Dr. Timothy Oladunni

I am a computer science professor focused on building interpretable, robust AI systems for cardiovascular and population health. My work combines multimodal ECG modeling, generative and anomaly-detection methods, and computational epidemiology to detect risk earlier, characterize physiologic mechanisms, and make model decisions transparent to clinicians and public-health leaders.

Explore Research Publications Oladunni Lab

Interpretable AI for Cardiovascular & Population Health

30+

Research Publications

20+

Multimodal AI Systems Built

2M+

ECG and Pysiologic Samples Processed

Research Vision

My research focuses on developing AI systems that can be trusted in high-stakes cardiovascular and population-health settings. I design multimodal models that integrate time, frequency, and time-frequency ECG representations with clinical and population data, with the goal of detecting risk earlier and supporting transparent decision making. Across projects, I study when models should or should not change their predictions, how to preserve physiologic invariance under noise and drift, and how to communicate evidence in forms that clinicians and public-health leaders can interrogate, critique, and refine.

Teaching Philosophy

My teaching philosophy is grounded in the belief that active engagement leads to deeper learning. I design learning experiences that emphasize practical application, analytical reasoning, and hands-on problem solving. By connecting technical concepts to real organizational and industry contexts and helping students to develop both competence and confidence. I continuously work to enhance the learning environment through thoughtful curriculum improvement and evidence-based instructional practices. I also champion computer science for all and strive to ensure that every student has equitable access to high-quality computing education.

Complementary Feature Domain (CFD) Theory

CFD theory formalizes how time-, frequency-, and time frequency-domain ECG representations carry
complementary – not redundant – diagnostic
information. Our work shows that multimodal models perform best when they balance these domains instead of simply stacking more features. We design and evaluate fusion architectures that explicitly preserve this complementarity, improving accuracy, robustness to noise, and data efficiency.

Physiologic Invariance Theory (PIT)

Physiologic Invariance Theory asks a simple question: when should a model change its mind? PIT defines robustness in terms of physiologic drift — normal variability in rhythm, amplitude, and morphology that should not flip a model’s prediction. We develop metrics and training strategies that keep latent representations stable under benign perturbations while remaining sensitive to truly pathological change.

Multimodal ECG Systems

Our lab builds hybrid deep learning systems that integrate time-series signals, frequency spectra, and time–frequency images from the same ECG. Using 1D-CNNs, 2D-CNNs, Transformers, and principled fusion strategies, we classify cardiovascular disease, detect anomalies, and estimate risk in noisy, real-world settings. These systems are designed to be deployable on resource-constrained devices and scalable to large clinical cohorts.

Explainable AI Trustworthiness (EAT)

The EAT framework combines saliency maps, stability analysis, and clinician-aligned explanations to evaluate whether model decisions are medically plausible. We study how explanations change under noise, perturbations, and domain shifts, and we design visualization tools that highlight which parts of the ECG each model truly relies on. The goal is not only high accuracy, but trustworthy, interpretable decisions that clinicians can question and refine.

Pattern Recognition and Machine Learning Laboratory

Lab Overview

The Oladunni Lab focuses on advanced research in artificial intelligence, machine learning, and computational modeling. We design methods that are interpretable, resilient, and grounded in real data, with a strong emphasis on biomedical signal analysis and trustworthy AI. Our work combines algorithm development, multimodal learning, and rigorous evaluation to solve meaningful problems in cardiovascular health and population-level risk detection.

Lab Mission

The mission of the Oladunni Lab is to create reliable and explainable AI systems that improve human health and decision making. We advance scientific understanding through innovative research, high-quality publications, and real-world impact. We also train students to become capable, ethical, and forward-thinking contributors to the computing field.

Student Opportunities

The Oladunni Lab provides hands-on research experiences in machine learning, biomedical AI, and applied data science. Students work on real datasets, contribute to ongoing publications, and build practical skills that prepare them for graduate study and industry careers. The lab welcomes motivated students who are curious, disciplined, and eager to explore computational approaches to scientific challenges.

Core Research Areas

Multimodal Biomedical Signal Processing

We analyze time, frequency, and time-frequency representations of ECG and other physiological signals to uncover patterns linked to clinical outcomes and health anomalies.

Trustworthy and Interpretable AI

We develop methods that make AI systems transparent, reliable, and aligned with human understanding. This includes saliency analysis, drift detection, and evaluation of model stability under noise.

Computational epidemiology & syndromic surveillance

We design machine learning models for cardiovascular health, public health surveillance, and data-driven risk assessment. Our work connects algorithmic innovation with real-world health applications.

Selected Publications

Advancements in Generative AI: A Comprehensive Review of GANs, GPT, Autoencoders, Diffusion Model, and Transformers
IEEE Access, 2024

ML-ECG-COVID: A Machine Learning-Electrocardiogram Signal Processing Technique for COVID-19 Predictive Modeling
IEEE Access, 2023

COVID-19 Vaccine Hesitancy: A Global Public Health and Risk Modelling Framework Using an Environmental Deep Neural Network, Sentiment Classification with Text Mining and Emotional Reactions from COVID-19 Vaccination Tweets
MDPI, 2023

A Machine Learning – Sentiment Analysis on Monkeypox Outbreak: An Extensive Dataset to Show the Polarity of Public Opinion from Twitter Tweets
IEEE Access, 2023

Text Mining and Emotion Classification on Monkeypox Twitter Dataset: A Deep Learning–Natural Language Processing (NLP) Approach
IEEE Access, 2023

A CNN Transfer Learning–Electrocardiogram (ECG) Signal Approach to Predict COVID-19
IEEE International Conference on Computer and Automation Engineering (ICCAE), 2023

Covid-19 Vaccine Hesitancy: Text Mining, Sentiment Analysis and Machine Learning on COVID-19 Vaccination Twitter Dataset
Journal of Expert Systems with Applications, 2022

COVID-19 Fatality Rate Classification Using Synthetic Minority Oversampling Technique (SMOTE) for Imbalanced Class
IEEE International Conference on Pattern Recognition and Machine Learning, 2021

Deep Neural Networks for Human Fall-Risk Prediction Using Force-Plate Time Series Signal
Journal of Expert Systems with Applications, 2021

COVID-19 County Level Severity Classification with Imbalanced Class: A NearMiss Under-sampling Approach
medRxiv, 2021

Automatic Electrocardiogram Detection of Suspected Hypertrophic Cardiomyopathy: Application to Wearable Heart Monitors
IEEE Sensors Letters, 2021

A Time Series Analysis and Forecast of COVID-19 Health Care Disparity
PLOS ONE, 2021 (Submitted)

Exponential Smoothening Forecast of African Americans’ COVID-19 Fatalities
International Conference on Computing and Data Science (CONF-CDS), January 28, 2021

Investigation of Data Size Variability in Wind Speed Prediction of AI Algorithms
Journal of Cybernetics and Systems, 2020

An Optimized ConvNet-LSTM Deep Learning Probabilistic Approach to Source Code Generation with Abstract Syntax Tree and Hyper-Parameter Tuning
Journal of Expert Systems with Applications, 2020 (Submitted)

A Machine Learning Approach to Epileptic Seizure Prediction Using Electroencephalogram (EEG) Signal
Biocybernetics & Biomedical Engineering, 2020

Wind Speed Prediction and Visualization Using Long Short-Term Memory Networks (LSTM)
10th IEEE International Conference on Information Science and Technology (ICIST), 2020

Data Augmentation for Mixed Spectral Signatures Coupled with Convolutional Neural Networks
9th IEEE International Conference on Information Science and Technology (ICIST, 2019)

A Deep Learning Model for Source Code Generation
IEEE SoutheastCon, 2019

A Gaussian Data Augmentation Technique on Highly Dimensional, Limited Labeled Data for Multiclass Classification Using Deep Learning
10th IEEE International Conference on Intelligent Control and Information Processing (ICICIP, 2019)

Homomorphic Encryption and Data Security in the Cloud
28th International Conference on Software Engineering and Data Engineering, 2019

A Data Augmentation-Assisted Deep Learning Model for High Dimensional and Highly Imbalanced Hyperspectral Imaging Data
9th IEEE International Conference on Information Science and Technology (ICIST, 2019)

H2O Deep Learning for Hedonic Pricing
International Journal of Computers and Their Applications, March 2018

Foreclosure Sale and House Value: Correlation or Causation?
16th IEEE International Conference on Machine Learning and Applications (ICMLA), December 18–21, 2017

A Spatio-Temporal Hedonic House Regression Model
16th IEEE International Conference on Machine Learning and Applications (ICMLA), December 18–21, 2017

An Occam’s Razor Approach to Hedonic Pricing Theory
4th IEEE International Conference on Computational Science and Computational Intelligence, December 14–16, 2017

Spatial Dependency and Hedonic Housing Regression Model
15th IEEE International Conference on Machine Learning and Applications (ICMLA, 2016), December 18–20, 2016

Hedonic Housing Theory – A Machine Learning Investigation
15th IEEE International Conference on Machine Learning and Applications (ICMLA, 2016), December 18–20, 2016

Predictive Real Estate Multiple Listing System Using MVC Architecture and Linear Regression
ISCA 24th International Conference on Software Engineering and Data Engineering, October 12–14, 2015

Predicting Fair Housing Market Value: A Machine Learning Investigation
International Journal of Computers and Their Applications, September 2016

Hedonic House Pricing Model Using Deep Learning with L1 Regularization
ISCA 26th International Conference on Software Engineering and Data Engineering, October 2–4, 2017

‍

Contact & Collaboration

For collaboration inquiries, student opportunities, or academic correspondence, please use the form below or contact the lab directly.

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.