| layout | title | nav_order | description |
|---|---|---|---|
page |
Syllabus |
2 |
Artificial Intelligence in Medicine II |
{:.no_toc}
{: .no_toc .text-delta }
- TOC {:toc}
You’ve likely encountered artificial intelligence (AI) before, either through your previous work in computer science, machine learning, or applications of AI in healthcare. You may have developed models, processed data, or explored how AI can transform research and medical practice. In this course, we’ll build on that foundation, but we’ll also challenge you to think differently about the role AI plays in the medical field.
Together, we’ll explore not just how AI models are developed, but why they matter in a real-world healthcare context. For instance, how can self-supervised learning or generative models enhance the way we analyze medical images or understand patient records? How do these techniques integrate into clinical workflows, and what are the implications of their use on patient outcomes and ethical decision-making in healthcare?
We’ll also dive deeply into the process of understanding data—how it’s collected, analyzed, and used in practice. You’ll work hands-on with multimodal data, ranging from natural language to medical imaging, and will develop the skills needed to interpret these data sources effectively, keeping in mind the ethical complexities of AI in healthcare.
Throughout the course, you’ll have opportunities to apply what you learn. You will collaborate on research projects, dive into the current literature, and work on practical tutorials that simulate real-world AI applications in medicine. We’ll be thinking about questions like: What makes an AI model interpretable for clinicians? How do we ensure that AI integrates smoothly into medical practice without disrupting care? And how can we be responsible stewards of AI technology in an ethically complex field like medicine?
By the end of this course, you’ll have sharpened your technical skills and deepened your understanding of the challenges and possibilities AI presents in medicine. You’ll not only know how to build and evaluate AI models but also how to think critically about their application in healthcare settings. The skills you gain will be invaluable in contributing to the future of AI-driven healthcare solutions.
{:.no_toc}
- Develop a comprehensive understanding of key advancements in AI as applied to medical informatics, including natural language processing, medical image analysis, and the use of relational and longitudinal data.
- Apply cutting-edge AI methodologies to real-world medical challenges, gaining hands-on experience in the development, implementation, and evaluation of AI models.
- Synthesize diverse AI techniques, learning how to integrate multiple data types—such as text, images, and clinical records—into cohesive solutions for healthcare applications.
- Reflect on the role of AI in clinical practice, understanding the impact of AI technologies on both healthcare professionals and patients, while exploring your own potential contributions to the field.
- Cultivate an appreciation for the ethical, legal, and social dimensions of AI in healthcare, focusing on building trustworthy AI models that prioritize fairness, interpretability, and patient outcomes.
- Develop the ability to critically evaluate and adapt AI methods over time, gaining skills in model monitoring, addressing real-world data challenges, and continuing your learning journey in AI and health informatics beyond the course.
Your final grade in this course will be based on a combination of semester-long course project, weekly assessments, and participation in focused tutorials.
-
All submissions related to course project (project proposal, midterm presentation materials, final report, final presentation materials) must be made through Canvas.
-
Weekly reading assessments are pre-class quizzes that open at 9:00am EST on Wednesday and are due at 2:00pm EST on Tuesday (pre-class quizzes close before lectures on Tuesday). Quizzes must be completed in Canvas.
-
Delayed beyond 24 hours of deadline: no credit. In the case of illness/absence, contact the course instructor. We will work with you to make up any missed assignments.
-
Questions/issues: Please contact the course instructor.
| Component | Percentage | Description |
|---|---|---|
| Project Proposal | 5% | A 2-page proposal outlining your project’s research question, methodology, dataset, and contingency plans, evaluated for clarity and feasibility. A third page is allowed for figures and tables. Unlimited space for references. |
| Peer-Reviewed Feedback on Proposal | 5% | Constructive feedback is provided to peers, following the criteria for effective research review. |
| Midterm Project Presentation | 10% | A presentation summarizing your progress, baseline results, and challenges. Assessed for clarity, engagement, and preparedness for feedback. Presentation file submitted through Canvas. |
| Final Project Report | 50% | A comprehensive, NeurIPS-style report detailing your research question, methods, results, and conclusions. Assessed for depth, accuracy, and insights. |
| Final Project Presentation | 13% | A conference-style presentation summarizing your project’s outcomes, strengths, and limitations. Evaluated on clarity, organization, and professionalism. Presentation file submitted through Canvas. |
| Focused Tutorials and Lectures | 5% | Participation in hands-on tutorials and lectures, demonstrating engagement with coding and model application exercises. |
| Weekly Reading Assessments | 12% | Completion of weekly assessments following assigned readings, ensuring ongoing engagement and comprehension. 1 point per quiz; there is no quiz for Lecture 1. |
Each week, you’ll be assigned 2-4 articles that explore important topics in medical AI. These readings are essential for building a strong understanding of the concepts we’ll discuss in class. After completing the readings, you’ll answer 3-5 short answer questions to reflect on what you’ve learned. These questions are graded on completion, so if you submit thoughtful responses, you’ll receive full credit. You’ll also receive model answers to compare with your own, helping you check your understanding and reinforce the material.
We encourage you to engage meaningfully with these questions, as your responses will guide us in identifying any misunderstood or overlooked concepts. This feedback will be used to tailor our lectures and ensure we’re covering areas that need more clarity.
These assessments are designed to help you consolidate your learning and ensure you’re mastering the critical topics each week.
- Students must attend all classes unless they have explicit permission from the course instructor. An unexcused absence can affect the participation grade. The course will be run in a in-person format.
- Students are highly encouraged to attend focused tutorials. We will offer a number of tutorials throughout the semester and expect that the student will attend at least some of them in-person.
- To see when lectures and office hours are scheduled, check the Weekly Schedule.
- To see when lectures, discussions, and assignments are released (and due), check the Home Page.
{:.no_toc}
Auditing the course is only permitted with explicit permission by the course faculty. Auditors must not increase the workload for instructors and TAs, or compete with enrolled students for other resources provided for students who are paying tuition, such as space in online classrooms or time during office hours.
If you are planning to audit the course but want to get more involved, i.e. submit homework assignments, use office hours, etc., you are encouraged to register as a “special student”, which provides access to this course at a per credit cost.
{:.no_toc}
- The office hours are listed on the Weekly Schedule, and will be held virtually and in-person.
- Students can come to office hours for any questions on course assignments or material.
- In-person office hours will be held in various locations specified in the Weekly Schedule.
- The instructor will also be hosting office hours. These will be reflected on the Weekly Schedule.
{:.no_toc}
Students will be allowed to submit regrade requests for the autograded and written portions of assignments in cases in which the rubric was incorrectly applied or the autograder scored their submission incorrectly.
Regrade requests will not be considered in cases in which:
- a student submits incorrect files and the student does not notify the course staff before the assignment deadline
- a student fails to save their notebook before exporting and uploads an old version to the online system
- a situation arises in which the course staff cannot ensure that the student's work was done before the assignment deadline
Students taking this course come from a wide range of backgrounds. We hope to foster an inclusive and safe learning environment based on curiosity rather than competition. All members of the course community—the instructor, TAs and students—are expected to treat each other with courtesy and respect. Some of the responsibility for that lies with the staff, but a lot of it ultimately rests with you, the students.
{:.no_toc}
Sometimes, the little things add up to creating an unwelcoming culture to some students. Your classmates may have medical conditions (physical or mental), personal situations (financial, family, etc.), or interests that aren’t common to most students in the course. Another aspect of professionalism is avoiding comments that (likely unintentionally) put down colleagues for situations they cannot control. Bragging in open space that an assignment is easy or “crazy,” for example, can send subtle cues that discourage classmates who are dealing with issues that you can’t see. Please take care, so we can create a class in which all students feel supported and respected.
Beyond the slips that many of us make unintentionally are a host of behaviors that the course staff, department, and university do not tolerate. These are generally classified under the term harassment; sexual harassment is a specific form that is governed by federal laws known as Title IX.
{:.no_toc}
Professionalism and respect for diversity are not just matters between students; they also apply to how the course staff treat the students. The staff of this course will treat you in a way that respects our differences. However, despite our best efforts, we might slip up, hopefully inadvertently. If you are concerned about classroom environment issues created by the staff or overall class dynamic, please feel free to talk to us about it. The instructor particular welcomes any comments or concerns regarding conduct of the course and the staff.
We are committed to creating a learning environment welcoming of all students that supports a diversity of thoughts, perspectives and experiences and respects your identities and backgrounds. If you feel like your performance in the class is being affected by your experiences outside of class (e.g., family matters, current events), please don’t hesitate to come and talk with us. We want to be resources for you.
{:.no_toc}
You are more than welcome to visit our office hours and talk with us. We know graduate school can be stressful and we want to help you succeed.
{:.no_toc}
Extensions are only provided in the case of exceptional circumstances. For that, email the course instructor to request an extension. If you make a request close to the deadline, we can not guarantee that you will receive a response before the deadline. Additionally, simply submitting a request does not guarantee you will receive an extension. Even if your work is incomplete, please submit before the deadline so you can receive credit for the work you did complete.
{:.no_toc}
AI research and applications can involve collaboration and group discussions. While you may talk with others about the homework, we ask that you write your solutions individually in your own words. If we suspect that you have submitted plagiarized work, we will call you in for a meeting. If we then determine that plagiarism has occurred, we reserve the right to give you a negative full score (-100%) or lower on the assignments in question, along with reporting your offense to the Center of Student Conduct.
Rather than copying someone else's work, ask for help. You are not alone in this course! The entire staff is here to help you succeed. If you invest the time to learn the material and complete the assignments, you won't need to copy any answers.
{:.no_toc}
The following policy outlines the guidelines for the use of generative AI, LLMs, and coding copilots in student assignments.
-
Responsibility for content: Students who use LLMs and generative AI tools in their assignments must take full responsibility for the content they submit. This includes ensuring the accuracy, relevance, and originality of the information provided by these tools.
-
Acknowledgment of AI use: Students must clearly acknowledge any use of LLMs or generative AI in their assignments. This acknowledgment should specify the nature and extent of the assistance received from these tools. LLMs and generative AI can be used to enhance the educational experience and help with ideation and understanding of complex concepts. However, students must perform the critical thinking, analysis, and synthesis of information.
-
Ethical use and originality: Students must use these tools ethically, following the principles of academic honesty. The use of AI to plagiarize, misrepresent original work, or fabricate data is strictly prohibited. Students are encouraged to use these tools to inspire and inform their work, not to undermine the learning process.
-
Coding copilots: Students may use coding copilots to assist with coding tasks. However, students bear ultimate responsibility for the code they submit, including its correctness, efficiency, and originality. Instructors have the discretion to assess students’ conceptual understanding of the code as part of grading.
-
Instructor discretion: Instructors may specify assignments where LLMs, generative AI, or coding copilots are particularly encouraged or prohibited, depending on the assignment's learning objectives.
This policy helps students get ready for a future with AI in jobs and ensures their education focuses on honesty and learning. Students are encouraged to read this NEJM AI editorial on why we support the use of LLMs and generative AI.
{:.no_toc}
All work in this course is governed by Harvard Medical School’s academic integrity policies. It is the students’ responsibility to be aware of these policies and to ensure that their work adheres to them both in detail and in spirit. Unless otherwise specified by the instructor, the assumption is that all work submitted must reflect the student’s own effort and understanding. Students are expected to clearly distinguish their own ideas and knowledge from information derived from other sources, including from conversations with other people. When working with others you must do so in the spirit of collaboration, not via a unidirectional transfer of information. Note that sharing or sending completed assignments to others will nearly always violate this collaborative standard. If you have a question about how best to complete an assignment in light of these policies, ask the instructor for clarification.
Students are expected to clearly distinguish their own ideas and knowledge from information derived from other sources, including from collaboration with other people. Specifically, this means that:
- Students must properly cite all submitted work appropriately.
- Unless noted otherwise, students are expected to complete assignments, quizzes, and projects individually, not as teams. Discussion about course content and materials is acceptable, but sharing solutions is not acceptable.
- Even though students are encouraged to consult websites for solutions to coding problems, they may never just copy code.
If you have a question about how best to complete an assignment in light of these policies, ask the instructor for clarification.
{:.no_toc}
Harvard Medical School is committed to supporting inclusive learning environments that value and affirm the diverse ideas and unique life experiences of all people. An equitable, inclusive classroom is a shared responsibility of both instructors and students, and both are encouraged to consider how their own experiences and biases may influence the learning environment. This requires an open mind and respect for differences of all kinds.
Students are encouraged to contact the course director if they are experiencing bias or feel that their learning experience – including a course’s content, manner of instruction, or learning environment – is not inclusive. Program administrators and directors, the Office for Gender Equity, and the Ombuds Office are also available to discuss your experiences and provide support. Additionally, students can utilize Harvard’s Anonymous Reporting Hotline to report issues related to bias.
{:.no_toc}
We value your well-being and recognize that as a graduate student you are asked to balance a variety of responsibilities and potential stressors: in class, in lab, and in life. If you are struggling with experiences either in- or outside of class, there are resources available to help. In addition to program leadership, master’s students can contact [email protected], HMS Director of Administration and Student Affairs for Master’s Programs and [email protected], Senior Associate Dean for Graduate Education, for support.
{:.no_toc}
Counseling and Mental Health Services (CAMHS) is a counseling and mental health support service that seeks to work collaboratively with students and the University to support individuals experiencing some measure of distress in their lives. It provides coverage to students year-round and is included in the student health fee, regardless of insurance, at no additional cost. More information is available on the CAMHS website or by calling the main office at 617-495-2042. Urgent care can be reached 24/7 at 617-495-5711.
CAMHS Care Line: The CAMHS Cares line 617-495-2042 is a 24/7 support line available to Harvard students who have mental health concerns, whether you are in immediate distress or not, on-campus or elsewhere. This the Line can also be used as resource for Harvard personnel who needs advice about a student who may be experiencing a mental health crisis. At all times, including evenings, weekends, and holidays, you can follow the prompts to speak directly with a CAMHS Cares Counselor about an urgent concern or if you just need to talk to someone about a difficult challenge.
TimelyCare, a virtual mental health and wellbeing platform for all Harvard students covered by the Student Health Fee, offers free virtual mental health care including scheduled counseling, psychiatry, and self-care content to support wellbeing and mental health any time. Scheduled therapy appointments are readily available.
{:.no_toc}
As an institution that values diversity and inclusion, our goal is to create learning environments that are usable, equitable, inclusive and welcoming. Harvard University complies with federal legislation for individuals with disabilities and offers reasonable accommodations to qualified students with documented disabilities and temporary impairments. To make a request for reasonable accommodations in a course, students must first connect with their local disability office. The HMS Director of Disability Services, Timothy Rogers ([email protected]), is the point of contact for accommodation information for HMS master’s and MD students.
Accommodations are determined through an interactive process and are not retroactive. Therefore, students should contact their local disability office as soon as possible, preferably at least two weeks before accommodations are needed in a course, or immediately following an injury or illness, in order to initiate the accommodation process. Students are strongly encouraged to discuss their access needs with their instructors; however, instructors cannot independently institute individual accommodations without prior approval from the disability office. Student privacy surrounding disability status is recognized under FERPA. Information about accommodations is shared on a need-to-know basis, and with only those individuals involved in instituting the accommodation.