Reading Summaries
You are expected to submit one written summary per assigned paper on Quercus. The summaries are due 24 hours before the corresponding lecture. Each paper summary should include the following five components:
| Component |
Required Length |
Description |
Notes |
| Brief summary |
4–5 sentences |
A brief summary of the contribution. |
This should not be a regurgitation of the abstract, but rather your own interpretation of the work. |
| Commendation |
4–5 sentences |
What did you find noteworthy or unique about the paper? Why do you think it was accepted? |
This should be a thoughtful reflection on what you believe is interesting about the work, not a restatement of the paper's contributions. |
| Two critiques |
4–5 sentences per critique |
What flaws do you see in the work? Why are these flaws problematic? How would you go about addressing them? |
These should be two distinct and coherent points. If your critiques cover more than two topics, only the first two will be graded. |
| Discussion point |
1 sentence |
A thought-provoking question and/or talking point that you would like to raise during the in-class discussion. |
Student presenters will refer to these talking points as they lead the in-class discussions. |
| Enjoyment rating |
A number |
A score from 1 (boring) to 5 (very interesting) indicating how much you enjoyed reading the paper. |
The instructor uses these scores to decide whether papers need to be rotated out of the syllabus for future course iterations. |
Your reading summary should uniquely apply to the corresponding paper. In other words, you will not be given marks for “cop-out” commendations and critiques that could apply to most papers. Some examples are provided below:
| Don't say... |
Instead... |
| "The paper was well written." |
Explain what made the paper readable with specific examples (e.g., informative figures highlighting clinical accuracy, explanations of domain-specific terminology). |
| "The paper was well motivated because they are tackling an important health topic.” |
Explain how the proposed solution can make a difference in the target domain. |
| “The authors had a small sample size. They should have recruited more participants.” |
Explain why the sample size isn't representative of the target population and why this shortcoming is important. |
Lecture Time and In-Class Presentations
Most weeks will have two required readings on a similar topic. You are expected to read the papers before each lecture to prepare for engaging and meaningful in-class discussions. The lecture time will be split into two parts: The corresponding lecture time will be split as follows:
| |
Presenter |
Agenda Item |
Duration |
| |
Instructor |
Topic Overview |
10 minutes |
| Paper 1 |
Student 1 |
Paper Presentation |
10 minutes |
| Student 1 |
Discussion |
40 minutes |
| Paper 2 |
Student 2 |
Paper Presentation |
10 minutes |
| Student 2 |
Discussion |
40 minutes |
Each student should sign up to present a paper at some point during the term. The expectations for presenters are summarized below:
- Paper Presentation: The purpose of this presentation is to refresh everyone's memory about the paper for the discussion that will follow. In other words, you need not and should not go over every single design detail, result, or figure in the paper, but rather remind people about the overall contribution and the way that the authors went about making that contribution in their paper. Note that the instructor will provide a brief introduction about the day's topic, so you should scan through the slides that are posted on the course's website ahead of time to ensure that you will not end up repeating material.
- Discussion: Presenters are also expected to lead a discussion about the paper with the rest of the class. You should prepare for this discussion in advance by having a list of potential talking points listed somewhere in your slide deck. You can refer to your classmates' reading summaries in order to identify topics that may be of interest to the group. Since it is usually easier for people to criticize other people's work, try to ensure that the discussion points also highlight the admirable aspects of the paper. Leading the discussion not only requires preparation but also active involvement and moderation during the lecture time itself. If there is a lull in the discussion, you should either provide additional encouragement to your peers to provide their thoughts, extend the conversation by providing your own thoughts, or introduce a new topic of discussion. Some students choose to incorporate parts of the discussion into their presentation in order to make the conversation more fluid and spread out. If you do this, reasonable adjustments will be for the timing expectations in the grading.
In-Class Participation
You are not expected to have something insightful to say about every paper we read during this course. However, being an active participant in the majority of the discussions is integral to your learning in this course. You will get a reasonable participation grade as long as the instructor is easily able to recall your name and face based on your level of engagement in the discussions.
Group Project
Throughout the semester, you will work in groups of 2 students on a project related to mobile health. The only requirement for the topic of your project is that it must involve some sort of contextual or sensor data from a portable device (e.g., smartphone, wearable, implant). Some examples of acceptable data sources may include, but are not limited to: IMU, camera, microphone, location, time, and app usage. Since the course is intended to satisfy curriculum requirements related to quantitative methods, the focus of the work should involve such techniques (e.g., machine learning, signal processing). Nevertheless, you are encouraged to use complementary methods when applicable.
You are encouraged to leverage any clinical collaborations that you might already have in place, but such a connection is not a requirement for your project given the limited time we have. Your project may involve a new data collection effort that can be completed during the semester, but it can also involve a new analysis of an existing dataset. You are highly encouraged to adapt your project to your research area to make it more relevant to your passions. Some examples of past projects are listed below:
- Evaluating Community Mobility Data for COVID-19 Forecasting: In this project, we aim to investigate the utility of community mobility data from Google's COVID-19 Open Data Repository for the task of COVID-19 infection forecasting. We evaluate the performance of common prediction models when given or withheld access to community mobility data. We perform this evaluation on caseload inflection points, where current methods perform worst, to investigate whether the inclusion of community mobility data as an input feature can improve performance.
- Mental Fatigue Detection in Computer-related Tasks Using Eye Tracking and Monitoring Input Devices: In this paper, we developed a computer task to induce fatigue while collecting user-input and eye-related measurements in addition to user-reported fatigue levels from 14 participants. Our goal was to find a correlation between collected user data and the user-reported fatigue level.
- Lab to Field Transitions in AI Models: Understanding Performance Deterioration and Shortcut Learning in Healthcare Data: Many deep learning-based tools are initially developed within smaller lab and research environments and then deployed for a larger population in the field. As a consequence of this, the AI models might pick up biases and spurious shortcuts from the lab data which does not generalize to the field. We introduce a novel framework that utilizes a shuffled input model to estimate data bias and predict performance deterioration.
- Evaluating the Effectiveness of Idle Games in Promoting Adherence to Deep Breathing Exercises: To motivate short and frequent practices of deep breathing exercises, we developed an idle game in which the core action is deep breathing. After validating the physiological efficacy of the breathing guide embedded in the idle game through an in-lab study with physiological sensors, we conducted a two-week study whose results demonstrate that idle games can better sustain deep breathing adherence than a standard guide.
If you are stuck trying to come up with an idea, the following resources and datasets may be helpful:
There will be deliverables throughout the semester to ensure that your group is making sufficient progress on your project:
- Proposal: A document proposing your idea and your execution plan (1–3 pages)
- Progress Report: A document describing your progress halfway through the semester (6–8 pages)
- Final Presentation: A presentation summarizing your project (~10 minutes)
- Final Report: A document akin to a journal or conference submission (10-14 pages)
All written documents should follow the
ACM Large format (single-space, single-column).
Late Policy
Late submissions will not be accepted for the weekly reading summaries. Late submissions for any of the project milestones will incur a penalty as follows:
- <24 hours late = –10% penalty
- <48 hours late = –30% penalty
- >48 hours late = –100% penalty
Academic Conduct
All the work you submit must be done by you (individually or within your group). Please refer to the university's
Code of Behaviour on Academic Matters for more information.