Abstract

The Lok Sabha Elections 2019 in the ’s largest democracy, India, was a significant electoral event.These elections are conducted to elect representatives across the country for a period of 5 years and arealso key to the process of selecting the Prime Minister, the highest authority in the Union cabinet. Inkeeping with global trends, the Indian elections saw a very prominent use of Online Social Media bypolitical parties to create discourses in their favour. We focus our study on Twitter, collecting over 45Million tweets, tracking more than 3,500 hashtags and over 2,500 political handles while monitoringtheir network interactions.We begin with a study of major political trends in the country since its independence in 1947. We lookat political alignments and realignments in the country that led to a shift from a Single Party dominantpolitical system to a strong two-party system we have today. Our study focuses on national politicaltrends, and not on individual states. We thus study two political blocs - the National Democratic Alliance(NDA, led by the Bharatiya Janata Party) and the United Progressive Alliance (UPA, led by the IndianNational Congress).We did a comparative analysis of various political parties in India and their Social media presenceduring the 2019 General Elections. We also perform a social network graph analysis for the verifiedpolitical handles who were active during the 2019 Elections. This helps us understand their relationshipswithin and across the parties. We are also able to see a direct correlation between the node-size of apolitical handle in its cluster, and its hierarchy in the party. This follower-following network shows anunmistakable relationship between politicians at an individual node level, and between the parties thatsupport each other at the cluster level.We then move on to analyse the nature of political discourse that parties attempted to propagate in theonline sphere. We analyze the Manifestos released by both the major political parties - BJP and INC. Weanalyze tweets collected during the 2019 elections to see how narratives were shaped and altered over time. We study the narratives propagated by the party already in power and how they were supportedor challenged by other parties. Spanning over 5 months, January to May 2019, we analyzed monthlychanges in the political messaging promoted by leading political parties and leaders. We then speculateon the impact of these changes on existing narratives during the campaigning and the elections.We further move onto a comparative discourse analysis between the 2014 and 2019 General Electionsin India. We begin by understanding how these elections were fundamentally different with regardsto the incumbency factor and various other politico-economic factors. Given the contrasting politicalscenarios, we observe how the ruling and the opposition party brought a change in the discourses duringthe two elections. We compare our results with the on-ground surveys (or the Mood of the Nation)conducted by Lokniti, a research programme run by the Delhi-based CSDS, across the nation.This thesis helps in getting a clear picture of the activity in the online eco-system during the 2019General Elections. The discourse analysis helps in understanding how the conversations were shaped onthe web and what impacts it had. Understanding the data and numbers from a socio-political perspective,we see how narratives that were formed online, steered the offline world and vice versa. The networkand social media analysis of the political handles aids in understanding how the political parties in Indiadrove their online battles. Adding on to a study conducted during the 2014 General Elections, we alsocontrasted the 2019 and 2014 General Elections in India.

Abstract

Recent advances in the semiconductor industry for the internet of things have increased the demand for novel, fast and robust circuits. The thirst for high-processing ICs is ever increasing. This has pushed the demand for high data rates in wireless and wireline communication systems in the multi-Gbps range. For high-speed transmission of data, there arises a need for high-speed on-chip clocking circuits making the use of Phase-Locked Loops (PLLs) crucial. Phase-Locked Loops are used for synchronization, synthesis of clock ,and jitter reduction in mobile or wireless communications. For demodulating a signal and recovering it to its original input signal, the PLL circuit is used. PLL has wide application in high frequency ranges (Giga Hertz) as crystal oscillators have a limited operating frequency range in the oder of MegaHertz. The current reference is an indispensable circuit in analog, digital, and power electronic systems. Since it is easy to implement a voltage reference by the bandgap circuit, current references are derived from voltage references by voltage to current conversion. However, this conversion may take more silicon area since it contains transistors, operational amplifier, and resistors. In this thesis, we present a robust on-chip clock generator circuit using a power efficient phase frequency detector and a low current mismatch dual adaptive regulated charge pump. Another technique for charge pump has been proposed to implement PLL using Dynamic Voltage Frequency Scaling (DVFS) scheme. To manifest the effectiveness of this DVFS, a 10-bit Successive Approximation ADC (SAR ADC) is built. The current reference circuit is proposed which is robust to process, temperature, and voltage variations. Ring oscillator and temperature sensor are built using the proposed current reference circuit. The simulations of all circuits are done by using the Cadence Virtuoso tool in the 180nm UMC MPW RF process.

Abstract

The word ‘Cloud’, in its traditional term, gives the illusion of immense and infinite capacity. This has been made very true by some of the leading public cloud providers. However, it is the subscription based, on-demand and flexible nature of the Cloud that attracts organizations. The sheer thought of running capacity comparable to a super-computer, spawning the entire infrastructure from scratch in minutes and paying at an hourly rate, is powerful. With virtual resources ready to start on the press of a button, cloud gave birth to programmable infrastructure. Gone are the days when only shell scripts were used to configure servers. Now an entire data center could be deployed using a few lines of code. This is so powerful and tempting that organizations are opting for a cloud-like model for handling their in-house resources, giving birth to the private cloud. Both public and private clouds offer great opportunities to start from scratch, but this is not an option for organizations which have been running their services on traditional infrastructure for years. Many organisations are already migrating their services and applications to cloud. This involves a lot of decision making and resources, both in terms of time and money. With plenty of public cloud providers available currently, choice isn’t straightforward. Apart from a few common features, each provider has offerings specific to their platform. With differences in APIs, features, cost etc. selecting a specific cloud becomes a long term commitment for the company. Due to inherent heterogeneity present in the cloud space, coming up with a unified solution which is compatible across all providers isn’t cheap and at times not even feasible. Our thesis focuses on designing and building solutions to facilitate migration of applications from a traditional setup to Cloud and enable a subset of applications and workload to use Multiple Clouds - across public providers or hybrid environments. A lift and shift strategy for migration could result in disruption of services. An approach to migrate a group of processes from one host to the other without ever stopping them is called live migration. To minimize the disruption and performance of services, we propose two live migration strategies - inter-hypervisor and inter-cloud. The former enables an organization to migrate their Virtual Machines to different physical machines allowing them to easily move to a Private Cloud. The latter allows live migration of containers across different nodes. Due to no dependence on hypervisors, the latter offers a solution to move to a public cloud. Live migration has a lot of inherent challenges and a majority of workloads don’t have such stringent downtime requirements which demand live migration. Not to mention that most production applications are clustered and have multiple replicas to be Highly Available. In such cases, it would be simpler and preferred to spawn a new process in the remote location and failover instead of live migration. Ideally, we should be using frameworks which support multiple clouds and migration of an application to a new cloud should simply be increasing its deployment scope. Our work has been focussed on building frameworks which leverage multiple cloud providers out of the box and reducing the heterogeneity between cloud providers. We’ve developed frameworks and prototypes for few standard applications and platforms to run them seamlessly across multiple cloud environments. To that end, we propose the design of MultiPaas, a platform capable of running traditional web applications across multiple cloud providers. We target the fundamental problem associated with running a PaaS across multiple clouds and design this as an extension to an existing PaaS, Deis. We built a platform, MultiStack, which enables Big Data frameworks to run across multiple public and private cloud providers. It currently supports Hadoop, a popular Big Data solution, and can be easily expanded to support others. It supports both EC2 API, allowing seamless adoption of AWS, Rackspace, Openstack and Eucalyptus cloud solutions

Abstract

Soft robotics is a sub-branch of robotics in which nature is an inspiration to develop robots with physically flexible bodies to mimic the movements of living organisms and their complex structures. All soft robots require an actuation system to generate forces that allow for the movements. In the field of soft robotics and its applications, Dielectric Elastomer Actuators (DEAs) are becoming popular as biomimetic (pseudo-muscular) actuators. The key working principle of DEAs is actuation based on the electrostatic pressure experienced due to applied voltage. Since the DEAs are electromechanical transducers, they must contract and stretch along with the polymer. Thus, the electrodes should be complaint. To get this desired stretch-ability, stretchable electrodes based on carbon nanotube (CNT) are used so that even at large strains, the electric conductivity of the electrode is retained. The high driving voltage of 0.3-4kV is one of the main drawbacks of DEAs. In order to reduce the operating voltage of the actuator, the relative permittivity can be increased without reducing the thickness. On the other hand, the thickness can be decreased to reduce the driving voltage, and a multiple-layered structure can be used to have the same actuation force. CNT based thin film electrode consists of a mesh of extensive network of interwoven CNTs. These networks of CNTs slide on each other when they get strained, increasing the dimensions of the thin film. However, this interwoven structure of CNTs does not resemble a continuous thin film but resembles a perforated one. This perforated thin film-like structure can lead to intricate patterns and distribution of the electric field, especially for multilayered DEAs. Since DEAs are inherently capacitors, they can be represented in the capacitor/resistor model. In the following work, the study of the electrostatics of such multilayered structures is carried, focusing on screening or lack of screening of electric fields generated by adjacent capacitors and the corresponding effect it has on the overall capacitance of the multilayered system. From the result of theoretical analysis, it has been observed that for a multilayered interdigitated structure with uniaxial perforation of the thin film electrode, the capacitance of such a system can be calculated by the following equation: 𝐶𝑛 = 𝐶0 [𝑛 − 2(𝑛 − 1)𝛽 + 2(𝑛 − 2)𝛽 2 − ⋯ + 2(−𝛽) 𝑛−1] Where Co is the capacitance of all solid electrode single layer system, β is the constant dependent on the perforation ratio of electrode, and n is the number of layers of the structure. Further, the result from the above theoretical analysis is verified by modeling the structures in 2D and 3D space using Finite Element Analysis (FEA). Further to get the multilayered structure, the analysis is carried by modeling them as an interdigitated structure, and the corresponding capacitance is obtained. From further analysis and their results, it has been discovered that β, which is a constant dependent on the perforation ratio, is independent of the number of layers of the structure. This acquired data closely matches the results of theoretical analysis obtained for normalized capacitance. Normalized capacitance is determined to show the change of percentage with that of an all-solid electrode-based capacitor. The loss of the electric field due to the discontinuity caused by perforations of the electrode in an interdigitated structure is minimal for a reasonable coverage area. Further, it can even be neglected for coverage of more than 90% in the area. The above analysis and the corresponding modeling give an estimate on the impact due to perforations and is essential for the development and manufacturing of reliable and robust CNTs based DEA structures.

Abstract

Buildings on hill slopes are highly vulnerable due to its less resistive lateral load capacity. These type of buildings are unsymmetrical and are of in irregular configuration. There are variations in the column height which gives eccentricity to the structure as the centre of mass and centre of rigidity doesn’t coincide. This irregularity leads to its torsional behavior and gives us the necessity for analysing these buildings for its base conditions and its behavior by changing its plan configuration. When lateral load is applied in plain building along one direction, it is observed that middle frames will take 30% of the loads and corner frames will take 20% of loads but this kind of distribution is not observed when lateral load is applied along the ridge direction (non- valley direction) on a building of hill slope. The building considered are step-back building. In these buildings the heights in the valley direction increases and all the floors points (connection with columns) are aligned in a plain parallel to ridge direction (non-valley direction). Two types of buildings are used, the first is Bare frame and then bracings are used to structure to check the effect of change in stiffness of frames on Twist, Axial Force on the road level Beam and on the load distribution. This thesis contains the calculation of Twist which is an effective parameter to decide the behavior of the building, Axial forces on the members that can lead to collapse in the structure and about the twist variation with change in the aspect ratio and with change in stiffness of the structure so that the suitability of the configuration can be known of the building. The building on hill slopes does not behave similar to the buildings which are on plane ground. Change in the Aspect ratio means changing the length of the building in ridge direction. Changing the length in valley direction will disturb the load path of gravity load for each and every change in the structure and after some increment of the length in valley direction will fail the structure for its gravity load path. So the length of the building can’t be extended along the valley direction but changing the length along ridge direction does not effect that much to the load path of gravity load and it can be obtain a best aspect ratio of the structure for that region. Change in the stiffness of the building means changing the stiffness of the frames and making the building stiffer. The more frames will be stiff the more buildings will be stiffer, stiffness of frames are considered so that the translation of the frame and the rotation of the floor can be minimized. The aim of the thesis is to produce a building on hilly slopes that will be less torsionally active by a procedure. The procedure may be applying a bracings or changing the aspect ratio or both in the structure and finally conclude with data’s and results by analysis.

Abstract

The social and political shifts being perpetuated by the Internet and social media are dominating contemporary discourse around politics, society and technology. The difficulties in regulating online practices, in advancing Internet literacy and moderating machine learning algorithms that govern online behaviour, and moreover in preventing malicious users from infiltrating the Internet have been highlighted by scholars in this domain. As the Internet and social media have become a ubiquitous and inseparable part of our lives, so have the social shifts brought on by them. The influence of modern technology on human behaviour, both personal and social, poses a great challenge to one of the most fundamental democratic principles; human free will. This thesis highlights the anti-democratic phenomena emerging out of the incessant use of the Internet in daily life, while also looking at precursors for these phenomena through history. This thesis employs formal analysis using natural language processing as well as actor-network theory to comprehend some of these phenomena. It also connects various theories proposed by thinkers like Kierkegaard, Chomsky and Foucault, and ideas expressed by researchers, authors and journalists in recent years, to sufficiently understand these phenomena. Through the discussions and analyses presented in this thesis, the dynamics of the social ills posed by the Internet can be better understood, and this can inform further research as well as policy-making aimed at combating these issues

Abstract

Hypernasality is a prominent characteristic of speech production observed in subjects with Cleft Lip and Palate (CLP). In the clinical diagnosis of individuals with CLP, it is essential to identify and assess the hypernasality in speech. The assessments techniques are invasive, which affect the natural way of speech production. So automatic objective assessments have gained popularity as they are repeatable, easily accessible, and cheap. Nevertheless, several works in literature that try to address this problem and try to extract fine features which capture the effects of hypernasality on the speech signal. However, almost all the current works try to address the problem in individual vowel sound units. In contrast, this work tries to address the problem of detection and assessment of hypernasality from continuous speech which involves complex articulatory movement. It is similar to the assessments carried out by speech and language pathologists. The hypernasality is a result of an abnormal vocal tract, where the air leaks through the clefts into the nasal tract, resulting in nasalization. The nasalization alters the spectrum of the speech signal in different ways depending on the configuration of the vocal tract. So accounting all different changes for all possible sounds is impossible. Generally, the nasal tract is different from oral tract as it is longer and has no moving articulators. This thesis proposes Single Frequency Filter Bank Based Long Term Average Spectral (SFFB-LTAS) features that enhance the contrast between nasal tract an oral tract for detection and assessment of hypernasality from continuous speech. Further, the impact of nasalization on excitation source has been studied. The experimental results show that the information of hypernasality is manifested even in excitation source evidence and it is complementary to vocal tract features. The results show that both the excitation source and vocal tract system features are useful for the evaluations. We also present a new database, called New Mexico Cleft Palate Center- Cleft Lip and Palate (NMCPC-CLP) database. It consists of continuous speech utterances from CLP subjects and normal subjects collected at New Mexico Cleft Lip and Palate Center.

Abstract

The advancements in hyperspectral remote sensing escalated the research in forestry-related to species identification, discrimination, and health assessment. The technology aided in the development of a spectral library of various species that enable researchers to map species without the need foran expensive field inventory. The research focus of this thesis is to generate a temporal spectral database for ten selected species of Araku forest, the Eastern Ghats using field inventories (multiple seasons), hyperspectral satellite data, and canopy upscaling models. The properties of leaf/bunch of leaves sampled from the bottom of trees may vary, and a direct analogy with canopy properties will lead to substantial overestimation/underestimation of the spectral reflectance of the tree.Hitherto researchers applied upscaling techniques with wellknown PROSPECT, SAIL, FLIGHT models, but so far, no model (in forestry) developed for leaf upscaling, within the country using field-collected leaf reflectance spectra and their biochemical parameters. Towards this objective, the current study proposes an upscaling model “SEMULS” (Semi-Empirical Model for Upscaling Leaf Spectra), a unique and first of its kind of model to upscale leaf spectra to canopy. The spectra extracted for the selected species from satellite data are correlated with canopy spectra derived from leaf reflectance spectra obtained from the field using SEMULS. The outputs of the model are validated for its applicability by comparing with upscaled spectra derived from popular PROSPECT and DART 3D models. The SEMULS model is unique in retaining the minute dips and peaks of the spectra, which are important in discriminating (80% matching at 5deg or less threshold) species, unlike other models that smoothen spectra using different filters. The upscaled canopy spectra (SEMULS) is further used to classify selected species in the study area from CHRIS-PROBA satellite data, using the SAM method and validated based on GPS locations observed from the field. The study also proposed few experiments such as upscaling spectra for chlorophyll region without field inventory or upscaling model, mapping species - from the spectra extracted from literature and use of spectra of one location to map same species in other regions, discriminating species using DART model and an algorithm to decompose Sentinel data to hyperspectral data. The developed database enhances the reliability and comparability of the datasets allowing multi-temporal and multi-scale studies. The database is uploaded in the NISA portal (spectral library) which will serve present and future researchers, in identifying the species without field inventory,and/or for any spectra related research.

Abstract

The English word ‘daedalian’, which means ‘ingenious, intricate, and confusing’, is a word that originates from the Greek mythological character Daedalus, who is believed to have built a labyrinth so complex, that he himself couldn’t escape its confines- but you would be forgiven for not knowing that. Daedalian is not a commonly used word, and its synonyms like ‘complex’ and ‘confusing’ are favored in day-to-day speech. But like the title of this thesis shows, rare words are still found in written text and this creates a need to simplify this text to make it accessible to the layperson. In this thesis, we study the problem of text simplification, the challenges it presents, and the different ways these challenges can be overcome. This thesis touches upon various topics along the vast expanse of text simplification research, starting from generic improvements to encoder-decoder models that are used for controlled Natural Language Generation (NLG) tasks (including text simplification), to domain-specific text simplification challenges and the specialized solutions that are required to solve them. Recent advances in deep learning and NLG research have enabled sophisticated text simplification models. The central idea behind most approaches to text simplification, and indeed most other controlled text generation tasks, is the encoder-decoder model. Thus, we believe that controlled text generation tasks like text simplification stand to gain by using improved encoder-decoder models. Encoder-decoder models have been vastly improved by the introduction of the attention mechanism, and architectures like the transformer, but there is scope to further improve these models by explicitly improving the encoder outputs. We do this by using a co-attention based loss in addition to the standard cross-entropy loss to train the model. We evaluate this proposal on multiple tasks and show that this not only benefits text simplification but also improves performance in adjacent tasks like style transfer and sentence compression. Another problem with current text simplification research is the lack of suitable datasets that are representative of real-world text simplification needs. We enumerate the problems with existing text simplification datasets and propose a new dataset that comes from diverse sources, has a richer vocabulary and is more representative of real-world text simplification tasks. We present various approaches to simplify text from our proposed dataset and recommend methods that can be used to make the generated text more readable and semantically similar to the source. We also study domain-specific text simplification, with a specific focus on the medical domain. We study the unique challenges that come with simplifying medical text and propose an unsupervised approach to solve the problem of medical text simplification, which goes a long way towards overcoming the challenge of the lack of parallel data in the medical domain.

Abstract

Positive counterions, such as metal cations, are necessary for the electrostatic neutralization of the polyanionic ribonucleic acids (RNA) as the negatively charged RNA backbone folds onto itself to adopt a 3D structure. Magnesium ions are the most frequently observed metal ion in experimentally determined RNA structures, and have been noted for their influence on the stability, dynamics and function of RNA molecules. Identifying motifs at magnesium-RNA sites is necessary to better characterize the role played by the ions in RNA molecules. Previously available tools to detect magnesium-RNA motifs are either based on manual annotation or rigidly defined ion-RNA binding characteristics that cannot be used to describe many recently identified motifs. In this work, a set of methods have been developed to help detect, characterize and discover magnesium-RNA motifs that overcome the limitations of previously available tools. A graph-based querying system “MagMotiRNA” has been developed to query for welldefined motifs in literature as well as user-defined patterns at magnesium-RNA sites. In keeping with the evidence regarding variability in magnesium-RNA binding patterns and the role played by magnesium in stabilizing certain inter-nucleotide contacts, the tool allows for defining queries with fuzzy ion-RNA binding descriptions and queries involving the local RNA structure at the site. The results from MagMotiRNA were benchmarked against MgRNA tool, the prevalent magnesium motif detection tool, and the discrepancies in the results could be largely attributed to the higher sensitivity in the hydrogen bond detection method used in MagMotiRNA. The occurrences of five distinct base-coordinated magnesium-RNA motifs in a redundant dataset were analyzed and the distinct structural contexts in which they occur were studied. The case study revealed that the ion-RNA binding patterns at these contexts are often variable, and hence it may not be appropriate to define magnesium-RNA motifs solely based on ion-RNA binding details. Instead, a bottom-up approach is proposed for defining motifs based on the invariant features in the local RNA structure at these sites that are indicative of the putative role played by magnesium ions. A set of heuristics has been developed as part of this work whereby conserved magnesiumRNA sites can be detected in an RNA molecule by analyzing multiple structures of the molecule. Building upon these heuristics, a metal-RNA binding profile was created for 19 RNA molecules by transposing all metal-RNA sites associated with these molecules in a redundant dataset onto a representative structure. The rather complex problem of studying magnesium-RNA sites given the sparsity of reliable experimental data would benefit from these profiles of frequently conserved metal-RNA sites provided by the MagMotiRNA-Atlas database. The Atlas equips users to perform top-down exploration of magnesium-RNA motifs, and can act as a complementary tool to the bottom-up analysis of magnesium-RNA sites developed in the other MagMotiRNA tools.