YouTube Courses

Federated Hyperparameter Tuning - Challenges, Baselines, and Connections

Explores challenges and solutions for hyperparameter tuning in federated learning, introducing FedEx method to accelerate the process. Connects to weight-sharing in neural architecture search and demonstrates improved accuracy on benchmarks.

• YouTube
• 57 minutes
• Self-Paced
• Free Video

Optimizing for Interpretability in Deep Neural Networks - Mike Wu

Explore techniques for making deep neural networks more interpretable through regularization, focusing on medical applications in critical care and HIV treatment. Insights on balancing model performance with human-understandable decision processes.

• YouTube
• 51 minutes
• Self-Paced
• Free Video

Deep Learning Methods for Electrocardiograms and Echocardiograms

Explore cutting-edge deep learning methods for ECG and echocardiogram analysis, outperforming traditional algorithms and offering insights into cardiovascular imaging interpretation.

• YouTube
• 1 hour 1 minute
• Self-Paced
• Free Video

GANs in Medical Image Synthesis, Translation, and Augmentation - Jason Jeong

Explore GANs in medical imaging for synthetic data generation, addressing scarcity and imbalance issues. Learn about applications in CT, MRI, and disease detection, with insights on challenges and future directions.

• YouTube
• 56 minutes
• Self-Paced
• Free Video

Adversarial Debiasing With Partial Learning - Medical Image Studies

Explore adversarial debiasing techniques to reduce racial disparities in medical AI models for chest X-rays and mammograms while maintaining overall performance and trustworthiness.

• YouTube
• 33 minutes
• Self-Paced
• Free Video

Assignment Control Plots for Causal Inference Study Design - Rocky Aikens

Explore visual tools for causal inference study design, focusing on assignment-control plots to analyze covariate distribution and improve observational and experimental studies.

• YouTube
• 57 minutes
• Self-Paced
• Free Video

Towards Generalist Imaging Using Multimodal Self-Supervised Learning - Mars Huang

Exploring multimodal self-supervised learning for generalist medical imaging AI, addressing limitations of supervised learning and clinical context integration in automated image analysis.

• YouTube
• 58 minutes
• Self-Paced
• Free Video

MedAI - Training Medical Image Segmentation Models with Less Labeled Data - Sarah Hooper

Innovative semi-supervised method for training medical image segmentation models using minimal labeled data, achieving comparable accuracy to fully supervised approaches while significantly reducing labeling requirements.

• YouTube
• 54 minutes
• Self-Paced
• Free Video

Observational Supervision for Medical Image Classification

Exploring gaze data as a novel supervision source for training deep learning models in medical image classification, focusing on chest X-rays and brain MRI scans for improved efficiency and accuracy.

• YouTube
• 36 minutes
• Self-Paced
• Free Video

Multimodal Medical Research of Vision and Language - Jean-Benoit Delbrouck

Explore multimodal medical research combining vision and language, focusing on innovative tasks like Medical Visual Question Answering and Radiology Report Generation using advanced AI architectures and pre-training techniques.

• YouTube
• 50 minutes
• Self-Paced
• Free Video

Self-Training - Weak Supervision Using Untrained Neural Nets for MR Reconstruction - Beliz Gunel

Exploring untrained neural networks for MR reconstruction, comparing self-training and weak supervision methods to improve performance with limited data while addressing slow inference times.

• YouTube
• 47 minutes
• Self-Paced
• Free Video

Multimodal Opportunistic Risk Assessment for Ischemic Heart Disease - Juan Manuel

Innovative approach integrating body composition biomarkers from CT scans with electronic medical records to enhance ischemic heart disease risk assessment, outperforming current clinical risk scores.

• YouTube
• 1 hour 1 minute
• Self-Paced
• Free Video

Mandoline - Model Evaluation under Distribution Shift

Explores a novel framework for evaluating machine learning models under distribution shifts, using slice-based reweighting to improve performance estimation on target distributions.

• YouTube
• 47 minutes
• Self-Paced
• Free Video

Rethinking Architecture Design for Data Heterogeneity in FL - Liangqiong Qu

Explores innovative architecture designs for federated learning in medical AI, focusing on Transformers to address data heterogeneity challenges and improve model performance across diverse healthcare institutions.

• YouTube
• 38 minutes
• Self-Paced
• Free Video

Frontiers of Medical AI - Therapeutics and Workflows

Explore cutting-edge AI applications in medicine, focusing on therapeutics and workflow optimization. Learn about AI-driven cancer treatment decisions, protein engineering, and specialized medical search engines.

• YouTube
• 46 minutes
• Self-Paced
• Free Video