HMX Short Courses
AI in Medicine

Recent advances in AI are enabling automated systems to detect patterns in biomedical signals with human-level or even superhuman performance, expanding access to timely and accurate interpretation of signal data for clinical use. This course will introduce you to the key ideas and approaches behind AI interpretation of biomedical signals, highlighting techniques that can directly analyze data to uncover intricate patterns. Topics include:

• Introduction to Biomedical Signals and Deep Learning
• EEG Fundamentals
• Clinical Use of the EEG
• Detecting Arrhythmias Using Deep Learning
• Evaluating Deep Learning Models for Arrhythmia Detection
• Deep Learning for Arrhythmia Detection
• Assessing the Potential of Deep Learning for Detection of Arrhythmias
• Predicting Psychiatric Disorder Symptoms with EEG and Machine Learning
• Evaluating the Predictive Capacity of the EEG Model
• In Practice: Evolving Applications of AI in Biomedical Signal Interpretation

AI has transformed the use of deep learning for interpreting medical images, such that computers can now effectively “learn from pixels”, along with other sources of clinical data, to achieve levels of diagnostic accuracy that rival (and sometimes exceed) those of human experts. Understanding these recent advances and the key developments that have enabled them will provide you with a strong basis for understanding future developments in this field. Topics include:

• Exploring Diverse Medical Imaging Modalities
• Retinal and Skin Photography
• Automating Lung Disease Detection from Chest X-Rays
• Results of Deep Learning for Chest X-Rays
• Predicting Cardiovascular Risk from Retinal Images
• Navigating Challenges in Retinal Image Analysis with Deep Learning
• Predicting Major Adverse Cardiovascular Events
• Enhancing Dermatologic Diagnosis with Deep Learning
• Unpacking the Potential and Challenges of Deep Learning in Dermatologic Analysis
• Results of Deep Learning in Dermatologic Diagnosis
• In Practice: Evolving Applications of AI in Medical Imaging

Advances in AI, including the development of large language models, have transformed natural language processing (NLP) such that computers can effectively work with text for a variety of tasks that include summarization, translation, extraction of key terms, and question-answering. Understanding these recent advances and the major developments that enabled them will provide anyone working in health care and related sectors with a strong basis for understanding future developments in this field. Topics include:

• Overview of the Importance of NLP and AI in Health Care
• The Value of Clinical Notes and NLP
• Example Note & Task Walkthrough
• Strengths and Weaknesses of Using NLP on Clinical Notes
• The Building Blocks for State-of-the-Art NLP
• From Sentences to Model Input
• Deep Learning for NLP
• The Large Language Model (LLM) “Revolution”
• Improving LLMs for Tasks of Interest
• Continuous Improvement of LLMs
• In Practice: Evolving Applications of AI and NLP in Health Care Systems

Pranav Rajpurkar, PhD

Assistant Professor of Biomedical Informatics, Harvard Medical School

Brett Beaulieu-Jones, PhD

Assistant Professor of Biomedical Data Science, University of Chicago