced theoretical and experimental pursuits in valleytronics and valley caloritonics.

cond-mat.mes-hall↗

arXiv2025

Dirac Fermions and Flat Bands in Phosphorus Carbide Nanotubes: Structural and Quantum Phase Transitions in a Quasi-One-Dimensional Material

Chemically realistic quasi-one-dimensional (1D) materials in which Dirac fermions and highly degenerate flat bands coexist intrinsically at the Fermi level are exceedingly rare, while representing a highly desirable platform for correlated and topological quantum phenomena. Here, using specialized symmetry-adapted first-principles calculations we predict a new class of nanomaterials -- phosphorus carbide nanotubes ($\text{P}_2\text{C}_3$NTs) -- obtained by rolling monolayer $\text{P}_2\text{C}_3$, a two-dimensional material shown in a previous letter to host "double Kagome bands". Both armchair and zigzag $\text{P}_2\text{C}_3$NTs are stable at room temperature and feature the rare coexistence of Dirac crossings and multiple flat bands at the Fermi level inherited from the underlying honeycomb-Kagome lattice, with the flat bands resilient to elastic deformations. Under large strain, the structure transforms from honeycomb-Kagome to "brick-wall," accompanied by multiple coupled structural and quantum phase transitions. We also uncover localized edge states, spin splitting from vacancies and dopants, and strain-tunable magnetism. Together, these results establish $\text{P}_2\text{C}_3$NTs as a chemically specific and mechanically tunable 1D material platform with potential applications in quantum hardware and spintronics.

cond-mat.mtrl-sci↗

arXiv2025

Strain-Tunable Topological Phase Transitions in Line- and Split-Graph Flat-Band Lattices

In recent years, materials with topological flat bands have attracted significant attention due to their association with extraordinary transport properties and strongly correlated electrons. Yet, generic principles linking lattice architecture, strain, and band topology remain scarce. Here, using a unified graph-theoretic framework we generate entire families of two-dimensional lattices and, using analytical tight-binding calculations, demonstrate that a single mechanical knob -- uniform in-plane strain -- drives universal transitions between trivial insulating, Dirac semimetal, and quantum spin-Hall phases across all lattices. The framework yields several flat band lattices that were hitherto absent or largely unexplored in the literature -- for example, the checkerboard split-graph and triangular-Kagome lattices -- whose strain-driven topological phase diagrams we establish here for the first time. The design rules implied by our studies provide a blueprint for engineering topological states in a wide variety of 2D materials, photonic crystals, and circuit lattices, and are anticipated to accelerate the discovery of strain-programmable quantum matter.

cond-mat.str-el↗

arXiv2024

SAFE-MEME: Structured Reasoning Framework for Robust Hate Speech Detection in Memes

Memes act as cryptic tools for sharing sensitive ideas, often requiring contextual knowledge to interpret them correctly. It makes multimodal meme moderation difficult, as existing work either lacks high-quality datasets for nuanced hate categories or relies on low-quality social media visuals. Here, we curate two novel multimodal hate speech datasets comprising English memes - MHS and MHS-Con, which capture fine-grained hateful abstractions in regular and confounding scenarios, respectively. We benchmark these datasets against several competing baselines. Furthermore, we introduce SAFE-MEME (Structured reAsoning FramEwork) with its two variants: a novel multimodal Chain-of-Thought based framework employing Q&A-style reasoning (SAFE-MEME-QA) and a hierarchical categorization (SAFE-MEME-H) to enable robust hate speech detection in memes. SAFE-MEME-QA outperforms the strongest open-source baseline model, showing an improvement of 2% on MHS and 4.7% on MHS-Con and closely follows the closed-source models, GPT-4o and Gemini 2.5. In contrast, SAFE-MEME-H surpasses SAFE-MEME-QA, marking a 3% improvement over the best open-source baseline, while achieving performance comparable to GPT-4o only on MHS. We show that fine-tuning a single-layer adapter within SAFE-MEME-H outperforms fully fine-tuned models in regular fine-grained hateful meme detection. However, the fully fine-tuning approach with a Q&A setup is more effective for handling confounding cases. We also systematically examine the error cases, offering valuable insights into the robustness and limitations of the proposed structured reasoning framework for analyzing hateful memes.

cs.CL↗

arXiv2024

Exploring the Benefits of Domain-Pretraining of Generative Large Language Models for Chemistry

A proliferation of Large Language Models (the GPT series, BLOOM, LLaMA, and more) are driving forward novel development of multipurpose AI for a variety of tasks, particularly natural language processing (NLP) tasks. These models demonstrate strong performance on a range of tasks; however, there has been evidence of brittleness when applied to more niche or narrow domains where hallucinations or fluent but incorrect responses reduce performance. Given the complex nature of scientific domains, it is prudent to investigate the trade-offs of leveraging off-the-shelf versus more targeted foundation models for scientific domains. In this work, we examine the benefits of in-domain pre-training for a given scientific domain, chemistry, and compare these to open-source, off-the-shelf models with zero-shot and few-shot prompting. Our results show that not only do in-domain base models perform reasonably well on in-domain tasks in a zero-shot setting but that further adaptation using instruction fine-tuning yields impressive performance on chemistry-specific tasks such as named entity recognition and molecular formula generation.

cs.CL↗

arXiv2024

Benchmarking LLMs for Environmental Review and Permitting

The National Environment Policy Act (NEPA) stands as a foundational piece of environmental legislation in the United States, requiring federal agencies to consider the environmental impacts of their proposed actions. The primary mechanism for achieving this is through the preparation of Environmental Assessments (EAs) and, for significant impacts, comprehensive Environmental Impact Statements (EIS). Large Language Model (LLM)s' effectiveness in specialized domains like NEPA remains untested for adoption in federal decision-making processes. To address this gap, we present NEPA Question and Answering Dataset (NEPAQuAD), the first comprehensive benchmark derived from EIS documents, along with a modular and transparent evaluation pipeline, MAPLE, to assess LLM performance on NEPA-focused regulatory reasoning tasks. Our benchmark leverages actual EIS documents to create diverse question types, ranging from factual to complex problem-solving ones. We built a modular and transparent evaluation pipeline to test both closed- and open-source models in zero-shot or context-driven QA benchmarks. We evaluate five state-of-the-art LLMs using our framework to assess both their prior knowledge and their ability to process NEPA-specific information. The experimental results reveal that all the models consistently achieve their highest performance when provided with the gold passage as context. While comparing the other context-driven approaches for each model, Retrieval Augmented Generation (RAG)-based approaches substantially outperform PDF document contexts, indicating that neither model is well suited for long-context question-answering tasks. Our analysis suggests that NEPA-focused regulatory reasoning tasks pose a significant challenge for LLMs, particularly in terms of understanding the complex semantics and effectively processing the lengthy regulatory documents.

cs.CL↗

arXiv2024

MemeMQA: Multimodal Question Answering for Memes via Rationale-Based Inferencing

Memes have evolved as a prevalent medium for diverse communication, ranging from humour to propaganda. With the rising popularity of image-focused content, there is a growing need to explore its potential harm from different aspects. Previous studies have analyzed memes in closed settings - detecting harm, applying semantic labels, and offering natural language explanations. To extend this research, we introduce MemeMQA, a multimodal question-answering framework aiming to solicit accurate responses to structured questions while providing coherent explanations. We curate MemeMQACorpus, a new dataset featuring 1,880 questions related to 1,122 memes with corresponding answer-explanation pairs. We further propose ARSENAL, a novel two-stage multimodal framework that leverages the reasoning capabilities of LLMs to address MemeMQA. We benchmark MemeMQA using competitive baselines and demonstrate its superiority - ~18% enhanced answer prediction accuracy and distinct text generation lead across various metrics measuring lexical and semantic alignment over the best baseline. We analyze ARSENAL's robustness through diversification of question-set, confounder-based evaluation regarding MemeMQA's generalizability, and modality-specific assessment, enhancing our understanding of meme interpretation in the multimodal communication landscape.

cs.CL↗

arXiv2024

Emotion-Aware Multimodal Fusion for Meme Emotion Detection

The ever-evolving social media discourse has witnessed an overwhelming use of memes to express opinions or dissent. Besides being misused for spreading malcontent, they are mined by corporations and political parties to glean the public's opinion. Therefore, memes predominantly offer affect-enriched insights towards ascertaining the societal psyche. However, the current approaches are yet to model the affective dimensions expressed in memes effectively. They rely extensively on large multimodal datasets for pre-training and do not generalize well due to constrained visual-linguistic grounding. In this paper, we introduce MOOD (Meme emOtiOns Dataset), which embodies six basic emotions. We then present ALFRED (emotion-Aware muLtimodal Fusion foR Emotion Detection), a novel multimodal neural framework that (i) explicitly models emotion-enriched visual cues, and (ii) employs an efficient cross-modal fusion via a gating mechanism. Our investigation establishes ALFRED's superiority over existing baselines by 4.94% F1. Additionally, ALFRED competes strongly with previous best approaches on the challenging Memotion task. We then discuss ALFRED's domain-agnostic generalizability by demonstrating its dominance on two recently-released datasets - HarMeme and Dank Memes, over other baselines. Further, we analyze ALFRED's interpretability using attention maps. Finally, we highlight the inherent challenges posed by the complex interplay of disparate modality-specific cues toward meme analysis.

cs.CY↗

arXiv2024

Quantum-Inspired Tensor Neural Networks for Option Pricing

Recent advances in deep learning have enabled us to address the curse of dimensionality (COD) by solving problems in higher dimensions. A subset of such approaches of addressing the COD has led us to solving high-dimensional PDEs. This has resulted in opening doors to solving a variety of real-world problems ranging from mathematical finance to stochastic control for industrial applications. Although feasible, these deep learning methods are still constrained by training time and memory. Tackling these shortcomings, Tensor Neural Networks (TNN) demonstrate that they can provide significant parameter savings while attaining the same accuracy as compared to the classical Dense Neural Network (DNN). In addition, we also show how TNN can be trained faster than DNN for the same accuracy. Besides TNN, we also introduce Tensor Network Initializer (TNN Init), a weight initialization scheme that leads to faster convergence with smaller variance for an equivalent parameter count as compared to a DNN. We benchmark TNN and TNN Init by applying them to solve the parabolic PDE associated with the Heston model, which is widely used in financial pricing theory.

q-fin.PR↗

arXiv2024

Recent Advances in Hate Speech Moderation: Multimodality and the Role of Large Models

In the evolving landscape of online communication, moderating hate speech (HS) presents an intricate challenge, compounded by the multimodal nature of digital content. This comprehensive survey delves into the recent strides in HS moderation, spotlighting the burgeoning role of large language models (LLMs) and large multimodal models (LMMs). Our exploration begins with a thorough analysis of current literature, revealing the nuanced interplay between textual, visual, and auditory elements in propagating HS. We uncover a notable trend towards integrating these modalities, primarily due to the complexity and subtlety with which HS is disseminated. A significant emphasis is placed on the advances facilitated by LLMs and LMMs, which have begun to redefine the boundaries of detection and moderation capabilities. We identify existing gaps in research, particularly in the context of underrepresented languages and cultures, and the need for solutions to handle low-resource settings. The survey concludes with a forward-looking perspective, outlining potential avenues for future research, including the exploration of novel AI methodologies, the ethical governance of AI in moderation, and the development of more nuanced, context-aware systems. This comprehensive overview aims to catalyze further research and foster a collaborative effort towards more sophisticated, responsible, and human-centric approaches to HS moderation in the digital era. WARNING: This paper contains offensive examples.

cs.CL↗

arXiv2024

Transparency in Sleep Staging: Deep Learning Method for EEG Sleep Stage Classification with Model Interpretability

Automated Sleep stage classification using raw single channel EEG is a critical tool for sleep quality assessment and disorder diagnosis. However, modelling the complexity and variability inherent in this signal is a challenging task, limiting their practicality and effectiveness in clinical settings. To mitigate these challenges, this study presents an end-to-end deep learning (DL) model which integrates squeeze and excitation blocks within the residual network to extract features and stacked Bi-LSTM to understand complex temporal dependencies. A distinctive aspect of this study is the adaptation of GradCam for sleep staging, marking the first instance of an explainable DL model in this domain with alignment of its decision-making with sleep expert's insights. We evaluated our model on the publically available datasets (SleepEDF-20, SleepEDF-78, and SHHS), achieving Macro-F1 scores of 82.5, 78.9, and 81.9, respectively. Additionally, a novel training efficiency enhancement strategy was implemented by increasing stride size, leading to 8x faster training times with minimal impact on performance. Comparative analyses underscore our model outperforms all existing baselines, indicating its potential for clinical usage.

eess.SP↗

arXiv2023

Dimension-free comparison estimates for suprema of some canonical processes

We obtain error approximation bounds between expected suprema of canonical processes that are generated by random vectors with independent coordinates and expected suprema of Gaussian processes. In particular, we obtain a sharper proximity estimate for Rademacher and Gaussian complexities. Our estimates are dimension-free, and depend only on the geometric parameters and the numerical complexity of the underlying index set.

math.PR↗

arXiv2023

Carbon Kagome Nanotubes -- quasi-one-dimensional nanostructures with flat bands

We introduce carbon Kagome nanotubes (CKNTs) -- a new allotrope of carbon formed by rolling up sheets of Kagome graphene, and investigate the properties of this material using first principles calculations. Based on the direction of rolling, we identify two principal varieties of CKNTs -- armchair and zigzag, and find that the bending stiffness associated with rolling Kagome graphene into either type of CKNT is about a third of that associated with rolling conventional graphene into carbon nanotubes (CNTs). Ab initio molecular dynamics simulations indicate that both types of CKNTs are likely to exist as stable structures at room temperature. Each CKNT explored here is metallic and features dispersionless states (i.e., flat bands) throughout its Brillouin zone, along with an associated singular peak in the electronic density of states, close to the Fermi level. We calculate the mechanical and electronic response of CKNTs to torsional and axial strains and compare against conventional CNTs. We show in particular, that upon twisting, degenerate dispersionless electronic states in CKNTs split, Dirac points and partially flat bands emerge from the quadratic band crossing point at the Fermi level, and that these features can be explained using a relatively simple tight-binding model. Overall, CKNTs appear to be unique and striking examples of realistic elemental quasi-one-dimensional (1D) materials that can potentially display fascinating collective material properties arising from the presence of strongly correlated electrons. Additionally, distorted CKNTs may provide an interesting material platform where flat band physics and chirality induced anomalous transport effects may be studied together.

cond-mat.mtrl-sci↗