Abstract

In drug discovery applications, high throughput virtual screening exercises are routinely performed to determine an initial set of candidate molecules referred to as “hits”. In such an experiment, each molecule from large small-molecule drug library are evaluated for physical properties such as the LogP, Topological Polar Surface Area (TPSA), and docking score against a target receptor. In real-life drug discovery experiments, the drug libraries are extremely large but still a minor representation of the essentially infinite chemical space , and evaluation of physical property for each molecule in the library is not computationally feasible. In the current study, a novel Machine learning framework for Enhanced MolEcular Screening (“MEMES”) based on Bayesian optimization is proposed for efficient sampling of chemical space. The proposed framework is demonstrated to identify 90% of top-1000 molecules from a molecular library of size about 100 million, while calculating the docking score only for about 6% of the complete library. We believe that such a framework would tremendously help to reduce the computational effort in not only drug-discovery but also areas that require such high-throughput experiments.

Abstract

Today’s society cannot be visualized without the dependency of humanity on technology. Every walk of life in human survival today is linked to computers and electronic systems, which undoubtedly improve the quality of life. This comfort comes with a restriction because of the vulnerability of the electronic system to anti-electronic threats. Not orienting towards offensive, it is indeed the need of the hour to protect our systems by characterizing their susceptibility levels. Many researchers formulate the threat scenarios, and international committees constituted for this purpose. High power electromagnetic (HPEM) systems that could generate and radiate few gigawatts were developed worldwide. Likewise, an antenna also plays a crucial role in the HPEM systems in any transmitting system. However, there are more challenges in designing and developing antennas for HPEM systems. Among the categories of antennas developed for this purpose, wideband/Ultra-wideband (WB/UWB) antennas that are capable of transmitting transient impulse-like high voltage signals are potential contenders. A well-designed WB/UWB antenna should have high fidelity in transmitting the transient signal without altering its characteristic parameters. This thesis focuses on the design, analysis, development, and measurements of a wideband antenna, spark gap integrated dipole (SGID) antenna. The main focus is placed on transient analysis, its construction, and performance evaluation. Spark gap integrated dipole (SGID) are simple in construction, compact, and reliable transient wideband antennas for high power applications. They offer the promise of utilizing them to very high powers without any complications and can easily be fabricated with low-cost materials. This makes them a viable alternative to complex horn-based antennas that are generally used in high-power scenarios. Despite the aforementioned advantages, the bandwidth offered by these antennas is limited to mesoband, whereas Transverse Electromagnetic (TEM) based horn antennas can scale up to hyper bands. As a result, their use is limited to Directed Energy Weapon (DEW) applications that demand damped sinusoidal radiation and, to some extent, in UWB radars. While horn-based antennas have matured technology, both experimentally and theoretically, only a few researchers have reported the analysis of spark gap integrated dipole antennas. In this thesis, we mathematically modelled the SGID antenna based on Maxwell’s equations, and a theoretical transient response is generated. The performance of SGID is thoroughly analysed using computational electromagnetics for its transient response. The physical dimensions of the dipole are optimized for maximum radiation efficiency. The spark gap efficiency was also studied by evaluating the performance for various electrode profiles. Modifications are proposed to the existing designs in order to enhance the performance and increase the figure of merit. This includes optimal dimensions visà-vis input signal parameters which help not only for the maximum radiation but also its spectral occupancy. In addition to this, the selection of Bruce’s profile has ensured uniform electric field distribution and has allowed using the dipole for higher voltages. To further enhance the reliability of the antenna, hydrogen gas is selected as the spark gap medium due to its high recovery rate property. These proposed methods are mathematically formulated, simulated, and analyzed, and a considerable improvement in the overall SGID performance is reported. The theoretical analysis is validated by measurements by subjecting the antenna to a 100kV pulsed input voltage. A logical measurement procedure is proposed that removes the ambiguity in measurements as the radiation from the spark gap antenna is highly probabilistic. The proposed design and development methodology of the SGID has resulted in an efficient transient radiator with a figure of merit (FoM) significant compared to those reported in the literature.

Abstract

There are several types of spatio-temporal dynamic processes, both naturally occurring and manmade like climate change, floods or droughts, crime-spread etc; and some which are influenced by multiple factors like agricultural crop performance, occurrence of diseases, etc. It is important to understand the behaviour of such phenomena as they have a significant effect on the places where they occur, having both direct and indirect impacts on the environment as well as the population in that region. These phenomenon have an intrinsic spatial as well as temporal context associated with them. In order to gain insights into these processes and their relationships, there is a need for a methodology that can capture the context in both the dimensions- spatial and temporal, simultaneously. At a particular location and time, the observed output value of a dynamic phenomenon can be affected by many causative factors that feed as an input to the process and are at play simultaneously. While many studies look at the correlations of input to output to understand the phenomena, it’s also important to understand the outcome exhibited with its intrinsic spatio-temporal context, deciphering which may help us better appreciate the factors at play, both known and unknown dynamics of the factors themselves. Hence there is a need to highlight these patterns as seen in the observed data and provide insights into the process that govern these phenomena. Different spatio-temporal statistical modelling methods have been developed, that try to approximate the process and produce a mathematically-formalized model that would generate the spatio-temporal data. However, such modelling techniques are more focused on accurate estimation of input parameters and finding a suitable model that fits the data. They often end up ignoring the spatially correlated parameters, or parameters with a weaker influence on the phenomena, thus affecting the overall predictive performance of the model. Alternatively, data mining approaches are focused on uncovering and describing unknown patterns and trends in the data. However, in the case of spatio-temporal datasets, the direct application of classical data mining algorithms is not suitable, since they are intended to work on transactional-like data which has independent attributes. The attributes in spatial data are not independent, and there is an implicit interaction between them, which cannot be captured directly by the classical data mining techniques. Therefore, the conventional data mining methods fail to effectively capture the implicit spatial or temporal characteristics of the geographic phenomenon and are unable to exploit the rich spatial temporal relationships/patterns embedded in the datasets. In this thesis, we make use of a generic Spatio-Temporal data mining algorithm called MiSTIC, as part of a generalized framework called MiSTIC++. Using MiSTIC, the goal of the framework is to segregates the entire area of analysis into different zones, where each zone is comprised of contiguous spatial samples. The method tried to create the zones in a way such that within each zone, the spatial interactions among its comprising elements remains preserved over time, thereby capturing the consistent spatio-temporal context, which cannot be observed directly from the data. These spatial interaction exist between two neighboring regions, the directionality of which remains the same across time, despite the changes in magnitude. There might be several input factors that are responsible for bringing a net change in the observed value. However, in spite of the various factors that are at play, the interactions between these individual spatial units remains very similar across time. The MiSTIC algorithm tries to capture this property of preserved spatial interactions, based on only the observed output data and the neighborhood information. The knowledge of such stable zones and their intra-zonal analysis helps understand the dynamic phenomena better, enabling us to respond better to the situations and to take better decisions with respect to the process being analyzed. The utility and generality of the proposed method has been demonstrated with the help a case study of crop yields in India. The study uses the district level yield dataset of Paddy and Maize in Kharif season across India for the period 2000 to 2010. The method helps cluster a set of districts into zones, with each zone further divided into nearly concentric sub-zones of similar performers based on the empirical rule. The results for paddy in the Kharif season show that there are 18 production zones across India and 23 for Maize, except the districts of North East and Jammu and Kashmir, with zonal averages ranging from 861.9 kg/ha to 3076.3 kg/ha for Paddy. MiSTIC is further extended, in this research, to indicate regions and sub-regions that are consistently high and low performers across the years, irrespective of the input simila

Abstract

Language is the most important medium of communication and education. In a multi-lingual country like India where there are 22 scheduled languages, it is a matter of national policy to endorse and promote the development of all these languages and reduce the language barrier. Giving universal access to information to people in their language is the primary concern and need of society. In this technological revolution, access to all information and communication technology (ICT) and Human language technology (HLT) will make it convenient for people to gain all the benefits of this. Though there are several NLP Tools and approaches available, due to the scarcity of good quality benchmark data, the development of practical usable systems is hampered. Machine Translation (MT) systems play a major role in bridging the language gap. However, there are several pit-falls in the MT due to various linguistic, cultural, and social reasons. Hence, there is a need for an Accessor system where there is a guarantee of faithful translation. In this thesis, we have proposed and implemented an architecture to develop a usable system called ’Anusaaraka.’ The use of an expert system to develop this Accessor + MT system is unique and helpful for the common man to contribute to further improvement of the system. As discussed without proper quality linguistic resources, no usable NLP applications can be developed. To address this issue, we have proposed a unique concept of ’mutual-bootstrapping’ where various linguistic/NLP tools interact with each other to further enhance/boot-strap the development of highquality linguistic resources. We also demonstrated the proposed model with the task of ’word/phrase alignment’ of a parallel corpus by building an ’Alignment Tool’. Methods to build high-quality linguistic resources like bilingual dictionaries, multiword expression dictionaries, treebank, domain dictionaries, technical dictionaries, etc. through mutual interactions by the Alignment Tool are discussed. Neural network based Machine Translation (NMT) systems are now producing state-of-the-art results. But still, there are errors and idiosyncrasies in the output. We have applied the ’mutual-bootstrapping’ approach to improve the accuracy of the NMT. Through direct pre-processing and post-processing techniques using the ’Alignment tool,’ the boot-strapping of NMT to produce high-quality translations will be discussed in detail. Throughout the thesis, we have emphasized the concept of ’mutual-bootstrapping’ in developing various NLP Tools/resources. The results of the tools are compared with other state-of-the-art results and the error analysis is done

Abstract

With quadrotors becoming immensely popular in applications such as relief operations, infrastructure maintenance etc., a key control design challenge arises when the quadrotor has to manoeuvre through constrained spaces during various operational scenarios: for example, inspecting a pipeline within pre- defined velocity and space, dropping relief material at a precise location under tight spaces etc., under the face of parametric uncertainties and external disturbances. To tackle such scenarios, a controller needs to ensure a predefined tracking accuracy so as not to violate the constraints while simultaneously tackling uncertainties and disturbances. However, state-of-the-art controllers dealing with constrained system motion are either not applicable for an underactuated system like quadrotor, or cannot tackle system uncertainties under full state constraints. This work attempts to fill such a gap in literature by designing Barrier Lyapunov Function (BLF) based robust controllers to satisfy multiple state-constraints while simultaneously negotiating parametric uncertainties and external disturbances. The superiority of the BLF control method over a typical unconstrained controller is demonstrated, followed by a ro- bust control design to satisfy position and orientation constraints on quadrotor dynamics. Finally, full state-constraints on a quadrotor(i.e., constraints on the position, orientation, linear velocity and angular velocity) are satisfied with robust control. For each control design, the closed-loop system stability is studied analytically and the efficacy of the design is validated extensively either via realistic simulation scenarios or via experiments performed on a real quadrotor.

Abstract

During Covid-19, we have seen experts from various fields come together and collaborate for collective social good. NLP and AI have been employed to create a positive impact in recent years, especially in public health research. In the work-from-home world, 500 thousand tweets are posted every day and 4 petabytes of data is created on Facebook1 . This plethora of data can be utilized for online biosurveillance to early insights into potential future epidemics. Today, smoking tobacco is one of the leading causes of preventable death. It causes more than 8 million deaths per year worldwide 2 . Real-time monitoring of public sentiment and trends helps us understand the gravity of potential health threats like smoking tobacco and help create necessary cessation and preventive campaigns. Contrary to previous tobacco studies, we have taken the language of the demographic affected, street words, and colloquial slang related to smoking into account. We released the smokeng dataset, a general tobacco-related dataset consisting of 3144 tweets along with a comprehensive annotation schema. Each class is created and annotated based on the content of the tweets such that further hierarchical methods can be easily applied. While cigarette smoking went down among high school students from 2011 to 2019, the number of students using e-cigarettes rose from 3.6 million to 5.4 million 3 . As the e-cigarette problem has exponentially increased, the need to treat smoking products individually presented itself. The target audience varies and might use different language and sentiments to describe the product they use. We extended Smokeng on tobacco product and product sentiment axes to release the SmokPro and SentiSmoke-Twitter datasets along with a comprehensive annotation schema for identifying tobacco products and their respective sentiments. Contemporary tobacco-related studies are primarily concerned with a single social media platform while missing a broader audience. Moreover, they are heavily reliant on labeled datasets, which are expensive to make. We explore sentiment and product identification on tobacco-related text from two social media platforms. Extending SentiSmoke-Twitter dataset, we created SentiSmoke-Reddit product and sentiment datasets by the application of transfer learning. To the best of our knowledge, this was first cross-OSMplatform attempt in topical tobacco research using semi-supervised learning. Further, we prove the efficacy of standard text classification methods on the above datasets by designing

Abstract

Underscoring the importance of storage and cache memory is a steady climb in the present generation, which is highly attributable to the presence of large number of video streaming devices, big data and the profusion of IoT devices. The placement of data across the system in a structured fashion aids in bringing down the number of transmissions within the system. Coded transmissions during the demand phase enables us to use the locally stored cache data most effectively. Coded Caching and Coded Data Rebalancing are two recent techniques to reduce the communication load in scenarios which require communication intended to multiple receivers, each of which store a fraction of the total data available in the system. Building on these two ideas, we focus on two main topics in this thesis. Our first contribution is a coded caching scheme for an interference channel which achieves subexponential subpacketization. Our second contribution consists of coded data rebalancing schemes for node addition and node removal scenarios in a decentralized distributed database. Although coded caching is an effective technique in obtaining large gains in cache-aided broadcast communication networks, its practical implementation is severely hampered by its exponential subpacketization levels. It is crucial in lowering this subpacketization level for various scenarios where coded caching is applied. We specifically choose the scenario of a cache aided interference channel with KT transmitters, KR receivers and N files stored in the system library. Each transmitter and receiver can cache upto MT and MR files from the library respectively. Initial coded caching scheme for this scenario had an exponential subpacketization with respect to the number of receivers KR for large KR and achieved a sum-DoF of L + KRMR N where L , KT MTN . Recent work on this scenario improvised on the required subpacketization to approximately its L th root and thereby boosted the achievable sum-DoF by a multiplicative factor of L. In the first contribution of this thesis, we present a projective geometry based coded caching scheme that achieves a subpacketization of F = 1 m! q m(m−1) 2 mQ−1 i=0 " k − i 1 # q , and a sum-DoF of L(m + 1) where k, m ∈ Z + with m ≤ k and q is some prime power. As the number of receivers grows large, this subpacketization is subexponential with respect to the number of receivers provided the receiver cache fraction MRN is upper bounded by a constant. We further present a comparison of this scheme with the recent work on this scenario that achieved a low subpacketization. Secondly, we focus on handling the problem of non-uniform storage across storage nodes (also known as data skew) in replication-based distributed databases. The performance of such replicationbased distributed databases is affected by data skew and reduction in the replication factor during operation, particularly due to node additions or removals. Data rebalancing refers to the communication involved between the nodes in correcting this data skew, while maintaining the replication factor. For carefully designed distributed databases, transmitting coded symbols during the rebalancing phase has been recently shown to reduce the communication load of rebalancing. In this thesis, we look at balanced distributed databases with random placement, in which each data segment is stored in a random subset of r nodes in the system, where r refers to the replication factor of the distributed database. We call these as decentralized databases. For a natural class of such decentralized databases, we propose rebalancing schemes for correcting data skew and the reduction in the replication factor arising due to a single node addition or removal. We give converse arguments which show that our proposed rebalancing schemes are optimal asymptotically in the size of the file.

Abstract

In this thesis, we propose a distributed, cooperative path planning technique for multiple drones (∼200) to explore an unknown region (∼10,000 connected units) in the presence of obstacles. The map of an unknown region is dynamically created based on the information obtained from sensors and other drones. The unknown region is considered as connected region split up into hexagonal unit regions. We use two types of communication, one long-range, and another more frequently used short-range communication. The short-range communication within drones in smaller proximity helps to avoid obstacles and re-exploration of cells already explored by companion drones located in the same sub-area as proven by the simulation results. The long range communication helps in deciding the next optimal sub-area to be targeted by individual drones based on weighted RNN (Reverse nearest neighbour). We have two types of communication in a weighted hexagonal representation of sub-areas which makes exploration more efficient and scalable. However, given the fact that drones are prone to failures due to a variety of causes, it is important that the strategies developed for these systems be failure resilient. Hence, failure resilience in multi-agent robot systems for area exploration and mapping is a critical aspect. Thus, in the thesis, we will introduce the failures in the simulation scenario and study the impact of failures on the present state of the art algorithms. Moreover, we will also study the present benchmarking metrics and show how the present metrics fail to incorporate the failures that exist in real-world scenarios

Abstract

Society has been evolving rapidly with the digital revolution. With the increase of Internet adoption in the last decade, multilingualism has proliferated as the Internet has provided unrestricted access to resources. In a multilingual society, people often mix multiple languages in an informal setting leading to Code-Mixed (CM) text. The CM text generated has been increasing with time. The analysis of this data may help us to derive potentially valuable conclusions. Since data is seen as the new oil, many public and private institutions have been moving towards digital governance, focusing on data-driven decision-making. This phenomenon is reflected in the enormous increase in the number of new and existing companies established to extract various valuable sources of information. But, analysis of low-resourced CM languages is still a long shot. While there have been significant advancements in Natural Language Processing with Machine Learning, processing low-resourced CM text presents new challenges because of its noisy nature and lack of reliable data. Thus, CM requires better data sets and robust pre-processing pipelines. This work aims to analyze a new approach towards building data sets and proposes a pre-processing pipeline for low-resourced CM text. We have chosen Telugu, a low-resourced Dravidian language spoken in the Southern part of India, as the primary matrix language and English as the secondary embedded language for this study. In this thesis, we first present the largest CM dataset for Code-Mixed Telugu-English Text (CMTET), annotated with a new, highly efficient chatbot-based annotation tool. We then propose a pre-processing pipeline to help model CM data using existing NLP models. The pre-processing pipeline starts with Language IDentification (LID) on raw text. It then presents a novel data normalization technique to normalize the spelling and transliteration errors. To analyze the performance of the proposed pipeline, we carried out Sentiment Analysis on the CM dataset with a reported 2.53% increase in accuracy with the pipeline included.

Abstract

With the growing data in terms of images and videos, it becomes imperative to derive a solution to filter out important/salient information in these data. Learning computational models for visual attention (saliency estimation) is an effort to inch machines/robots closer to human visual cognitive abilities. Developing such a solution can help reduce human effort and can be used in various applications like automatic cropping, segmentation etc. We approached this problem by investigating saliency estimation in images which is predicting the stimuli of the human visual system when exposed to an image. We start by identifying four key components of saliency models, i.e. , input features, multi-level integration, readout architecture, and loss functions. Data-driven efforts have dominated the landscape since the introduction of deep neural network architectures. In deep learning research, the choices in architecture design are often empirical and frequently lead to more complex models than necessary. The complexity, in turn, hinders the application requirements. We review other state-of-the-art models on these four components and propose two novel and simpler end-to-end architectures - SimpleNet and MDNSal. They are straightforward, neater, minimal, and more interpretable than other architectures achieving state-of-the-art performance on public saliency benchmarks. SimpleNet is an optimized encoder-decoder architecture. MDNSal is a parametric model that directly predicts parameters of a GMM distribution and aims to bring more interpretability to the prediction maps. Conclusively, we suggest that the way to move forward is not necessarily to design complex architectures but a modular analysis to optimize each component and possibly explore novel (and simpler) alternatives. After exploring these components, we shifted our focus to saliency estimation in videos where the stimuli of users are captured when exposed to a dynamic scenario. We propose ViNet architecture for the task of video saliency prediction. ViNet is a fully convolutional encoder-decoder architecture. The encoder uses visual features from a network trained for action recognition, and the decoder infers a saliency map via trilinear interpolation and 3D convolutions, combining features from multiple hierarchies. The overall architecture of ViNet is conceptually simple; it is causal and runs in real-time. ViNet does not use audio as input and still outperforms the state-of-the-art audio-visual saliency prediction models on nine different datasets (three visual-only and six audio-visual datasets). We also explore a variation of ViNet architecture by augmenting audio features into the decoder. To our surprise, upon sufficient training, the network becomes agnostic to the input audio and provides the same output irrespective of the input. Interestingly, we also observe similar behavior in the previous state-of-the-art models for audio-visual saliency prediction. Our findings contrast with earlier works on deep learningbased audio-visual saliency prediction, suggesting a clear avenue for future explorations incorporating audio more effectively