Publications

Explore our cutting-edge research contributions to machine learning, deep learning, and high-dimensional statistics

Showing 15 of 15 publications

Information preservation with wasserstein autoencoders: generation consistency and adversarial robustness

AC
AB
SD
3 authors
2025
journal

Amongst the numerous variants Variational Autoencoder (VAE) has inspired, the Wasserstein Autoencoder (WAE) stands out due to its heightened generative quality and intriguing theoretical properties.

Deep generative modelsWasserstein autoencoderMaximum Mean DiscrepancyMinimum distance estimationRobustness
Statistics and Computing
View

Sampling-Tailored Two-Pronged Network for Long-Tailed Class Imbalance Learning

FA
AP
SD
3 authors
2025
journal
Class ImbalanceLong-Tail
Engineering Application of Artifical Intelligence

Force of Attraction-Based Distribution Calibration for Enhancing Minority Class Representation

PM
FA
SD
PS
4 authors
2025
conference

Imbalanced image datasets pose significant challenges for developing robust classifiers, particularly when certain classes are heavily underrepresented. To tackle this issue, we propose Density-Driven Attraction

Class ImbalanceNeural NetworksImage ClassificationDensity-Driven Attraction
IEEE International Joint Conference on Neural Networks (IJCNN)

Can Graph Neural Networks Tackle Heterophily? Yes, With a Label-Guided Graph Rewiring Approach!

KB
SB
SD
3 authors
2025
journal

Graph neural networks (GNNs) witness impressive performances on homophilic graphs characterized by a higher number of edges connecting nodes of similar class labels. A decline

AutoencodersTrainingTopologyGraph neural networksStandardsNetwork topologyMessage passingLearning systemsHistogramsHandsConvolutiongraphheterophilymessage passing (MP)rewiring
IEEE Transactions on Neural Networks and Learning Systems
View

The Goldilocks Principle: Achieving Just Right Boundary Fidelity for Long-Tailed Classification

FA
AP
SD
3 authors
2025
journal

This study addresses the challenges of learning from long-tailed class imbalances in deep neural networks, particularly for image recognition. Long-tailed class imbalances occur when a

Heavily-tailed distributionTrainingImage recognitionFeature extractionAdaptation modelsRobustnessData augmentationClass imbalancelong-tailed classificationfeature regularizationmixup
IEEE Transactions on Emerging Topics in Computational Intelligence
View

Deep multi-view clustering: A comprehensive survey of the contemporary techniques

RA
AG
SD
3 authors
2025
journal

Data can be represented by multiple sets of features, where each semantically coherent set of features is called a view. For example, an image can

Multi-view clusteringDeep learningAutoencodersSubspace clusteringMetric learning
Information Fusion
View

Enhancing Medical Image Analysis with MA-DTNet: A Dual Task Network Guided by Morphological Attention

SG
SD
2 authors
2025
conference

Accurate breast tumor segmentation and malignancy detection are crucial for early cancer diagnosis. In this context, we propose a novel lightweight multi-task learning framework, MA-DTNet,

CV
International Conference on Pattern Recognition (ICPR)
View
Featured Publication

Improved Alzheimer’s Disease Detection with Dynamic Attention Guided Multi-modal Fusion

AB
SR
AG
SD
4 authors
2025
conference

The early detection of neurodegenerative disorders such as Alzheimer’s disease is crucial to providing effective healthcare for management and recovery. We address the task of

ApplicationAttention
International Conference on Pattern Recognition (ICPR)
View
Featured Publication

Enhancing Contrastive Clustering with Negative Pair-guided Regularization

AK
AC
SS
SD
4 authors
2024
journal

Contrastive Learning (CL) aims to create effective embeddings for input data by minimizing the distance between positive pairs, i.e., different augmentations or views of the

Constrative LearningCLusteringRegularization
Transactions on Machine Learning Research
View

Multi-scale morphology-aided deep medical image segmentation

SG
SD
2 authors
2024
journal

Medical image segmentation serves as a critical tool for healthcare professionals, enabling the precise extraction of Regions of Interest (ROIs) from clinical images at the

Medical image segmentationMathematical morphologyMultiscale trainable morphological modulesDeep convolutional networksUNet
Engineering Applications of Artificial Intelligence
View
Featured Publication

Algorithmic Fairness in Lesion Classification by Mitigating Class Imbalance and Skin Tone Bias

FA
TC
SD
3 authors
2024
conference

Deep learning models have shown considerable promise in the classification of skin lesions. However, a notable challenge arises from their inherent bias towards dominant skin

Algorithmic FairnessClass Imbalance
Medical Image Computing and Computer Assisted Intervention (MICCAI)

CCO: A Cluster Core-Based Oversampling Technique for Improved Class-Imbalanced Learning

PM
FA
SD
3 authors
2025
journal

Supervised classification problems from the real world typically face a challenge characterized by the scarcity of samples in one or more target classes compared to

Clustering algorithmsNoise measurementInterpolationNoiseComputational intelligenceClassification algorithmsTask analysisClassificationimbalanced dataoversamplingsynthetic minority oversampling technique
IEEE Transactions on Emerging Topics in Computational Intelligence
View

Mo2E: Mixture of Two Experts for Class-Imbalanced Learning from Medical Images

FA
AB
BS
SD
4 authors
2024
conference

Class imbalance in the medical image dataset is almost inherent due to the limited availability of clinical data for certain diseases and patient populations. Under-represented

TrainingHeadAccuracySociologyTailInference algorithmsTask analysisImbalanced classificationAugmentation TechniqueResampling
2024 IEEE International Symposium on Biomedical Imaging (ISBI)
View

Affinity-based Homophily: Can we measure homophily of a graph without using node labels?

IO
KB
SD
3 authors
2024
conference

The homophily (heterophily) ratio in a graph represents the proportion of edges connecting nodes with similar (dissimilar) class labels. Existing methods for estimating the homophily

GraphsConvolutionHomophilyHeterophilyAffinity
The Second Tiny Papers Track at ICLR 2024
View

Lost in Translation: GANs' Inability to Generate Simple Probability Distributions

DD
AC
SD
3 authors
2024
conference

Since its inception, Generative Adversarial Networks (GAN) have marked a triumph in generative modeling. Its impeccable capacity to mimic observations from unknown probability distributions has

Generative Adversarial NetworksGenerative ModelsStatistical Simulation
The Second Tiny Papers Track at ICLR 2024
View