Medical AI Courses

These courses are available for the Medical AI Program. For courses for CIDS, CI, HPO, and MedEd Programs, see this page.

Mathematical Foundations of AI

This course can be summarized as “What mathematical background do you need to know in order to understand how to create and evaluate AI models?” The course will provide an overview of probability, statistics, optimization, classical machine learning (classification, regression, clustering, feature selection), model evaluation and calibration.

Medical AI Context

This course can be summarized as “What does one need to know to appreciate how a complex modern healthcare system or organization operates, and how and where medical data (the input to AI systems) are generated.” The course will present an introduction to healthcare from a systems’ perspective, with an emphasis on units/procedures where data is generated. We will pay particular attention to the different data and the characteristics, challenges and limitations of each of these modalities. This will include an introduction to electronic health record systems, to medical imaging, and to ethical and regulatory requirements for medical AI applications.

Computational Foundations of AI

The summary statement for this course is “What does medical data look like (from a software perspective) and how do we analyze it?” We will first present a description of the standards used to store and transfer medical data (e.g., HL7, DICOM, database schemas) and then provide core background to the programming environments/tools/libraries used to analyze such data. We will make heavy use of AI-assisted coding tools to ensure that no significant programming background is required for this course. We will also provide more advanced material for students with a strong programming background.

Medical AI Lab

This will consist of 14 interactive assignments offered via a variety of online platforms that will familiarize students with the use of software tools and libraries to visualize, curate, and analyze medical data including (in a subset of cases) the creation and evaluation of medical AI algorithms. We will leverage open-source and simulated data to simplify privacy and security requirements.

Electives

Analyzing Medical Data I: Text + Clinical Data (Technical Track)

This course will present techniques for the analysis of clinical data that is typically stored in the electronic health record system. This will include a detailed view of the data formats/databases used, classical analysis techniques (such as natural language processing) and the creation of AI models for these types of data.

Analyzing Medical Data II: Imaging, Sensor and Omics Data (Technical Track)

This is a companion course to the Analyzing Medical Data I course and will focus on data sources that are typically stored outside the electronic health record system. This will include medical imaging data, sensor/wearable data, and various genetics and multi-omics datasets (including some that are currently research only but are anticipated to become part of clinical practice in the future.)

Deep Neural Networks, Transformers, and Generative Models (Technical Track)

This course will offer a more advanced take on the cutting-edge technologies that underlie modern medical AI. We will discuss deep neural network architectures, transformers and their use in generative models. Students will learn how to create and train both large-language models (based on existing models), imaging models, multimodal models, and custom language models for non-text languages (e.g., medical event-based models.)

AI Software Engineering in a Regulated Context (Both Tracks)

This course will present a complete view of the process of designing and evaluating medical AI solutions in a regulated context. The course will introduce the students to FDA regulations that govern medical software and medical AI, core international standards that undergird these, and the overall regulatory process. With this background, we will review the AI software engineering lifecycle and related risk management techniques. We will discuss how these regulations impact how software is designed, implemented, tested and monitored (post-release) and the particular challenges that the integration of generative AI techniques present.

Security and Privacy

This course will present a unified view of security and privacy challenges in healthcare in general and medical AI in particular. We will first review core issues in cybersecurity (including FDA regulations) and techniques for preventing and managing cyberattacks. We will also discuss AI specific challenges such as data poisoning, prompt injection attacks, and adversarial attacks. The second half of this course will review privacy both from a regulatory perspective (HIPAA, GDPR) and then review and discuss the privacy challenges unique to each particular type of medical data as it becomes a part of the medical AI lifecycle with a particular emphasis on de-identification techniques.

Clinical Decision Support and Digital Health Systems

In this course we will discuss the creation, implementation, evaluation, monitoring and governance of clinical decision support systems and digital health technologies within healthcare organizations. Students will learn how AI-enabled tools assist clinicians at the point of care and improve workflow and service delivery. The course also explores emerging digital health technologies such as telehealth, remote patient monitoring, wearable devices, digital therapeutics, and generative AI applications. We will also discuss critical appraisal, challenges and emerging opportunities in this area.

Healthcare Systems, Quality and Operational Efficiency

In this course we will present a more advanced view of how complex healthcare systems function across clinical, operational, fiscal, and regulatory domains. Topics will include the application of AI technologies to streamline and optimize patient flow, utilization management, healthcare payment operations, and healthcare quality and value measurement. We will address the ongoing evolution of these technologies within healthcare organizations including governance, interoperability, and return on investment.