Deeksha M Shama

Deeksha M Shama

PhD student at Johns Hopkins University

Research

Project 1 Icon

Uncertainty-Aware Bayesian Deep Learning with Noisy Training Labels for Epileptic Seizure Detection

Novel strategy to handle label noise with the target application of tackling inter-rater variability in seizure detection. Accepted at UNSURE@MICCAI'24. Long article coming soon.

GitHub | Paper
Project 2 Icon

DeepSOZ: A Robust Deep Model for Joint Temporal and Spatial Seizure Onset Localization from Multichannel EEG Data from Multichannel EEG Data

Transformer-based network for multi-task framework with uncertainty quantification. Presented at MICCAI'23.

GitHub | Paper
Project 3 Icon

DeepBreath—automated detection of respiratory pathology from lung auscultation in 572 pediatric outpatients across 5 countries

Involved in the initial model developement and the first draft of paper. Extended it for COVID-19 detection presented as my bachelor's thesis.

GitHub | Paper | Thesis

Mentored projects

Project 3 Icon

DeepSOZ-HEM at Seizure Detection Challenge

Presented by Amruth, an undergraduate researcher in our lab, at the International Conference on Artificial Intelligence in Epilepsy and Other Neurological Disorders (2025).

Leaderboard
Project 3 Icon

Interpretable and Lightweight Machine Learning Approach for Autism Classification Using Biomarkers Derived from Multi-trial Resting EEG

Presented by Michelle, an intern in our lab, at BU's RISE symposium. Short article coming soon.

Github | Paper

Other projects

Synthetic Telepathy: Inner Speech Recognition using EEG

Employed machine learning, clustering, and deep leanring methods to decode imagined speech from EEG. Group project for Machine learning and Signal processing course JHU'21

Link

Neural style conversion with Generative models

Developed cycle-GAN and diffusion models for synthetic art generation. Group project for deep learning course JHU'22

Link

Multi-Atlas Brain Segmentation And Age Prediction

Developed 3D CNN model for brain age prediction after preprocessing 3D MR images for Medical Image Analysis course JHU'22

Link