Abstract

Coreference resolution (CR) is the task of identifying text spans that refer to the same entity. It is a fundamental component of natural language understanding with applications in various downstream NLP tasks, such as question answering, knowledge graph construction, and summarization. Despite its significance and the advancements made by neural coreference models, CR models face a major bottleneck: their limited generalization capability. Prior work attributes this generalization gap to differences in annotations, such as what constitutes a mention (or entity) and varying preferences to span boundaries. For a model to have strong referential capabilities, it must adapt to these annotation-specific nuances. However, achieving this level of adaptability remains a significant challenge, even for state-of-the-art (SOTA) models. This challenge is further amplified when evaluating the referential capabilities of large language models (LLMs) in a few-shot setting, where replicating nuanced annotations with just a few examples is highly unrealistic. We observe that these annotation-specific nuances, can be beneficial but are not essential for downstream tasks or for evaluating the core referential capabilities of an LLM. We describe these nuances as bells and whistles. In this work, we redefine the traditional formulation of coreference resolution by shifting focus away from its bells and whistles. Instead, we propose task formulations more aligned with practical applications and demonstrate improved generalizability across domains. Our first contribution introduces an alternative referential task, Major Entity Identification (MEI). MEI simplifies referential tasks by:(a) assuming that target entities are explicitly provided in the input, and (b) focusing exclusively on frequent entities. Assuming entities to be part of the input shifts the responsibility for domain-specific annotation adaptation—determining which entities are annotated—from the training phase to inference. Through extensive experiments, we show that MEI models generalize effectively across domains using both supervised approaches and LLM-based few-shot prompting across multiple datasets. Importantly, MEI aligns with the classification framework, enabling the use of robust, intuitive, and well-understood classification-based evaluation metrics. Beyond its theoretical appeal, MEI also has practical utility as it allows users to efficiently search for all mentions of a specific entity or a group of entities of interest. Our second major contribution addresses critical shortcomings identified in recent evaluations of large language models (LLMs) on coreference resolution. These studies revealed that traditional output formats and evaluation metrics fail to capture models’ referential understanding fully. Traditional evaluation methods require reproducing the entire document along with annotated cluster information or precisely replicating the antecedent span. This introduces additional bells and whistles, such as ensuring the accurate reproduction of spans and documents. To tackle this issue, we introduce IdentifyMe, a new benchmark for mention resolution that adopts a multiple-choice question (MCQ) format—a widely used evaluation approach for LLMs. With this simplified task design, any failure can now, be attributed exclusively to issues with mention resolution. IdentifyMe presents long narratives and applies heuristics to eliminate easily identifiable mentions, resulting in a more challenging and rigorous task. The benchmark incorporates a curated mix of various mention types and their corresponding entities, enabling fine-grained analysis of model performance. Notably, LLM performance remains substantially below human-level performance on IdentifyMe, highlighting considerable room for improvement even for advanced models like GPT-4. The evaluation also reveals key weaknesses in current LLMs, particularly with pronominal mentions, nested mentions, and other nuanced cases. Overall, this work moves beyond traditional coreference resolution formulations, focusing on tasks with practical applicability and providing fresh insights into the referential strengths and weaknesses of current models. We term this approach Coreference Without Bells and Whistles — a streamlined perspective that prioritizes utility and understanding of model capabilities over tailored annotation adaptation

Abstract

This thesis presents novel contributions to the field of Private Information Retrieval (PIR), a fundamental cryptographic technique that enables users to query public databases without revealing the identity of the accessed data. The study begins by revisiting classical PIR schemes introduced by Chor et al. [1], establishing a foundational understanding of how user privacy is preserved through query obfuscation. Building on this foundation, the thesis advances PIR design by incorporating modern dis- tributed storage systems and powerful coding-theoretic tools, focusing on the application of Berman codes—binary linear codes of length nm (for n ≥ 2, m ≥ 1)—which are known to achieve the capacity of the binary erasure channel. A significant portion of the work explores (D, E)-retrieval schemes, which utilize linear codes for both storage and retrieval, enabling ro- bustness against server collusion and improving data access efficiency. A key contribution of this thesis is the development of a new PIR framework that supports up to t colluding servers. This construction employs the star product of Berman codes and achieves strong privacy guarantees while maintaining optimality in terms of PIR capacity. The proposed schemes provide flexible control over system parameters, including the number of servers, the storage overhead, the PIR rate, and the collusion threshold t. Our results introduce a new class of capacity-achieving PIR protocols that ensure the privacy of the retrieved file’s identity, even in adversarial settings with colluding databases. Notably, these schemes are constructed entirely over binary fields, offering a practical advantage by reducing implementation complexity compared to prior approaches that require large finite fields. Through rigorous theoretical analysis and constructive design, this work demonstrates the prac- tical viability and theoretical strength of Berman-code-based PIR systems. The methodologies proposed herein significantly enhance the privacy, efficiency, and deployability of PIR schemes, with far-reaching implications for secure data access in privacy-sensitive domains such as health- care, finance, and cloud computing.

Abstract

This thesis is based on a research problem carried out in [1]. The problem statement deals with the identification of secret sharable state in three qubit system for which maximum secret reconstruction is possible. We classify this set of states as MSR states. For a given value of maximum of the tele- portation fidelities that can be achieved by the bipartite channels of dealer-receiver and dealer-assistant, these states achieve the maximal secret reconstruction fidelity. In other words, it deals with finding a potential resource for secret sharing that maximizes the potential for secret sharing. A tripartite state can be considered a resource for secret sharing if it imposes limits on the teleportation fidelities in both bipartitions involving the dealer while still being useful to reconstruct the secret. Similarly, for a specific maximum value of the Bell-CHSH measurement in the bipartitions specified above, we determine the highest possible reconstruction fidelity. Another interesting discovery is that all secret-sharable states satisfy Bell’s inequality in both dealer- reconstructor and dealer-assistant channels. This reveals a novel form of mutual exclusivity between secret-sharable states and violations of Bell’s inequality. This thesis is based on our findings in [1] and it helps to identify the best candidates among secret sharing resource states to achieve maximum recon- struction fidelity, setting a practical limit on information transfer in a resource theoretic extension of secret sharing. Moreover, it highlights a new form of exclusivity between bipartite correlations and the ability to perform secret sharing in a tripartite setting

Abstract

In the rapidly evolving landscape of Very-Large-Scale Integration (VLSI) design, the integration of machine learning (ML) techniques has emerged as a powerful tool to enhance automation and optimization processes. However, the effectiveness of these ML applications is often hampered by a critical challenge: data scarcity. As VLSI systems grow in complexity, the demand for high-quality training datasets becomes increasingly vital for the development of robust and accurate ML models. This thesis addresses this pressing issue by exploring innovative strategies for data augmentation through the application of generative models, specifically Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and diffusion models. The initial segment of this research investigates the use of GANs to generate supplementary circuit data based on simulations performed in established design environments, including Cadence Virtuoso, HSPICE, and Microcap. By leveraging GANs, we aim to synthesize artificial circuit data that accurately reflects the characteristics of real-world data, thereby enhancing the training sets available for machine learning algorithms. A comprehensive comparative analysis is conducted to evaluate the performance of GANs against VAEs, highlighting their respective strengths and weaknesses in the context of data augmentation specifically tailored for VLSI applications. Subsequently, the thesis delves into the capabilities of diffusion models, which have recently gained prominence for their effectiveness in generating high-fidelity synthetic data. By utilizing these advanced generative techniques, we demonstrate the potential to produce artificial datasets that closely resemble real electronic circuit behavior, effectively addressing the limitations posed by traditional data collection methods. Through simulations conducted in the HSPICE design environment, we establish the quality and reliability of the synthetic data generated, thereby validating its applicability in enhancing ML model performance. A key focus of this research is the rigorous evaluation of the authenticity of the synthetic data and its impact on the predictive accuracy of machine learning models. The results reveal a significant reduction in prediction errors for circuit performance assessments when models are trained on augmented datasets, showcasing the transformative potential of generative models in the VLSI design domain. This improvement in accuracy not only highlights the effectiveness of the proposed methodologies but also underscores the importance of data-driven approaches in advancing the capabilities of electronic design automation In addition to presenting the main contributions of this thesis, this work also sheds light on the challenges encountered during the implementation of generative techniques. Issues such as the calibration of model parameters, the validation of synthetic data fidelity, and the integration of these data solutions into existing design workflows are critically examined. Furthermore, we provide insights into the lessons learned from experiments that did not yield successful results, offering valuable perspectives for future research endeavors in this field. By addressing the data scarcity issue through innovative generative modeling approaches, this thesis contributes to the ongoing dialogue within the VLSI community regarding the future of machine learning applications in electronic design automation. The findings of this research not only pave the way for more effective ML solutions in VLSI design but also serve as a foundation for future advancements that may ultimately reshape the landscape of electronic design and technology

Abstract

The emergence of AI-driven systems has transformed smart grid networks, enabling widespread automation in decision-making. A smart grid is an advanced electricity distribution system that empowers customers to actively participate through smart meters. These grids operate across three key markets: wholesale, tariff, and balancing markets. In this ecosystem, distribution companies, or electricity brokers, play a pivotal role by procuring electricity in the wholesale market through double auctions and selling it to customers in the tariff market. Meanwhile, the balancing market oversees real-time supply and demand, penalizing brokers responsible for imbalances. The primary goal of brokers is to maximize profits, which involves addressing three critical challenges: (i) minimizing procurement costs in the wholesale market, (ii) offering competitive yet profitable tariffs in the tariff market, and (iii) mitigating peak demand scenarios to avoid penalties. Achieving this requires brokers to develop intelligent strategies leveraging AI techniques. This thesis develops AI-based strategies that enhance the efficiency of electricity brokers, tested using a close-to-real-world simulation platform, PowerTAC. Specifically, we design various bidding strategies for brokers operating in the wholesale market, where electricity is traded through day-ahead periodic double auctions (PDAs). The first strategy draws inspiration from game theory, leveraging Nash equilibrium principles for a single-buyer and single-seller setup. This is extended to real-world scenarios consisting of multiple buyers and sellers using reinforcement learning techniques. The second strategy reduces procurement costs by exploiting the supply curve information of prominent sellers to inform bidding decisions. The third strategy employs Markov Perfect Nash Equilibrium (MPNE) policies to model buyer behavior with known supply curves and introduces an algorithm to address scenarios where supply curve information is unavailable. The fourth strategy leverages Monte Carlo Tree Search (MCTS) to operate in the continuous action space of bid prices for optimized bidding in PDAs. In the tariff market, we develop strategies for generating attractive tariff contracts to build and retain a robust customer base. Using game theory, we demonstrate that maintaining an optimal market share significantly boosts net revenue. To achieve this, we propose both heuristic and learning-based approaches, with the latter employing multiarmed bandit (MAB) techniques for tariff generation. To address peak demand scenarios, we propose a demand response mechanism that incentivizes customers to shift their electricity usage away from peak hours. We present an optimal algorithm for allocating discounts that maximize expected peak demand reduction while adhering to budget constraints. Additionally, we introduce an MAB-based online algorithm to handle cases where customer responses to incentives are initially unknown. Finally, we integrate these strategies to develop our autonomous broker, VidyutVanika, for the PowerTAC simulation-based tournament. VidyutVanika competed against other brokers with the objective of maximizing profits. The results from the PowerTAC 2021 and 2022 tournaments demonstrate the effectiveness of our approach, with VidyutVanika emerging as the champion in both years.

Abstract

The field of sketch generation and recognition has seen significant advancements through the innovative application of generative models. This thesis presents a comprehensive exploration of face stylization , artistic portrait generation and forensic sketch synthesis, leveraging the power of stateof-the-art generative models like StyleGAN and StableDiffusion. Our work addresses key challenges in preserving identity, accommodating various poses, and bridging the modality gap between sketches and photographs. Through three interconnected studies, we demonstrate significant advancements in generating high-quality sketches and improving forensic applications. We begin by introducing a novel approach to face cartoonization that preserves identity and accommodates various poses. Unlike conditional-GAN methods, our technique utilizes an encoder to capture pose and identity information, generating embeddings within StyleGAN’s latent space. This approach uniquely adapts a pre-trained StyleGAN model, originally designed for realistic facial images, to produce cartoonized outputs without requiring a dedicated fine-tuned model. Building upon this foundation, we present Portrait Sketching StyleGAN (PS-StyleGAN), a style transfer approach tailored for portrait sketch synthesis. PS-StyleGAN leverages StyleGAN’s semantic W+ latent space to generate portrait sketches while allowing meaningful edits such as pose and expression alterations. By introducing Attentive Affine transform blocks and a specialized training strategy, we achieve high-quality sketch generation without fine-tuning StyleGAN itself. This method demonstrates superior performance over current state-of-the-art techniques, requiring only a small number of paired examples and minimal training time. Finally, we address the challenging task of forensic sketch-to-mugshot matching with CLIP4Sketch, a novel approach that uses diffusion models to generate diverse sketch images. By combining CLIP and Adaface embeddings of reference mugshots with textual-style descriptions, we create a comprehensive dataset of sketches corresponding to mugshots. This synthetic data significantly improves the accuracy of sketch-to-mugshot matching in face recognition systems, outperforming training on limited real face sketch data and datasets made by GAN-based methods. Collectively, these contributions push the boundaries of sketch generation and recognition, offering promising applications in both the artistic and forensic domains.

Abstract

This thesis addresses the critical challenge of improving road safety by introducing novel approaches to predictive modeling of accident-prone zones and action recognition in critical traffic scenarios. It makes two key contributions: the early identification of accident-prone zones using Advance Driving Assistance System (ADAS) data and the development of IDD-CRS, a comprehensive dataset for action recognition in unstructured road environments. In the first study, geo-tagged collision alert data from a fleet of 200 ADAS-equipped city buses in Nagpur, India, is leveraged to proactively identify high-risk zones across urban road networks. Using Kernel Density Estimation (KDE), this study captures the spatiotemporal distribution of collision alerts, enabling the detection of emerging blackspots before accidents occur. A novel recall-based metric evaluates the alignment of these predicted zones with historical blackspots, while Earth Mover Distance (EMD)-based analysis identifies previously unreported accident-prone areas. This predictive framework provides civic authorities with actionable insights for targeted interventions, such as traffic-calming measures and infrastructure improvements, thereby enhancing public safety. The second part of the thesis introduces the IDD-CRS dataset, a large-scale collection of traffic scenarios recorded using ADAS and dash cameras. IDD-CRS fills a critical gap in existing datasets by focusing on complex interactions between vehicles and pedestrians, with scenarios such as high-speed lane changes, unsafe vehicle approaches, and near-miss incidents. With precise temporal annotations powered by ADAS technology, the dataset ensures accurate event boundaries, providing a robust benchmark for action recognition and long-tail action recognition tasks. It includes 90 hours of footage spanning 5,400 one-minute videos and 135,000 frames, with hard negative examples to challenge existing models. Initial benchmarks highlight the limitations of current video backbones in recognizing rare events, emphasizing the need for further advancements. Together, these contributions provide a holistic framework for improving road safety through proactive accident prevention and robust action recognition in traffic scenarios. By addressing both spatial accident prediction and temporal event recognition, this work offers foundational resources and actionable insights to advance research and practical solutions for safer road environments.

Abstract

Understanding what makes certain stimuli memorable is a critical area of research that has implications in two major industries: advertisements and education. The ability to create memorable essays and other textual materials on different topics are of great importance in education. Similarly, in the entertainment industry, there is a need to identify short video clips that are memorable to content consumers. This thesis investigates the concept of memorability in two distinct modalities: words and videos. The general topic under study is the factors influencing memorability, with a specific focus on how features such as semantic network properties and word properties affect word memorability, and how spatio-temporal attention mechanisms influence video memorability. The central questions addressed are: What makes certain words and videos more memorable than others? How do cognitive and cultural factors impact memorability across different languages? Previous research has explored various aspects of word and video memorability, but there is a gap in understanding the cross-linguistic differences in word memorability and the underlying human vs. model mechanisms for video memorability. In our first exploration, we conduct a cross-linguistic study comparing the memorability of words in English and Hindi. Our analysis reveals that centrality metrics in semantic networks are weakly indicative of word memorability in English, as well as in Hindi. Moreover, the memorability of words in Hindi and English were found to be uncorrelated, suggesting that language-specific cognitive and cultural factors might play a significant role in memorability. Word frequency remains a strong factor influencing memorability in both English and Hindi. This research highlights the need to consider both intrinsic and extrinsic effects when studying memorability. In the second exploration, we leveraged a simple CNN+Transformer model to predict video memorability. This model not only achieved state-of-the-art performance but also provided insights into spatio-temporal attention patterns. Human gaze data was collected to compare with the model’s attention mechanisms, showing that both humans and the model focus on similar regions of the videos, both spatially and temporally. This finding mirrors human attention patterns and provides a deeper understanding of the temporal dynamics in video memorability. Additionally, our analysis identified critical issues in existing video memorability datasets, offering guidelines for future dataset construction and validation.The significance of this research lies in its contribution to the understanding of memorability across different modalities and languages. Future research can build on these findings to enhance the robustness and generalizability of memorability models, as well as build foundations for endto-end products. A few examples include helping students learn better by making memorable educational content, big companies understanding their consumer base better by analysing advertisement content, and so on

Abstract

This work focuses on applying task-specific deep learning models to the brain encoding for both language and speech processing. Brain encoding involves developing models that predict the response of the brain to stimuli, which could reveal some understanding of how the brain processes the information presented. Previous work showed it was possible to predict brain responses successfully using selfsupervised language and speech models like Bert, Wav2Vec2.0, etc., but little examination has been conducted to determine whether further improvement can be achieved by using fine-tuned task-specific models. In this work, we rely on a variety of transformer-based models fine-tuned for different speech and language tasks—such as automatic speech recognition (ASR), phoneme recognition (PR), and sentiment analysis (SA)—to make predictions on neural activity. Such task-specific models should better capture the fine details of brain activity during listening to stories or reading sentences. Our method trains encoding models using these fine-tuned representations to predict responses in fMRI for different regions of the brain, including the whole brain, as well as specific language and auditory regions. This work introduces an overarching framework that systematically benchmarks the performance of task-specific representations in predicting brain activity. This is then demonstrated with two very different datasets—one from the reading of sentences (Pereira dataset) and the other from listening to spoken stories (Narratives dataset). For both datasets, we observe that the best performance in terms of predicting whole brain activity as well as language regions corresponds to models fine-tuned on ASR for speech encoding and Named Entity Recognition (NER) for language encoding. Further differences were observed in task representations that were maximally associated with brain responses related to reading versus listening. We also explored cross-modal functioning by comparing how the brain responds differently when processing linguistic stimuli through reading versus listening. The key contributions of this work are brain encoding taskonomy, which categorizes tasks along the lines of how well they align with neural responses. We analyze abstract versus concrete concepts and bring out different neural pathways for the same. Our results demonstrate the unique contributions of semantic tasks such as paraphrase generation and summarization and syntactic tasks such as coreference resolution and named entity recognition towards brain encoding while providing insight into neural correlates for language comprehension and auditory processing. This research advances cognitive neuroscience by illuminating how representations, especially taskspecific representations derived from deep learning models, can further elucidate encoding mechanisms in the brain. It bridges AI and neuroscience; potential applications could actually mean a breakthrough in brain-computer interfaces and cognitive technologies

Abstract

Current pedagogical systems in dance majorly focus on performance training which has a major leaning towards entertainment service. As a result, the repertoire of movements has been restricted in catering to what is popular. This has led to a repetitive template-based training fitting into the audio-visual norms of the aesthetics of the time, which restricts the scope for further evolution of the dance form and curtails the possibilities of newer dance forms to emerge. Historically dance as a knowledge system was fundamentally taught in a form-neutral, styleneutral fashion with the ulterior motive of serving broad societal functions including ritualistic, devotional, and entertainment purposes. It is important to observe that only with the universal understanding of dance, the masters of yesteryears were able to design and formulate newer artforms in order to meet newer requirements of the clients and times. Any knowledge system has its universal fundamentals through which any of its regional practices can be taught. For example, music is generally taught through musical notes, considered to be universal and allows any musical system to be taught and practised. However, dance does not have common fundamentals through which all forms of dances can be taught. This research aims at deriving these universal fundamentals of dance systematically and devise a pedagogical system to disseminate these universals. To this end, the objectives are as follows. First, to establish the possibility of having universal dance fundamentals by reviewing the pertinent Dance Texts and understanding Dance Practices of Traditional Practitioners. Second, to systematically arrive upon categories of universal dance fundamentals to create form-neutral, style-neutral holistic Indian Dance Pedagogy catering to children and beginners. Third, to create a Pedagogical method to teach the universal dance fundamentals. Fourth, to develop a pedagogical tool as an aid to teach the universal dance fundamentals. Since the lower half body training serves as the fundamental training offered in any dance form, this research is focussed in creating a pedagogical system to understand the dance movements of the lower half of the body. In order to identify the universal fundamentals of dance, a survey of the lower-half bodypostures and movements mentioned in the Ancient Indian Dance texts such as Nāṭyaśāstra, Saṅgītaratnākara and Abhinaya Darpaṇa, and specific postures where the foot/feet are in contact with the floor were collected. As the major dance texts are not form- or style-specific, the postures mentioned in the dance texts are form-neutral and style-neutral. Based on analysis of these texts and from practice, the logical progression from one posture to another is identified. This gives rise to the logical prediction of the unmentioned postures in between the already mentioned postures of the texts. This enables a learner not to memorise the postures but gives a pedagogical method to understand movements as a progression based on logical parameters. Similarly the movements from the dance texts and practices were collected and logically arranged. In the process, it is observed that movements have three major components, that is, Posture, Transition and Rhythm in Transition. Posture is the static body position which serves as a marker to the movements. This way of understanding Postures springs out of an anatomical approach and the possible affordances of the joints, to make a posture. The postures are systematically arrived at based on eight parameters namely, part of the body in contact with the earth, Knee Positions, Knee angle, Feet angle, Body weight orientation, Foot position, Part of the foot touching the point and Feet Distance. Transition is dynamic. It is understood as a transitory movement from one posture to another. Despite similar postures across sports, martial arts, yogasana, dance and theatre could be common, the Transition, or the movement in between the postures is the factor which decides whether the act is of sports, martial arts, yogasana, dance or theatre. It is in fact the Transition which decides the form or style of dance. The Transitions are systematically arrived at based on seven parameters namely, lifted leg, point of lifting the leg, foot variation which is touching the point of lifting, part of the foot which is touching the point, trajectory traced by the foot as it reaches to the point, trajectory traced by the foot as it comes out of the point of lifting and foot activity. Furthermore, the parameter Rhythm in Transition determines the specific dance form and dance style. By understanding Postures, Transitions and Rhythm in Transition one can systematically understand the fundamentals of the movements, and more specifically dance movements. The mentioned three parameters served as a basis for a pedagogical framework, based on which a pedagogical meth