Abstract

The exchange of information between humans involves the use of speech and text and is known as Natural Language. Each day, people communicate with each other in various languages through speech or text, sharing a vast amount of information. The data produced through natural language communication can offer valuable insights despite being ambiguous, unstructured, and noisy. However, computers cannot interpret natural language on their own yet. To fully comprehend this data and respond intelligently, computers need the ability to understand and emulate human language. This is where Natural Language Processing (NLP) comes in - it is a branch of Artificial Intelligence that focuses on enabling machines to read, comprehend and derive meaning from human languages. NLP integrates the disciplines of linguistics and computer science. It decodes language, its structure, rules, and creates models capable of comprehending, analyzing, and extracting important information from both text and speech. An abundance of information is available in the form of text, including books, documents, articles, social media posts, and more. A document, one of the oldest forms of information exchange, refers to written, printed, or electronic material that is created to facilitate the exchange of information from its author to its intended audience. These files contain valuable information that can significantly benefit business activities. With the use of NLP applications, insights can be extracted from text data. Enterprises utilize NLP applications for various purposes, ranging from document understanding, information extraction, or providing answers to common questions. In this thesis, we develop techniques for deeper analysis and understanding of documents that are commonly used in enterprises. A contract is a frequently used type of document in the corporate world. Contracts are agreements between two or more parties, that govern what each party can or cannot do and are usually dense in information. Automatically extracting key components or components that contain rare or novel information from these large documents makes reviewing contracts easier. Nevertheless, it can be a challenging task as the key and novel components are not present in isolation within the contract. Extraction of significant components (key components + novel components) from contracts aims to simplify the end user’s comprehension and reduce dependency on legal experts for reviewing contracts. In this thesis, we introduce approaches for the automatic identification and extraction of significant components from a contract. We propose a Bidirectional Encoder Representations from Transformers (BERT) based model that automatically identifies or highlights significant components of a contract. In the corporate world, reports are also a frequently encountered type of document. A report is a document that provides information and analysis on a particular topic or issue. Reports are used to convey important information to stakeholders, such as managers, executives, investors, and customers. The vast data available in these reports have the potential to revolutionize datadriven analysis. Causality identification and span detection is one such data-driven task. The relationship between two entities where one causes another event to happen is known as cause and effect. We explored various transformer-based models that help in classifying sentences as well as identifying spans in a sentence.

Abstract

Hospital buildings are expected to remain Occupiable after earthquakes to cater to the post-disaster response and rescue efforts. This requires a stiffer and stronger structural system compared to the commonly adopted moment frames in normal buildings; wallframe systems are recommended in these buildings. But, in addition to merely providing the wall-frame structural system, it is also essential to adequately design the wall-frames to meet the desirable Occupiable seismic performance. Appropriate seismic structural configuration, Structural Plan Density (SPD) of structural walls, and seismic design parameters are presented of typical hospital building located in Seismic Zone IV, that helped meet the preferred performance. A displacement-based limit state of structural damage in line with classic displacement demand estimation is proposed, to confirm the performance. Fiber inelasticity in walls and moment frames in study frames and buildings are defined, for performing displacement controlled nonlinear static pushover analyses and nonlinear time history analyses in commercial software PERFORM3D. Alongside, limit states of structural damages, namely yielding of longitudinal reinforcement in beams or columns in tension, yielding of vertical longitudinal reinforcement in structural walls in tension, crushing of confined concrete in compression in beams, crushing of confined concrete in compression in columns, and spalling of unconfined concrete in compression in structural walls are also monitored to grade the damages, and in turn the seismic performance of wall-frames. Results suggest, numerical models with fiber inelasticity help predict nonlinear behaviour reasonably well. Moment frames do not provide preferred seismic performance, but wall-frames with plan-aspect ratio more than 4 enhances lateral stiffness, strength, and ductility of wall-frames, thereby meeting the preferred performance. The major findings of the study are at least 3% SPD of structural walls is essential in typical wall-frame hospital buildings to ensure Occupiability performance after earthquakes. Furthermore, Column-to-Beam Strength Ratio of minimum 2 is also required in hospital buildings to remain Occupiable.

Abstract

We, humans, are social beings, and our communication is the most evolved and wellstructured form of a communication system that we are aware of. An essential aspect of human communication that helps humans bond faster and develop a sense of closeness is the use of humor. Humor that occurs as part of a conversation is known as conversational humor. Conversational humor is a type of humor that is unique and contrary to what it may seem like. It is more than just plugging canned jokes into a conversation. It requires the use of certain techniques and the presence of at least two interlocutors who understand the context of the conversation. The first step towards understanding conversational humor is to identify the different types into which it can be categorized and the techniques that are used to generate each type of humor. Current studies on this front either consider only a subset of these types and techniques or are domain specific. To tackle these challenges, we first propose a hierarchical annotation schema which allows us to get a comprehensive overview of conversational humor. For this task, we use a famous Telugu play, Kanyasulkam, and consider humorous utterances from this play as the dataset. This schema includes tags for type, technique, and benignity and considers cultural nuances in the text, making it an extensive schema for conversational humor. Further, to test the universality of the schema, we built a dataset of a different domain (Covid-19-based humor) and language (English). This dataset was annotated using a part of the annotation schema containing the type and technique tags. Two more tags viz. “Situation” and “Relevance” were added in the schema to help make the dataset more valuable as a standalone dataset which can be used by researchers from other fields like marketing, sociology, etc. The effectiveness of this dataset is tested with the help of various experiments for binary as well as multi-label multi-class classification using state-of-the-art ML models including but not limited to BERT, RoBERTa, BerTweet, etc. Based on the accuracy and analysis from the experiments, we can show that the annotation schema is universal in terms of language and domain. Such a classification of data can be used to accelerate the annotation process for humor data, and this annotated data can be used for various purposes like marketing, connecting with a target audience based on the relevance tag, and aiding research in the field of conversational humor for building humorous chatbots, and more human-like interactive systems.

Abstract

This thesis presents a comprehensive survey of blockchain layer 2 scaling solutions and investigates the phenomenon of market manipulation within the Bored Ape Yacht Club (BAYC) non-fungible token (NFT) collection. The emergence of blockchain technology has revolutionized various industries, and NFTs have gained significant attention due to their unique properties. However, scalability remains a major challenge for blockchain systems, limiting their widespread adoption. Layer 2 scaling solutions have emerged as potential remedies to address this scalability issue. The primary objective of this research is to provide a thorough analysis of the different layer 2 scaling solutions available in the blockchain ecosystem. A comparative study is conducted, assessing their strengths, weaknesses, and overall effectiveness in improving transaction throughput and reducing fees. The survey encompasses various approaches, including sidechains, state channels, and off-chain computation protocols. Each solution is evaluated based on criteria such as scalability, security, decentralization, and interoperability. The findings provide valuable insights into the current state of layer 2 scaling and guide future developments in this field. Additionally, this thesis delves into the BAYC NFT collection, which has gained significant popularity and market value. However, concerns regarding market manipulation within the BAYC ecosystem have arisen, raising questions about its fairness and integrity. Therefore, the study investigates potential market manipulation practices that may occur within the BAYC NFT market. It explores various forms of manipulation, including pump-and-dump schemes, wash trading, and front-running. The analysis employs data-driven methodologies, including statistical analysis and pattern recognition, to identify and understand such manipulative activities. The data for analysis is collected from reputable blockchain data sources, such as Etherscan and Moralis, including on-chain transaction records, smart contract events, and market trading data. The results of this research contribute to the existing body of knowledge in the fields of blockchain scalability and NFT market analysis. The survey of layer 2 scaling solutions provides a comprehensive overview for practitioners and researchers, facilitating informed decision-making when choosing appropriate solutions for blockchain applications. Furthermore, the investigation of market manipulation practices within the BAYC NFT collection raises awareness about potential vulnerabilities in the NFT market and suggests measures to mitigate such risks. Overall, this thesis advances our understanding of blockchain layer 2 scaling and market manipulation in the context of NFTs. It provides valuable insights into the technical and economic aspects of these topics, paving the way for further research and development in the blockchain ecosystem

Abstract

During the past few decades, the semiconductor VLSI industry has distinguished itself both by the rapid pace of performance improvements, and by a steady path of constantly shrinking device geometries. MOSFETs upon downsizing, have met the world’s growing needs for electronic devices like computing, communication, entertainment, automotive, and other applications with steady improvements in cost, speed, and area. But the device density has led to increased power consumption. Hence, Leakage, along with propagation Delay are the key metrics to evaluate the performance of any digital circuit. In nano-scale technologies(<45nm), static power occupies a significant share in the total power budget thereby prioritizing the need for leakage reduction techniques. Performance optimization of CMOS based circuits gains more significance for nano scale technology nodes. Variations in operating parameters such as supply voltage, temperature etc. have profound effects on power-delays specifications. The non-homogeneity in process parameters at such scaled technologies hampers the yield of the final designs. Performance degradation over time is another important factor determining the lifetime and reliability of an IC. In this work we develop a framework of algorithms to optimize digital circuits for low power and high performance applications. A wide range of analysis such as sensitivity of the circuit, correlation between parameters etc has been performed to understand the functioning in response to transistor sizing of digital cells. The first stage of this work aims to optimize basic digital cells through transistor sizing using the proposed optimization algorithms like Pareto Harris Hawk optimization algorithm, Glowworm Swarm optimization algorithm, Strength Pareto evolutionary algorithm-II and Neighbourhood Cultivation Genetic algorithm. The resultant system design becomes robust to withstand fluctuations caused by process, aging and operating parameters apart from providing highly improved leakage-delay performances. The second stage of this work involves optimization complex circuits. We have proposed a framework to optimize the complex circuits by selectively replacing the basic cells with the optimized sizing. The proposed framework deconstructs a given circuit into its constituent basic cells. It does a wide range of analysis such as sensitivity, the correlation between process parameters, load analysis, screening of insignificant parameters, and mismatch analysis to generate robust sizing. The algorithms generate sizing that can be used at multiple instances across different circuits. Thus the total number of design variables remains within a limit even when the cell count increases across circuits. The proposed framework then reconstructs the circuit with these optimized cells to improve the power-delay front. While replacing the nominal with the optimized sizing for each cell, the tool identifies the load it has to drive, the path (critical/ non-critical) it is present in and its driver modules in the path. The basic cells are selectively replaced with the optimized sizing using the proposed techniques: Backward traversal replacement technique and Partitioning large basic cells. Results have shown a substantial reduction in leakage power and propagation delays in addition to minimizing human effort

Abstract

Pressure sensors are popular in a large variety of industries. For some applications, it is critical for these sensors to come in a flexible form factor. With the development of new synthetic polymers and novel fabrication techniques, flexible pressure sensing arrays are more easily accessible and can serve a variety of applications. As part of this dissertation, we demonstrate one such application of the same by developing a low-cost flexible writing pad and doing crosstalk analysis on sensors with similar working principles. Digital multimedia tools are becoming increasingly important as the world slowly shifts to an online mode. We present a low-cost, flexible writing pad that uses a 16×16 pressure sensing matrix based on the piezoresistive thin film of velostat. The writing area is 5 cm × 5 cm with an effective pixel area of 0.06 mm2 . A read-out circuit is designed to detect the change in resistance of the velostat pixel using a voltage divider. A microprocessor raster scans through the sensor pixel matrix to obtain a data frame of 256 numbers. This data is processed using techniques like squaring and normalising (S&N), Gaussian blurring, and adaptive thresholding to generate a more readable output. The writing pad is able to resolve characters larger than 2 cm in length. The flexible writing pad produces legible output while flexed at a bending radius of up to 4 cm. Such flexibility promises to enhance the usability and portability of the writing pad significantly. We noticed that the raw data produced by the writing pad had a lot of crosstalk which we were subsequently able to resolve using the algorithms mentioned above. Such crosstalk has been reported in literature multiple times and is common, especially for sensors of the crossbar architecture. Crosstalk, in a sensor matrix, is the unwanted signal obtained at a sensor pixel that is not directly related to the stimulus. This paper presents a novel approach towards quantifying the crosstalk characteristics of a sensor matrix. The method involves obtaining the normalized sum of all the neighboring pixel readings, weighted by their distance from the stimulated pixel. We have used this methodology to characterise the crosstalk for a 5×5 velostat-based flexible pressure sensing matrix that uses a crossbar electrode architecture. We characterized sensor matrices with three different pitch lengths: 3mm, 4mm, and 5mm. We observe the crosstalk values to lie between 0.032 to 0.17 across different weights and pitches, which indicates a low measure of crosstalk. We observed that the 5 mm pitch matrix has the least crosstalk, which was expected due to the larger spacing between the pixels. We also characterised the crosstalk for all 25 pixels of a 4 mm pitch matrix and found the mean to be 0.081±0.002, within the range from 0.0071 to 0.1656

Abstract

Precomputed Radiance Transfer (PRT) is widely used for real-time photorealistic effects. PRT disentangles the rendering equation into transfer and lighting, enabling their precomputation. Transfer accounts for the cosine-weighted visibility of points in the scene, while Lighting is usually a distant emitted lighting, e.g., environment. Transfer computation involves tracing several rays into the scene from every point on the surface. For every ray, the binary visibility is calculated, and a spherical function is obtained. The spherical function is projected into Spherical Harmonic(SH) domain. SH is a band-limited representation of spherical functions, and the order of SH decides the representation capacity of the SH (the higher the SH order better the approximation of a spherical function). The SH domain also facilitates fast and efficient integral computation by simplifying the integral into simple dot products and convolutions. The original formulation of PRT by Sloan et al. 2002 provides different storage requirements for the transfer—vectors in the case of diffuse materials and matrices in the case of glossy materials. Using matrices for Transfer representation makes it infeasible as the SH orders increase. The work of Triple Product Formulation by Ng et al. in 2004 extended the formulation to allow simple vector-based Transfer storage even for the case of glossy materials. Prior art stored precomputed transfer in a tabulated manner in vertex space. These values are fetched with interpolation at each point for shading. Since the barycentric interpolation is finally employed to calculate the final color across the geometry apart from the vertex locations, the vertex space methods require densely tessellated mesh vertices to obtain accurate radiance. Sometimes high-density(tessellated) meshes adversely affect runtimes and memory requirements. This is mainly observed in simple geometries with no additional detailing but still demanding higher triangle counts (e.g., planes, walls, etc.). The first work provides a solution by leveraging Texture space, which is more continuous than the Vertex space. We also added additional functionality to obtain inter-reflection effects in the texture space. While Texture space methods provide faithful results in meshes, they require non-overlapping, areapreserving UV mapping, and a high-resolution texture to avoid artifacts. In the subsequent work, we propose a compact transfer representation that is learnt directly on scene geometry points. Specifically, we train a small multi-layer perceptron (MLP) to predict the transfer at sampled surface points. Our approach is most beneficial where inherent mesh storage structure and natural UV mapping are unavailable, such as Implicit Surfaces, as it learns the transfer values directly on the surface. Using our approach, we demonstrate real-time, photorealistic renderings of diffuse and glossy materials on SDF geometries with PRT.

Abstract

Pseudowords are a part of language that are not translatable to another, as they have no meaning attached to them while also having the constraint of sounding like a phonologically valid sequence under the desired language’s native phonotactics. This thesis thus explores automated language-agnostic pseudoword generation, evaluation of them, and use of them outside psycholinguistics research and clinical use. As the thesis progresses, we highlight current research, draw inspiration from close topics of study, build a pipeline to generate pseudowords and generate Hindi and English pseudoword candidates for further experiemntation. We make this reusable pipeline available on a public repository, as one of the deliverables of this work. Then we show how the current evaluation work in this field is very scarce and sew an evaluation framework with reproducible details on how to design and analyse a human-inthe-loop experiment for something as tricky as pseudoword judgement, conducted for a layman native speaker. After showing various ways to prod a pseudoword set for quality, we compare notes against past sets in English and present observations summarising how comparable they are. However as there is no Hindi pseudoword dataset yet, we add in psycholinguistic features on top of results of evaluation metrics per Hindi pseudoword and release “Soodkosh” another fully public and usable for research resource. Finally, we conduct two separate studies involving pseudowords to show the application, impact, and importance of them across fields. The first study uses pseudowords to establish gradient between high-frequency words, low-frequency words, and non-sensical sequences of alphanumerics used as passwords. The aim of this study is to find correlation and its strength between the perceived security and memorability of a password/phrase. The other part of this chapter is an exploration into language models’ performance on Aphasia classification and if replacing pseudowords can help them. This is as pseudowords like neologisms, mis-pronunciations, and other novel forms generated by Aphasic speakers are largely out-of-vocabulary to a standard languge model that functions off of a pile of mostly well-formed and coherent data. As these are not directly helpful to the field of Aphasia, this work replaces one possible hurdle to see if it is a feasible solution. However the results show that pseudowords are passively used as features and cannot be replaced directly

Abstract

Smart cities leverage IoT to improve citizens’ quality of life by providing better infrastructure, enhanced transportation systems, and efficient public services. With IoT-enabled smart city applications receiving traction, mechanisms must be implemented to cater to the diverse requirements of the use cases. The increasing number of connected devices increases the risk of cyber attacks and breaches. Further, smart cities rely heavily on integrating critical infrastructure such as power grids and water treatment plants. Securing these systems is crucial to protect citizens and institutions from potential harm caused by attacks. However, securing the IoT ecosystem is a challenging task. There are constraints on processing, memory, and energy consumption in addition to interoperability challenges, cost, and complexity. Designing and implementing the appropriate security controls for the devices and services requires on-ground testing and analysis. This thesis explores and analyzes the security of IoT-enabled smart cities. The work consists of two main contributions. First, IoT security requirements, potential threats, and vulnerabilities to the various layers of IoT are analyzed. Vulnerability assessment and modeling of threats are performed on the air quality monitoring vertical of the smart city deployment of IIIT-H to gain visibility into the baseline security requirements. The recommendations and state-of-the-art solutions are extensible to applications that require Wi-Fi and mobile network-based security for large-scale IoT deployments. Second, the entire smart city deployment was examined, covering applications such as air quality monitoring, water quantity management, weather monitoring, and energy monitoring. With each application operating using different hardware components, communication technologies, and software dependencies, requirements such as interoperability, scalability, resiliency, and security are essential. This study focused on the provisions of the oneM2M standard in catering to these requirements and its role in smart cities. Experiments were conducted on the OM2M platform, an implementation of oneM2M, used in the smart city deployment of IIIT-H. The recommendations are a foundation for the security of oneM2Mbased smart city applications.

Abstract

In today’s digital era, biometric authentication has become increasingly widespread for verifying a user across a range of applications, from unlocking a smartphone to securing high-end systems. Various biometric modalities such as fingerprint, face, and iris offer a distinct way to recognize a person automatically. Fingerprints are one of the most prevalent biometric modalities. They are widely utilized in security systems owing to their remarkable reliability, distinctiveness, invariance over time and user convenience. Nowadays, automatic fingerprint recognition systems have become a prime target for attackers. Attackers fabricate fingerprints using materials like playdoh and gelatin, making it hard to distinguish them from live fingerprints. This way of circumventing biometric systems is called a presentation attack (PA). To identify such attacks, a PA detector is added to these systems. Deep learning-based PA detectors require large amounts of data to distinguish PA fingerprints from live ones. However, there exists significantly less training data with novel sensors and materials. Due to this, PA detectors do not generalize well on introducing unknown sensors or materials. It is incredibly challenging to physically fabricate an extensive train dataset of high-quality counterfeit fingerprints generated with novel materials captured across multiple sensors. Existing fingerprint presentation attack detection (FPAD) solutions improve cross-sensor and cross-material generalization by utilizing styletransfer-based augmentation wrappers over a two-class PA classifier. These solutions generate large artificial datasets for training by using style transfer which learns the style properties from a few samples obtained from the attacker. They synthesize data by learning the style as a single entity, containing both sensor and material characteristics. However, these strategies necessitate learning the entire style upon adding a new sensor for an already known material or vice versa. This thesis proposes a decomposition-based approach to improve cross-sensor and cross-material FPAD generalization. We model presentation attacks as a combination of two underlying components, i.e., material and sensor, rather than the entire style. By utilizing this approach, our method can generate synthetic patches upon introducing either a new sensor, a new material, or both. We perform two different methods of fingerprint factorization - traditional and deep-learning based. Traditional factorization of fingerprints into sensor and material representations using tensor decomposition establishes a baseline using machine learning for our hypothesis. The deep-learning method uses a decompositionbased augmentation wrapper for disentangling fingerprint style. The wrapper improves cross-sensor and cross-material FPAD, utilizing one fingerprint image of the target sensor and material. We also reduce vi vii computational complexity by generating compact representations and utilizing lesser combinations of sensors and materials to produce several styles. Our approach enables us to generate a large variety of samples using a limited amount of data, which helps improve generalization