arXiv subjects

Anil A Bharath

Publications and source records attributed to Anil A Bharath.

10 recordsLinked to original sources

arXiv2026

Privacy-Preserving Generation of Clinical Narratives from Medical Terminologies

In high-stakes domains such as healthcare, privacy concerns severely limit the use of real-world training data. Differentially private (DP) synthetic data offers a promising alternative with formal privacy guarantees, but achieving strong utility remains challenging for clinical note generation due to domain specificity and long-form text complexity. We present Term2Note, a method for synthesising full-length clinical notes under DP constraints. By structurally separating content and form, Term2Note generates section-wise note content conditioned on medical terms, with terms and notes privatised under separate DP constraints, and applies a DP quality maximiser to improve outputs. Experiments demonstrate that Term2Note produces synthetic notes with statistical properties closely aligned with real clinical notes, and that downstream models trained on these notes achieve performance comparable to those trained on real clinical data. Compared to existing DP text generation baselines, Term2Note substantially improves both fidelity and utility, without relying on label distribution assumptions, highlighting its effectiveness as a practical privacy-preserving alternative to real clinical notes.

cs.CL↗

arXiv2025

Evaluating Differentially Private Generation of Domain-Specific Text

Generative AI offers transformative potential for high-stakes domains such as healthcare and finance, yet privacy and regulatory barriers hinder the use of real-world data. To address this, differentially private synthetic data generation has emerged as a promising alternative. In this work, we introduce a unified benchmark to systematically evaluate the utility and fidelity of text datasets generated under formal Differential Privacy (DP) guarantees. Our benchmark addresses key challenges in domain-specific benchmarking, including choice of representative data and realistic privacy budgets, accounting for pre-training and a variety of evaluation metrics. We assess state-of-the-art privacy-preserving generation methods across five domain-specific datasets, revealing significant utility and fidelity degradation compared to real data, especially under strict privacy constraints. These findings underscore the limitations of current approaches, outline the need for advanced privacy-preserving data sharing methods and set a precedent regarding their evaluation in realistic scenarios.

cs.LG↗

arXiv2025

Generating Synthetic Data with Formal Privacy Guarantees: State of the Art and the Road Ahead

Privacy-preserving synthetic data offers a promising solution to harness segregated data in high-stakes domains where information is compartmentalized for regulatory, privacy, or institutional reasons. This survey provides a comprehensive framework for understanding the landscape of privacy-preserving synthetic data, presenting the theoretical foundations of generative models and differential privacy followed by a review of state-of-the-art methods across tabular data, images, and text. Our synthesis of evaluation approaches highlights the fundamental trade-off between utility for down-stream tasks and privacy guarantees, while identifying critical research gaps: the lack of realistic benchmarks representing specialized domains and insufficient empirical evaluations required to contextualise formal guarantees. Through empirical analysis of four leading methods on five real-world datasets from specialized domains, we demonstrate significant performance degradation under realistic privacy constraints ($ε\leq 4$), revealing a substantial gap between results reported on general domain benchmarks and performance on domain-specific data. %Our findings highlight key challenges including unaccounted privacy leakage, insufficient empirical verification of formal guarantees, and a critical deficit of realistic benchmarks. These challenges underscore the need for robust evaluation frameworks, standardized benchmarks for specialized domains, and improved techniques to address the unique requirements of privacy-sensitive fields such that this technology can deliver on its considerable potential.

cs.CR↗

arXiv2023

Prototype of a Cardiac MRI Simulator for the Training of Supervised Neural Networks

Supervised deep learning methods typically rely on large datasets for training. Ethical and practical considerations usually make it difficult to access large amounts of healthcare data, such as medical images, with known task-specific ground truth. This hampers the development of adequate, unbiased and robust deep learning methods for clinical tasks. Magnetic Resonance Images (MRI) are the result of several complex physical and engineering processes and the generation of synthetic MR images provides a formidable challenge. Here, we present the first results of ongoing work to create a generator for large synthetic cardiac MR image datasets. As an application for the simulator, we show how the synthetic images can be used to help train a supervised neural network that estimates the volume of the left ventricular myocardium directly from cardiac MR images. Despite its current limitations, our generator may in the future help address the current shortage of labelled cardiac MRI needed for the development of supervised deep learning tools. It is likely to also find applications in the development of image reconstruction methods and tools to improve robustness, verification and interpretability of deep networks in this setting.

physics.med-ph↗

arXiv2021

EP-PINNs: Cardiac Electrophysiology Characterisation using Physics-Informed Neural Networks

Accurately inferring underlying electrophysiological (EP) tissue properties from action potential recordings is expected to be clinically useful in the diagnosis and treatment of arrhythmias such as atrial fibrillation, but it is notoriously difficult to perform. We present EP-PINNs (Physics-Informed Neural Networks), a novel tool for accurate action potential simulation and EP parameter estimation, from sparse amounts of EP data. We demonstrate, using 1D and 2D in silico data, how EP-PINNs are able to reconstruct the spatio-temporal evolution of action potentials, whilst predicting parameters related to action potential duration (APD), excitability and diffusion coefficients. EP-PINNs are additionally able to identify heterogeneities in EP properties, making them potentially useful for the detection of fibrosis and other localised pathology linked to arrhythmias. Finally, we show EP-PINNs effectiveness on biological in vitro preparations, by characterising the effect of anti-arrhythmic drugs on APD using optical mapping data. EP-PINNs are a promising clinical tool for the characterisation and potential treatment guidance of arrhythmias.

physics.med-ph↗

arXiv2020

Left atrial ejection fraction estimation using SEGANet for fully automated segmentation of CINE MRI

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, characterised by a rapid and irregular electrical activation of the atria. Treatments for AF are often ineffective and few atrial biomarkers exist to automatically characterise atrial function and aid in treatment selection for AF. Clinical metrics of left atrial (LA) function, such as ejection fraction (EF) and active atrial contraction ejection fraction (aEF), are promising, but have until now typically relied on volume estimations extrapolated from single-slice images. In this work, we study volumetric functional biomarkers of the LA using a fully automatic SEGmentation of the left Atrium based on a convolutional neural Network (SEGANet). SEGANet was trained using a dedicated data augmentation scheme to segment the LA, across all cardiac phases, in short axis dynamic (CINE) Magnetic Resonance Images (MRI) acquired with full cardiac coverage. Using the automatic segmentations, we plotted volumetric time curves for the LA and estimated LA EF and aEF automatically. The proposed method yields high quality segmentations that compare well with manual segmentations (Dice scores [$0.93 \pm 0.04$], median contour [$0.75 \pm 0.31$] mm and Hausdorff distances [$4.59 \pm 2.06$] mm). LA EF and aEF are also in agreement with literature values and are significantly higher in AF patients than in healthy volunteers. Our work opens up the possibility of automatically estimating LA volumes and functional biomarkers from multi-slice CINE MRI, bypassing the limitations of current single-slice methods and improving the characterisation of atrial function in AF patients.

eess.IV↗

arXiv2018

Adversarial Information Factorization

We propose a novel generative model architecture designed to learn representations for images that factor out a single attribute from the rest of the representation. A single object may have many attributes which when altered do not change the identity of the object itself. Consider the human face; the identity of a particular person is independent of whether or not they happen to be wearing glasses. The attribute of wearing glasses can be changed without changing the identity of the person. However, the ability to manipulate and alter image attributes without altering the object identity is not a trivial task. Here, we are interested in learning a representation of the image that separates the identity of an object (such as a human face) from an attribute (such as 'wearing glasses'). We demonstrate the success of our factorization approach by using the learned representation to synthesize the same face with and without a chosen attribute. We refer to this specific synthesis process as image attribute manipulation. We further demonstrate that our model achieves competitive scores, with state of the art, on a facial attribute classification task.

cs.CV↗

arXiv2018

Inverting The Generator Of A Generative Adversarial Network (II)

Generative adversarial networks (GANs) learn a deep generative model that is able to synthesise novel, high-dimensional data samples. New data samples are synthesised by passing latent samples, drawn from a chosen prior distribution, through the generative model. Once trained, the latent space exhibits interesting properties, that may be useful for down stream tasks such as classification or retrieval. Unfortunately, GANs do not offer an "inverse model", a mapping from data space back to latent space, making it difficult to infer a latent representation for a given data sample. In this paper, we introduce a technique, inversion, to project data samples, specifically images, to the latent space using a pre-trained GAN. Using our proposed inversion technique, we are able to identify which attributes of a dataset a trained GAN is able to model and quantify GAN performance, based on a reconstruction loss. We demonstrate how our proposed inversion technique may be used to quantitatively compare performance of various GAN models trained on three image datasets. We provide code for all of our experiments, https://github.com/ToniCreswell/InvertingGAN.

cs.CV↗

arXiv2018

Denoising Adversarial Autoencoders: Classifying Skin Lesions Using Limited Labelled Training Data

We propose a novel deep learning model for classifying medical images in the setting where there is a large amount of unlabelled medical data available, but labelled data is in limited supply. We consider the specific case of classifying skin lesions as either malignant or benign. In this setting, the proposed approach -- the semi-supervised, denoising adversarial autoencoder -- is able to utilise vast amounts of unlabelled data to learn a representation for skin lesions, and small amounts of labelled data to assign class labels based on the learned representation. We analyse the contributions of both the adversarial and denoising components of the model and find that the combination yields superior classification performance in the setting of limited labelled training data.

cs.CV↗

arXiv2017

Generative Adversarial Networks: An Overview

Generative adversarial networks (GANs) provide a way to learn deep representations without extensively annotated training data. They achieve this through deriving backpropagation signals through a competitive process involving a pair of networks. The representations that can be learned by GANs may be used in a variety of applications, including image synthesis, semantic image editing, style transfer, image super-resolution and classification. The aim of this review paper is to provide an overview of GANs for the signal processing community, drawing on familiar analogies and concepts where possible. In addition to identifying different methods for training and constructing GANs, we also point to remaining challenges in their theory and application.

cs.CV↗