Abstract

Traffic flow prediction is a critical component of intelligent transportation systems, with direct implications for congestion mitigation, safety enhancement, and efficient travel planning. Given the complex spatio-temporal nature of traffic data, a wide range of hybrid deep learning models—such as CNN-LSTM, ConvLSTM, and Temporal Convolutional Networks (TCNs)— have been explored to capture spatial, temporal, and periodic dependencies effectively. In this work, we perform a comprehensive comparison of deep learning models with and without periodicity to evaluate their prediction accuracies. To address the challenge of hyperparameter tuning in these models, we propose a Genetic Algorithm (GA)-based optimization framework for CNN-LSTM, ConvLSTM, and a novel GA-TCN architecture, all of which demonstrate improved performance on benchmark datasets. Furthermore, we introduce a purely temporal deep learning model named Grid LSTM-based Attention Modelling for Traffic Flow Prediction (GLSTM-A), which leverages a combination of Grid LSTM for long-term dependencies, a standard LSTM for recent trends, and a custom attention mechanism to automatically prioritize significant temporal features. GLSTM-A exhibits superior prediction accuracy and memory efficiency compared to existing temporal models such as TCN, LSTM, and Bi-LSTM. Finally, to tackle the challenge posed by limited long-term traffic datasets, we develop TrafficFlowGAN—a GAN-based time-series data generation framework that effectively captures temporal patterns through a joint supervised and adversarial learning process. The synthetic data generated by TrafficFlowGAN closely resembles real-world traffic patterns, thereby enhancing the robustness and accuracy of downstream prediction models. Extensive experimental evaluations and ablation studies validate the efficiency of the proposed models across various traffic forecasting scenarios.

Abstract

The rapid adoption of Large Language Models (LLMs) as tools for annotation and evaluation tasks in Natural Language Processing (NLP) has led to important questions about their reliability, robustness, and alignment with human expectations. This thesis investigates these concerns across three connected lines of inquiry: • the indistinguishability of LLM-generated annotations from human annotations, • human alignment and reliability of LLMs as subjective judges of language quality, and • the susceptibility of LLM-judges to misinformation attacks in evaluation settings. In the first part of this thesis, we study whether annotations made by LLMs can be reliably distinguished from those created by humans. Previous research has claimed that LLMs behave as ’human-like annotators’, motivating our work to rigorously test this hypothesis. We frame this as a classification task: Given a dataset of annotations, can a model detect their origin? Surprisingly, our findings indicate that even state-of-the-art classifiers achieve near-chance performance, with accuracy not exceeding 51%. These results offer strong empirical evidence that LLMs produce annotations that are indistinguishable from human annotators in most settings, validating the ’human-like’ claim from a discriminative modeling point of view. Building on this, we explore whether LLMs can function as reliable evaluators: a more nuanced role that goes beyond labeling and involves subjective scoring of text quality (e.g., summaries, open-ended responses). In this phase, we test how well LLM-generated judgments align with human evaluations and how robust they are to subtle changes in prompt phrasing. We introduce controlled prompt perturbations and evaluate the consistency of LLM scores and textual justifications. The results show a significant lack of robustness: Small, often misleading perturbations lead to large changes in judgment outputs. These inconsistencies expose limitations in using LLMs as trustworthy evaluators, especially in high-stakes or subjective settings. The final part of this thesis presents a preliminary exploration into the robustness of LLM-based evaluators when exposed to misinformation, marking a conceptual shift from controlled perturbations to real-world-inspired adversarial scenarios. We introduce a novel framework that systematically injects misinformation into prompts, grounded in a taxonomy of ten misinformation types. This study probes how these manipulations affect LLM judgments, analyzing both score alignment and justification consistency. Contrary to our initial expectations, the results suggest that LLM-judges demonstrate a surprising degree of resilience. Scores and justifications remain largely consistent across many misinformation types, even when factual content is altered. However, this robustness appears to be superficial, with the LLM-Judge often failing to explicitly recognize or reflect awareness of being misled in its justifications. Rather than detecting and correcting misinformation, they tend to preserve their prior judgments, raising concerns about unconscious robustness without epistemic awareness. These early findings open new avenues for future research. Comparative assessment, misinformation detection, and deeper semantic analysis may reveal subtler vulnerabilities. More broadly, this work lays the groundwork for building misinformation-aware evaluation systems and motivates the development of LLMs that not only evaluate well but do so with fact-sensitive reasoning. Collectively, this thesis explores the evolving role of LLMs as annotation and evaluation agents. While they show promise as ’human-like’ annotators, their behavior as evaluators under adversarial and misinformation-rich contexts highlights both strengths and blind spots. By offering early insights into their robustness and limitations, we provide a foundation for future work aimed at building trustworthy, interpretable, and context-aware LLM judges.

Abstract

Large Language Models (LLMs) are powerful AI tools with broad abilities, but they still struggle with tasks that require deep understanding and complex, non-standard reasoning. This thesis examines two major challenges: how well LLMs understand and interpret cross-cultural communication, and how to improve their advanced problem-solving skills like lateral thinking and context-aware support. It evaluates how well current LLMs handle these tasks and explores two main strategies to improve their performance: advanced prompting and reinforcement learning-based fine-tuning. The first part of the research focuses on cultural understanding. A user study of book reviews shows that most (83%) contain Culture-Specific Items (CSIs), often making cross-cultural understanding difficult. We test different LLMs—including GPT-4o and smaller open models like Aya and Gemma—on how well they can detect and categorize these CSIs. The results show mixed strengths across models (e.g., Gemma-2 does well with social references, Aya with customs), but all models show a noticeable Western bias compared to human evaluations. To support future work, we also provide standardized datasets. The second part explores LLM performance in complex reasoning tasks such as puzzles that require creative and unconventional thinking. We find that advanced prompting—using both static and dynamic examples, along with model-generated reasoning steps—greatly improves results with Gemini 1.0 Pro over basic methods. However, models still do not reach human-level performance, suggesting that prompting is a promising way to improve complex reasoning, even if it is not yet perfect. Finally, the thesis focuses on generating context-aware assistance for solving challenging math problems, particularly those at the Olympiad level. We propose a framework for producing diverse, multistage synthetic hints tailored to different student needs. These hints cover a range of instructional types, including: beginning steps to help students start; relevant theorems and definitions to build conceptual grounding; equivalent examples for analogical learning; next steps and summaries of remaining steps based on student progress; identification of unused information from the problem; and correctionoriented hints to help students recognize and fix mistakes. Using this framework, we fine-tuned Small Language Models (SLMs) through several reinforcement learning approaches—Supervised Fine-Tuning (SFT), Kahneman-Tversky Optimization (KTO), and Odds Ratio Policy Optimization (ORPO)—to train models to generate helpful, targeted mathematical hints. This not only addresses a key gap in automated pedagogical support for difficult problems but also tackles an important research question: how effectively can these reinforcement learning methods align a smaller model with significantly fewer parameters to synthetic data generated by a much larger LLM? Overall, we work on multi-dimensional evaluation of LLMs in cultural understanding and advanced problem-solving. It highlights current limitations, reveals existing biases, and demonstrates that prompting and fine-tuning together can build more capable, adaptive, and culturally aware AI systems.

Abstract

Re-identification (ReID) is a critical challenge in computer vision, predominantly studied in the context of pedestrians and vehicles. However, robust object-instance ReID, which has significant implications for tasks such as autonomous exploration, long-term perception, and scene understanding, remains underexplored. In this work, we address this gap by proposing a novel dual-path object-instance re-identification transformer architecture that integrates multimodal RGB and depth information. By leveraging depth data, we demonstrate improvements in ReID across scenes that are cluttered or have varying illumination conditions. Additionally, we develop a ReID-based localization framework that enables accurate camera localization and pose identification across different viewpoints. We validate our methods using two custom-built RGB-D datasets, as well as multiple sequences from the open-source TUM RGB-D datasets. Our approach demonstrates significant improvements in both object instance ReID (mAP of 75.18) and localization accuracy (success rate of 83% on TUM-RGBD), highlighting the essential role of object ReID in advancing robotic perception. Our models, frameworks, and datasets have been made publicly available.

Abstract

Flapping-wing robots (FWRs) have garnered significant attention in bio-inspired robotics due to their ability to replicate the highly efficient and manoeuvrable flight of birds and insects. Unlike traditional aerial vehicles, FWRs generate lift and thrust through unsteady aerodynamics, allowing for improved agility and safer operation in confined or complex environments. While early research focused on fundamental aerodynamic modelling and control, recent advancements have explored hybrid propulsion systems, morphing wing designs, and lightweight, high-efficiency actuation mechanisms to enhance endurance and payload. However, most existing research follows a trade-off between range and payload capacity, limiting adaptability for mission-specific tasks. This thesis presents an innovative approach to addressing this limitation by developing a modular flapping-wing robot (FWR) with a swappable powertrain module. The proposed system enables dynamic switching between payloadfocused and endurance-optimized flight profiles by designing interchangeable motor-gearbox configurations and varying flapping kinematics. The research investigates the effects of flapping frequency, stroke amplitude, and tandem auxiliary propulsion on flight performance, leveraging simulation tools like PteraSoftware to refine aerodynamic efficiency. Two powertrain modules are fabricated and tested—one optimized for high-frequency flapping to support payload capacities and another with a tandem propeller to reduce flapping effort and extend range. The flight data collected over multiple experimental flight tests validate the adaptability of this modular design, demonstrating its effectiveness in altering mission capabilities without redesigning the entire platform. Beyond enhancing versatility in FWRs, this research contributes to the broader field by addressing key challenges in adaptive aerial robotics. The findings have direct applications in search-and-rescue missions, environmental monitoring, and silent surveillance, where traditional UAVs face operational constraints. Furthermore, the modularity principle introduced here opens pathways for integrating emerging actuation technologies, such as dielectric elastomer actuators (DEA), for future iterations of highly efficient, autonomous flapping-wing robots. Through this thesis, we aim to push the boundaries of biomimetic flight, offering a scalable and reconfigurable approach to mission-specific aerial robotics.

Abstract

The rapid adoption of electric vehicles (EVs) has accelerated the need for scalable, cost-effective charging infrastructure, particularly in urban environments. This thesis presents a curb-side EVcharging solution that is low-cost, plug-and-play, and interoperable with existing smart-city platforms while requiring no continuous Internet connectivity at each charger. This thesis presents an integrated approach to EV charging, combining Wireless Smart Ubiquitous Network (Wi-SUN) technology with Bluetooth Low Energy (BLE) for enhanced connectivity and cost-effective deployment. The proposed architecture employs Wi-SUN, a long-range, mesh-based sub-GHz communication protocol, as a robust and scalable backbone for charger-to-cloud connectivity, thus reducing operational costs by eliminating reliance on cellular networks. BLE is utilized for efficient short-range interactions, enabling user authentication, session management, and rapid connectivity between the user’s smartphone and the EV charger. The charging system adheres to Level-2 AC charging standards (220-240 V) and follows the Open Charge Point Protocol (OCPP) to ensure interoperability, while also complying with oneM2M middleware standards for seamless integration into broader smart city frameworks. The developed system, encapsulated in an IP65-rated polycarbonate enclosure, integrates a Raspberry Pi as the central processor, Arduino Nano 33 IoT for BLE communication, Wi-SUN transceivers, a solid-state relay, and a energy-metering unit. The field evaluation conducted within the Smart City Living Lab at IIIT Hyderabad validated key performance metrics, demonstrating a reliable communication range of approximately 370 meters between two consecutive Wi-SUN nodes, BLE connection establishment in under 300 milliseconds, and an end-to-end authentication time of less than a second. Furthermore, a proof-of-concept emergency vehicle detection feature underscores the versatility and extensibility of the developed architecture. Overall, this integrated Wi-SUN and BLE-based streetlight-mounted EV charger provides a practical, scalable solution to support the widespread adoption of electric vehicles in urban centers, paving the way for advanced smart-city applications.

Abstract

r exaggerate content to drive user engagement. While they may boost clicks, such headlines and articles compromise reader trust and contribute to the spread of misinformation. Tackling clickbait thus represents not just a technical challenge but also a crucial step toward improving the quality and transparency of online information. To address this, recent research has introduced the task of clickbait spoiling, where we attempt to ”spoil” the headline by generating the key information the article withholds. Our work contributes to this growing area by firstly engaging with the clickbait spoiling task in English using modern Large Language Models (LLMs) and secondly, in Hindi, where we extend this task into a vastly underexplored linguistic space. The spoiling task is especially meaningful because it does not attempt to censor or suppress content instead it just simply makes it more honest, giving users the context they need upfront to make informed choices about what they click. From the evaluations on the English dataset it is revealed that LLMs, particularly when prompted with a few-shot setup, can be highly effective at generating relevant and informative spoilers. This demonstrates the potential of prompting-based systems to perform sophisticated reasoning and summarization tasks, even in cases where training data may be limited. Building on these insights, we make a significant contribution by shifting focus to Hindi, a language that has been critically underrepresented in this field of NLP research despite being spoken by hundreds of millions. We curate the first-of-its-kind Hindi clickbait dataset, using automatically generated annotations from multiple powerful LLMs. This dataset will not only address a huge gap in multilingual clickbait research but also lay a foundation for future work in Hindi misinformation and media literacy. We also observed notable inconsistencies in the behavior of certain large language models. Hermes-3, in particular, exhibited significant divergence in both its clickbait classification judgments and its assigned ratings. It was unable to understand explicit instructions and hence in light of this unreliability, we made a principled decision to take it off the annotator lineup. This screams for quality control whenever using LLMs for such tasks. We then extended the spoiling task to Hindi using this dataset by introducing a selection process based on agreement across models, combined with manual curation, resulting in a dataset that captures the diverse forms of Hindi clickbait. We also classify spoiler types and evaluate LLM performance in zero-shot and few-shot settings. Our findings reveal clear patterns: model performance varies widely, prompting strategies significantly affect outcomes, and certain models (like Deepseek-R1) demonstrate robust generalization even in zero-shot settings.

Abstract

Telugu, a Dravidian language spoken by approximately 95.7 million people, faces significant barriers in digital ecosystems due to its agglutinative morphology and limited computational and informational resources. This thesis addresses these challenges through two transformative contributions: a contextaware morphological analyzer and the generation of 25,000 high-quality Telugu Wikipedia articles to enhance knowledge accessibility. To tackle the scarcity of annotated datasets, a novel 10,000-sentence Telugu corpus was developed, meticulously annotated with morphological features such as lexical category, gender, number, person, case, tense-aspect-mood (TAM), and clause markers. This dataset, capturing complex forms like adav ¯ allalMdark ¯ ¯ı (‘to all women’) and ambiguities like pagalu (‘day’ vs. ‘to break’), powers a Transformer-based analyzer. Fine-tuned on BERT-Te, a monolingual model pretrained on 80,15,588 Telugu sentences, the analyzer achieves F1 scores of 0.778 for gender and 0.602 for lexical category, outperforming multilingual models (mBERT, IndicBERT, XLM-R) by 20-25%. This enables precise parsing for applications like machine translation and sentiment analysis, addressing a critical gap in Telugu NLP infrastructure. Concurrently, a scalable pipeline leveraging Jinja2 templates and structured data from IMDb, USDA, IUCN, and Smithsonian generated 25,000 Wikipedia articles across movies (8,929), plants (2,140), animals (4,928), and volcanoes (8,676). These articles, incorporating infoboxes, tables, and citations, achieve a BERTScore of 0.79 for semantic similarity and a Fleiss’ Kappa of 0.82 for human-rated quality, adhering to Wikipedia’s five pillars (neutrality, verifiability, no original research, free content, civility). Expanding Telugu Wikipedia’s 95,599-article base by over 25%, they empower native speakers with accessible knowledge in entertainment, science, and geography. Both contributions are opensourced at https://github.com/parameshkrishnaa/Telugu-Morph-Dataset and https://tewiki.iiit.ac.in, ensuring reproducibility and adaptability to other low-resource languages. Despite challenges like sandhi splitting (8% manual corrections) and transliteration errors (e.g., “The” as Te, 8% manual fixes), the ˆ research establishes robust benchmarks—F1 scores, BERTScore, and Fleiss’ Kappa—for Telugu NLP. By enhancing computational tools and public knowledge access, this work elevates Telugu’s digital presence, fosters linguistic equity, and provides a scalable model for global low-resource language advancement. Future directions include expanding datasets, automating preprocessing, and deploying articles on Wikipedia to further bridge the digital divide.

Abstract

Examining limitations is a crucial step in the scholarly research review process, revealing aspects where a study might lack decisiveness or require enhancement. This practice aids readers in considering broader implications for further research and helps in identifying potential areas for improvement and future investigation. In this thesis, we introduce a novel and challenging task of Suggestive Limitation Generation (SLG) for research papers. To facilitate this task, we compile a comprehensive dataset calling it LimGen, which encompasses 4068 research papers along with their associated limitations from the ACL anthology. The LimGen dataset represents a significant resource for the academic community, providing a structured and extensive collection of limitations that can be used to train and evaluate models aimed at generating meaningful and constructive limitations. We investigate several approaches to harness large language models (LLMs) for producing suggestive limitations. This involves a thorough examination of related challenges, practical insights, and potential opportunities. Our research delves into various techniques and methodologies, assessing their effectiveness and exploring ways to enhance the generation of suggestive limitations. The challenges addressed in this thesis include the inherent difficulty of understanding complex research content, the need for generating limitations that are both relevant and insightful, and the integration of domain-specific knowledge to improve the quality of the generated limitations. We also discuss the practical implications of our findings, providing valuable insights for researchers and practitioners looking to leverage LLMs for this purpose. We experimented with various setups, including using summarization-specific pre-trained models like BART and PEGASUS, and generative models such as T5, Cerebras-GPT, and Llama 2. To address the length constraint of input papers, we employed the Dense Passage Retrieval (DPR) approach, which processes relevant passages to maintain context. Additionally, we implemented a chain modeling technique to refine the generated limitations for better coherence and relevance. A comprehensive evaluation of the proposed models was conducted using automatic (ROUGE-based), human (manual expert-based), and model assistive (LLM-based) approaches. The results indicate that models trained specifically for summarization do not effectively generate insightful limitations, highlighting the complexity of the SLG task compared to traditional summarization tasks. Our contributions include the development of the LimGen dataset, the proposal of several LLM-based schemes for SLG, and a thorough evaluation of these methods. The LimGen dataset and accompanying code are made available to the research community to foster further exploration and development in this area. Our work aims to contribute to the advancement of automated tools for research evaluation, ultimately aiding in the improvement of research quality and the acceleration of scientific discovery

Abstract

In recent years, the widespread adoption of the internet and the growth in user-base of various social media platforms have led to a notable increase in the proliferation of extreme speech online. While traditional Natural Language Processing (NLP) models have demonstrated proficiency in distinguishing between neutral text and non-neutral text (i.e. extreme speech), categorizing the diverse types of extreme speech presents significant challenges. The task of extreme speech classification is particularly nuanced, as it requires a deep understanding of socio-cultural contexts to accurately interpret the intent of the language used by the speaker. Even human annotators often disagree on the appropriate classification of such content, emphasizing the complex and subjective nature of this task. The use of human moderators also presents a scaling issue, necessitating the need for automated systems for extreme speech classification. The recent launch of ChatGPT has drawn global attention to the potential applications of Large Language Models (LLMs) across a diverse array of tasks. Due to their training on vast and diverse corpora, coupled with their demonstrated ability to capture and encode contextual information effectively, LLMs emerge as highly promising tools for tackling this specific task of extreme speech classification. Their capacity to generalize across linguistic patterns and contextual nuances makes them particularly well-suited for applications requiring sophisticated language understanding. LLMs have the capability to understand complex language and context, making them useful tools for identifying and controlling extreme speech on online platforms. In this thesis, we leverage the Indian subset of the extreme speech dataset (XTREMESPEECH) introduced by Maronikolakis et al. [63] to evaluate zero-shot classification, finetuning, preference optimization, and ensemble methods. We address two primary research questions: (1) How do LLMs perform on multilingual extreme speech datasets, particularly within India’s linguistically and culturally diverse context? (2) To what extent can open-source LLMs serve as viable alternatives to proprietary GPT models, and does fine-tuning bridge any performance gap? Our findings indicate that while pre-trained LLMs demonstrate moderate efficacy in detecting extreme speech in a zero-shot setting, their performance is substantially enhanced through fine-tuning with domain-specific data. This improvement underscores the adaptability of these models in capturing linguistic subtleties and contextual nuances. Furthermore, we note that while GPT-based models exhibit superior performance compared to open-source Llama models in a zero-shot context, the performance gap vanishes after fine-tuning. This suggests that open-source LLMs, when appropriately fine-tuned, can achieve comparable results, thereby presenting a viable alternative for extreme speech classification tasks.