Abstract

Molecular dynamics (MD) simulations provide a powerful framework for studying the time evolution of molecular systems, enabling the characterization of both thermodynamic and kinetic properties.
However, many biologically and chemically relevant processes, including protein conformational transitions and ligand-binding events, occur on timescales that remain inaccessible to conventional atomistic
simulations. Monte Carlo and enhanced sampling methods can efficiently explore configurational space
and accurately reproduce equilibrium distributions, but they do so at the expense of physical time,
thereby limiting access to kinetic information. This thesis investigates whether physically meaningful molecular trajectories and dynamical information can be recovered without performing extensive
molecular dynamics simulations. Three complementary approaches are developed and evaluated, each
relying on data that is substantially less expensive to generate than fully converged MD trajectories.
The first approach introduces a velocity-based trajectory reconstruction scheme that seeks to recover
the temporal ordering of shuffled molecular dynamics frames. The method combines Maxwell-Boltzmann
velocity likelihoods with a velocity continuity criterion, and a greedy optimization algorithm is employed to identify the most probable sequence of configurations. The approach performs well for simple, low-dimensional systems such as n-pentane, although its effectiveness decreases for more complex
molecular systems.
The second and principal contribution of this thesis is DIME (Dynamics Inferred from Monte Carlo
Ensembles), a framework that reconstructs molecular kinetics from static, Boltzmann-weighted ensembles. Metastable states are identified using density-based clustering, and a rate matrix satisfying detailed
balance is generated by construction, with the transition rate between two states set by the free energy
barrier separating them, read off the free energy surface of the same ensemble. Absolute timescales
are calibrated using a short MD reference trajectory. DIME is validated on six systems of increasing
difficulty: n-pentane and alanine dipeptide as low-dimensional benchmarks, a WLALL pentapeptide,
the achiral peptide AIB9, whose left- and right-handed helices are related by an exact symmetry that
provides a reference-free test of the inferred kinetics, wild-type chignolin as a beta-hairpin folder, and
the millisecond native-state dynamics of bovine pancreatic trypsin inhibitor. Across systems whose
slow processes span picoseconds to a millisecond, the inferred models reproduce MD-derived rate matrices, mean first-passage times, state occupancy autocorrelation functions, and implied timescales with
quantitative accuracy following calibration.
The third component of this work explores the use of deep generative models for trajectory generation in continuous configurational space. Normalizing flows trained on molecular configurations are
v
vi
employed to interpolate within latent space and generate low-barrier transition pathways. In a second approach, continuous normalizing flows are combined with Hamiltonian Neural Networks in latent
space to produce dynamical trajectories. Both methods are demonstrated on the Muller-Brown potential ¨
energy surface.
Overall, the results presented in this thesis demonstrate that meaningful kinetic information can
be extracted from computational approaches that require substantially less effort than fully converged
molecular dynamics simulations. Among the methods developed, DIME provides the most comprehensive and rigorously validated framework for recovering molecular kinetics from equilibrium ensembles.

Abstract

In this thesis, we study extinction in dynamical systems generated by reaction networks. We introduce two notions: weak extinction and strong extinction, and relate them to the structure of the underlying network through Lyapunov functions and LaSalle's invariance principle. In particular, for all
deficiency-zero networks that are not weakly reversible, we provide a geometric construction of linear
Lyapunov functions. Using these functions, we establish that if these networks have bounded invariant
subspaces, then they must exhibit weak extinction within every such subspace. Also, for linear networks
that are not weakly reversible, we show that every species outside a terminal strongly connected component undergoes strong extinction. Moreover, in order to further emphasize the difference between weak
and strong extinction, we construct an example of a reaction system (based on the Ivanova network) that
exhibits weak extinction for all the species, but does not exhibit strong extinction in any species.

Abstract

Modern speech technology often relies on orthographic text such as transcripts, spellings, and characterlevel edit distances, even though human listeners judge speech by phonological and semantic content.
For morphologically rich and low-resource languages, this orthographic bottleneck produces unreliable evaluation and weak pronunciation modelling. This thesis explores phonetic abstraction: the
deliberate reduction of surface script to shared phonological structure before comparison or transfer.
Automatic Speech Recognition (ASR) evaluation for morphologically complex languages remains
an open and practically significant problem. Standard metrics such as Word Error Rate (WER) and
Character Error Rate (CER) operate purely at the lexical level, assigning uniform penalties to all substitutions regardless of whether the resulting output is intelligible to a human listener. In agglutinative and
fusional languages with rich phonological inventories, such as Tamil, Telugu, Hindi, and Marathi, these
metrics routinely misalign with human intelligibility.
This thesis proposes and evaluates a semantic-phonetic hybrid evaluation framework that combines a deterministic phonetic skeleton, derived by mapping consonant characters across languagespecific Indic scripts into shared articulatory equivalence classes, with contextual semantic embeddings
from large pre-trained multilingual models. A linear combination of these two components, optimised
via Ordinary Least Squares regression against expert human scores on 1,287 challenging transcription
pairs, yields a metric that consistently outperforms string-edit and LLM-based baselines across all evaluated languages.
Arabic, Persian, and Urdu share phonology but not computational resources. As a complementary contribution, the thesis also presents a Common Phone Space (CPS-APU) for Arabic, Persian,
and Urdu, which applies cross-lingual phonological transfer to the problem of Grapheme-to-Phoneme
conversion for low-resource Perso-Arabic languages. It is a three-stage Grapheme-to-Phoneme (G2P)
pipeline (expert rules, corpus-mined rules, Large Language Model (LLM)-powered generation), and
shows that cross-lingual transfer from high-resource Arabic improves G2P for low-resource Persian and
Urdu.
Together, these works argue that underlying phonetic and phonological structures offer a robust computational resource that can be systematically exploited to overcome text-level limitations. Both contributions instantiate the same principle: related languages share phonological structure that can be
encoded as equivalence classes (evaluation-time bucketing in Indic scripts; CPS labels in Perso-Arabic
G2P). Phonetic abstraction is a transferable resource for building fair and effective speech technology
for the world's linguistic diversity beyond surface orthography.

Abstract

The classical setting of private information retrieval (PIR) involves the retrieval of one desired message by a client, from a library of K messages stored across N distributed servers, without revealing the
identity of the desired file from any particular server. In graph-based replication, the storage configuration is described by a (simple, undirected) graph with N vertices and K edges, where the edges denote
the files and the nodes denote the servers. Each edge corresponds to a single distinct file, stored exactly
in the two servers indexed by the vertices that are its end points. A query-response protocol between the
client and the servers which enforces (a) the client is able to retrieve his desired file and (b) the identity
of the desired file is not revealed to any individual server constitutes a valid PIR scheme. Such protocols
involve utilizing private randomness at the client, which enables the masking of the index of the desired
file from the server.
The rate of PIR for a given storage scenario is the ratio of the number of symbols in a file, to the
number of symbols downloaded from all servers by the client for successful recovery of the desired
file. The optimal (largest) rate of PIR for a given setting is known as the PIR capacity. Prior work
in high-rate PIR for graph-based systems typically require large subpacketization (which indicates the
number of parts each file has to be divided into, indicating the complexity of the scheme). A good
number of these schemes are fixed-download schemes, i.e., they have the property that the number of
bits downloaded from each server remains unchanged, irrespective of the query realization. In this
thesis, we present four high-rate PIR schemes for graph-based storage systems.
• A unit-subpacketization (which is minimal possible) variable-download scheme on star graphs.
• A fixed-download scheme on star graphs.
• A variable-download scheme for general graphs.
• A fixed-download scheme for general graphs.
The fixed-download scheme for general graphs is obtained from the variable-download scheme, via a
generic method of transformation, known from literature.
Our variable-download PIR scheme on star graphs is the first scheme in the literature that employs
unit subpacketization (L = 1) while also having orderwise optimal rate. Our PIR schemes for general
graphs use a decomposition of the graph via independent sets. They achieve rates lower than prior
schemes for the complete graph, however they can achieve higher rates than previously known for some
specific graph classes.

Abstract

Functional materials with tailored properties are essential for advancements in crucial areas such
as energy storage and semiconductor technologies. Despite progress in generative models, producing
materials for specific target properties remains challenging because of their vast chemical space and
complexities introduced by atomic structures and material properties.
In this work, an approach to the generative design of inorganic materials is introduced, employing a string-based invertible representation called the Simplified Line-Input Crystal Encoding System
(SLICES). This representation adheres to crystallographic invariances, facilitating accurate modeling
and reversible generation of crystal structures. A transformer decoder model with approximately 100
million parameters is integrated with a property-guided generation mechanism based on scaled propertytoken embeddings, following the continuous value encoding methodology. For the first time, special
tokens corresponding to material properties (formation energy, band gap, bulk modulus, and crystal
system) are introduced as control codes, and the embedding of each token is scaled by the desired
property value, allowing fine-grained and continuous control over the generated structures. Unlike
prior approaches such as MatterGPT, which concatenated property embeddings with SLICES tokens,
or SLICES-PLUS, which required architectural modifications and an enhanced crystal representation
for symmetry control, the present work achieves both property and crystal system conditioning within a
single unified framework.
Trained on the Alex-20 dataset (280,033 structures) for formation energy and band gap, and on the
Materials Project dataset (13,024 structures) for bulk modulus, the model consistently achieved validity
exceeding 90% across single-property generation experiments, with competitive uniqueness and novelty.
The framework was extended to dual-property conditioning, demonstrating simultaneous targeting of
formation energy with bulk modulus and formation energy with band gap across broad grids of target
combinations. Crystal system conditioning was further introduced as a new capability, achieving match
rates of up to 93.44% for triclinic systems and 57.06% for cubic systems when co-conditioned with
formation energy, and generalizing consistently across band gap and bulk modulus conditioning as
well. A multi-step validation strategy incorporating structural and compositional plausibility checks,
predictive property assessments via the MEGNet model followed by Density Functional Theory (DFT)
calculations confirmed the physical realism of the generated structures. The results show that MatGPT
generates valid, unique, and novel inorganic crystal structures while enabling controlled generation
toward prescribed formation energy, band gap, bulk modulus, and crystal system targets. Representative
DFT relaxations further indicate that the generated structures can retain their structural integrity after optimization, supporting the physical plausibility of the proposed materials. These results establish
scaled property-token conditioning as a simple and unified strategy for controllable inverse design of
inorganic crystals.
In addition, an agentic workflow for materials science is presented, integrating the Qwen language
model with domain-specific computational tools through a modular Model Context Protocol (MCP)
architecture. The framework augments the language model with tools for structure relaxation, thermodynamic calculations, and retrieval-augmented generation, and is evaluated on the MaScQA benchmark
for materials science question answering. Together, the generative MatGPT framework and the agentic
verification system establish foundations for end-to-end AI-driven materials discovery pipelines.

Abstract

Automated interpretation of high-resolution satellite imagery using deep learning has advanced
rapidly, yet its application to aviation surface mapping remains limited. Existing studies primarily focus
on roads, buildings, or coarse runway-level features, while the fine-grained surface markings required
for airport ground navigation and aerodrome mapping remain underexplored.
This thesis addresses the automated extraction of fine-grained airport surface features from satellite
imagery, with emphasis on supporting digital moving maps and Aerodrome Mapping Database update
workflows within the aeronautical data update (AIRAC) cycle. Unlike conventional object detection
tasks, airport feature extraction requires grouping spatially contiguous and repeating markings--such
as runway thresholds or chevrons--into semantically meaningful aviation objects.
Two case studies first establish feasibility and expose the core challenges. An initial single-class
detection study demonstrates that machine learning can localize airport features from satellite imagery,
but reveals the limitations of small, domain-specific training data and weak generalization. A second
study extends this to multi-class detection and segmentation, achieving strong performance on image
patches yet failing at full-scene scale, where false positives proliferate and conventional sliding-window
inference breaks down. These failures motivate the Airport Ground Navigation Object Detection
Dataset (AGNODD), which addresses the granularity and domain-specific data gap in existing datasets.
To overcome the limitations of sliding-window inference on large satellite scenes, a multi-level-ofdetail tiling strategy is then developed and extended into RAHI (Raster-Aware Hyper Inference),
a memory-mapped, geometry-preserving inference framework that delivers substantial tiling speedups
while improving detection fidelity over conventional slicing. An exploratory generative-AI-based validation stage is also investigated but found to be unreliable and computationally expensive at scale,
motivating a dedicated model-driven solution.
The proposed AERO-DETR pipeline introduces a dual-perspective approach for runway surface
extraction. A top-down stage employs oriented object detection to identify candidate runway corridors,
while a bottom-up stage uses transformer-based set prediction to assemble precise runway polygons
from orientation-standardized markings. Spatial post-processing enforces shape and adjacency constraints aligned with ICAO aviation standards. Experimental results demonstrate improved full-scene
robustness, reduced false positives compared to tiling-only baselines, and a mean Intersection-overUnion of approximately 0.89 against ground-truth data.
Overall, the thesis demonstrates that aviation-aware computer vision pipelines can enable more reliable and scalable mapping of airport surface features from satellite imagery.

Abstract

Change-point detection (CPD) is the problem of detecting abrupt changes in the statistical properties
of a sequence of observations. Given a sequence whose underlying distribution changes at an unknown
time instant--called the change-point--the goal is to estimate when this change occurred. In the offline
(retrospective) setting, the entire sequence is available before processing begins. In the online (sequential) setting, data is processed as it arrives and the goal is to detect the change as quickly as possible
while controlling false alarms. CPD has a rich theoretical foundation, with applications spanning quality
control, financial time series analysis, climate monitoring, and biomedical signal processing.
This thesis studies the CPD problem in the context of channel-coded and modulated communication
systems. In modern wireless communication, adaptive modulation and coding (AMC) is employed to
improve spectral efficiency: the transmitter dynamically adjusts its channel code and modulation scheme
based on the prevailing channel conditions. For the receiver to correctly decode the transmitted data,
it must identify the transmission parameters in use. While blind identification methods can estimate
these parameters directly from the received noise-affected data, they assume that the parameters remain
fixed throughout the observation. In practice, the transmitter may switch its parameters at an unknown
time instant, violating this assumption. It is therefore essential to first detect the change-point--the time
at which the transmitter switches to a different set of parameters--before applying blind identification
algorithms to each segment separately. To the best of our knowledge, CPD has not been previously
investigated in the context of channel coding, and this thesis presents the first systematic study of this
problem.
The primary contribution is a two-stage framework for estimating the change-point when the transmitter switches from one binary linear block code to another. In Stage 1, we apply a data processing
strategy based on inner-product projections with vectors from the dual code, which transforms the received high-dimensional codeword sequence into a one-dimensional sequence of independent Bernoulli
random variables while preserving the change-point signature. We then compute the maximum likelihood estimate of the change-point on this projected sequence and analytically construct a confidence
interval around the estimate such that the true change-point lies within the interval with a desired probability, using the asymptotic distribution of the estimator derived by Hinkley. In Stage 2, we restrict
attention to the unprocessed received data within this confidence interval and employ a trained onedimensional convolutional neural network to refine the change-point estimate. The key insight is that the
coarse statistical estimation narrows the search region, while the neural network provides fine-grained localization within this region. The combined two-stage pipeline significantly outperforms either stage
in isolation: the coarse maximum likelihood estimate alone achieves a detection probability of 0.54 at
low noise levels, while the combined pipeline achieves 0.81 under the same conditions. We also address
the practical scenario where the channel noise parameter is unknown, demonstrating that the framework maintains consistent detection accuracy across varying channel conditions without requiring prior
knowledge of the noise level. The validity and performance of the proposed method are demonstrated
through extensive simulations under both the binary symmetric channel and additive white Gaussian
noise channel, evaluating detection accuracy, localization precision, scalability with sequence length,
and sensitivity to design parameters.
We also investigate two additional CPD-related problems. First, we study the problem of detecting
when the transmitter switches between modulation schemes. We develop two online sliding-window
detection methods: a feature-based method using higher-order cumulants as a discriminating statistic,
and a likelihood-based method adapting the quasi Average Likelihood Ratio Test to a sliding-window
framework. Both methods employ a confirmation mechanism to control the tradeoff between detection
delay and false alarm rate. This investigation is ongoing, and simulation results will be reported in a
forthcoming publication.
Finally, we study a cross-domain application: detecting authorship changes in multi-author documents. For each adjacent paragraph pair, three contrastively fine-tuned transformer encoders--DeBERTav3, RoBERTa-large, and ELECTRA-large--produce style-sensitive pair representations that are classified by encoder-specific multilayer perceptrons. Their boundary-change probabilities are combined
through a validation-tuned weighted ensemble. This completed system achieves macro-F1 scores of
0.995, 0.92, and 0.81 on the ea

Abstract

Although topological mapping and navigation have been extensively studied, the specific role and
downstream impact of loop closures in purely topological representations remains relatively underexplored. Importantly, loop closures in topological maps are fundamentally different from those in
globally referenced metric SLAM systems, where they primarily serve to correct accumulated drift.
Building on recent advances in dense topologies grounded in pixel-level relative 3D geometry, we
propose PixelLoop, a navigation framework that introduces loop closures directly in pixel space. Unlike sparse image-level edges or pose-graph corrections, our formulation treats loop closures as dense
geometric shortcuts that explicitly modify graph connectivity and cost propagation.
By establishing zero-cost correspondences between matched pixels across revisited views, PixelLoop
enables fine-grained topological connectivity that more accurately reflects the underlying spatial structure of the environment. This dense connectivity allows for stable any-point-to-any-point navigation
and produces costmaps that closely approximate true geometric shortest paths. In particular, we highlight the advantages of applying loop closures at the pixel level over traditional image-level topological
representations.
Across extensive simulated experiments in the Habitat simulator, PixelLoop achieves over 35% absolute improvement in both Success Rate (SR) and Success weighted by Path Length (SPL) compared to
strong image-relative baselines, with the largest gains observed in scenarios requiring shortcut exploitation. These improvements are further validated through real-world mobile robot deployments using a
ROS Noetic middleware stack, demonstrating that dense pixel-level loop closures provide a practical,
scalable, and robust foundation for topological visual navigation

Abstract

Legal Information Systems are key tools that improve access to and management of legal information. They help legal professionals, researchers, and the public find case laws, statutes, and regulations
quickly and accurately. These systems save time and reduce errors, making legal decision-making more
effective. They also promote transparency by making legal information available to everyone. They
automate routine tasks, which reduces the workload for legal staff and boosts efficiency. Overall, Legal
Information Systems are vital for making the legal process faster and more accurate. Precedent retrieval
is an important challenge in Legal Information Systems. It involves finding past legal decisions that are
relevant to current cases. Precedents are crucial because they help guide legal reasoning and decisions.
However, finding the right precedent can be difficult because of the large volume of case law. Legal
professionals must use search tools to find cases with similar facts or issues. If precedent retrieval is
inaccurate, it can affect the quality of legal arguments and case outcomes. Improving this process is
vital for making Legal Information Systems more effective.
Developing methods to improve the performance of precedent retrieval systems is an active area of
research. In the literature, efforts are being made to exploit metadata, catchphrases, citations, sentenceand paragraph-level features, etc. Several efforts are also being made to exploit citations in legal documents. In this paper, we propose that the text surrounding the citation, which we refer to as Citation
Anchor Text (CAT), can be used to meaningfully improve the document representation of the judgment
the citation points to (i.e., the referenced judgment). It is reported in the literature that the text surrounding a citation not only provides information to identify the referenced judgment but also supplies
additional information about the referenced judgment and its connection to the argument being formulated. Notably, in the case of the Web, anchor text associated with hyperlinks has been widely exploited
to improve the performance of search engines. It has also been used to index non-textual entities like
images and videos.
In this work, we have analysed the resourcefulness of text surrounding a citation in enhancing the
document representation of the referenced judgment by conducting experiments with various document
representation approaches. Experiments conducted on two Indian SC datasets show that CAT alone has
the potential to capture information that can be leveraged to establish the relevance of the referenced
judgment for a given query case. Moreover, the experiments show that certain proposed document representation approaches capture certain nuances that are not captured by the text present in the referenced judgment, indicating that there is scope to exploit text surrounding citations to improve the performance
of precedent retrieval systems.

Abstract

The rapid adoption of Machine Learning (ML) models in safety-critical and commercial applications has raised fundamental questions about their security. We investigate two complementary facets
of ML model security: Intellectual Property (IP) protection of Graph Neural Networks (GNNs) and
adversarial robustness of merged models.
In the first part, we present GENIE, the first watermarking scheme designed specifically for GNNs
trained on the Link Prediction (LP) task. Existing GNN watermarking techniques address only node or
graph classification, leaving LP--a task central to recommendation systems, social network analysis,
and knowledge graph completion--entirely unprotected. GENIE introduces a novel backdoor-based
watermark that supports both node-representation-based (GCN, GraphSAGE, NeoGNN) and subgraphbased (SEAL) LP methods. It constructs a secret trigger set and watermark vector to embed an ownership signature during training. We further propose Dynamic Watermark Thresholding (DWT), a statistically principled verification procedure that bounds the misclassification probability with high confidence under minimal distributional assumptions. Extensive experiments across 4 model architectures, 7
real-world datasets, and 21 watermark removal attacks demonstrate that GENIE preserves model utility
(less than 2% AUC degradation), achieves near-perfect watermark accuracy, and remains robust against
black-box, white-box, combination, and adaptive attacks.
In the second part, we study the adversarial robustness implications of Model Merging (MM), a
popular technique that combines multiple fine-tuned models into a single multi-task model without requiring access to training data. While recent work suggests that MM confers "free" adversarial robustness by mitigating backdoor attacks, we challenge this notion by demonstrating that MM significantly
increases vulnerability to adversarial transfer attacks. Through comprehensive evaluations spanning 8
MM methods, 7 image classification datasets, 6 attack methods, and 336 distinct attack settings, we establish three statistically validated findings: (1) stronger MM methods lead to higher adversarial transfer
rates (exceeding 80% on average); (2) mitigating representation bias through post-hoc surgery increases
vulnerability to transfer attacks from fine-tuned surrogates; and (3) weight averaging, despite being the
weakest MM method, is paradoxically the most vulnerable to transfer attacks. We provide theoretical
justification via gradient alignment analysis and the Cross-Task Linearity property.
Together, these contributions advance our understanding of ML model security from both defensive
(IP protection) and offensive (adversarial vulnerability) perspectives, highlighting the complex interplay
between model utility, robustness, and security in modern ML systems.