11-631: Data Science Seminar - Syllabus

Course Learning Outcomes

This course introduces students to the breadth of data science, covering human-centered, analytic, and systems approaches, through exposure to a wide variety of research topics and literature. Emphasis is placed on developing core academic skills: reading, writing, presenting, critiquing, discussing, and researching in data science. Students will collaborate to analyze publications, synthesize ideas from diverse subfields, and effectively communicate their insights both individually and in groups.

• Gain exposure to the breadth of data science literature, including human-centered, analytic, and systems-oriented research, as well as relevant application areas, venues, and writing styles.
• Learn how to critically read, evaluate, and discuss data science publications, justifying academic assessments of specific works.
• Develop skills for writing academic papers and reviews, by synthesizing research content from multiple perspectives.
• Present research papers in a clear, comprehensive, and collaborative manner, connecting a given publication to related works and broader themes in the field.
• Learn utility and limitations of using GenAI to help with distilling and understanding data science literature.

Time & Location

TR 05:00PM- 06:20PM, TEP 1403

Course Format

In-Person. The course opens with an initial overview of the Data Science literature and tutorials on how to analyze and critique Data Science publications. The course also provides tutorials on preparing and presenting reviews of Data Science publications and related literature.

Course Organization

The main objective of the course is to get familiar with critically reading, reviewing, and presenting data science papers, to prepare you for the capstone courses. As such, the course will involve the following activities which each student is expected to participate in (N: number of times the activity occurs):

• Attend lectures and guest lectures (6 or 7)
• Read paper, write a short summary according to a specialist role, and discuss it with a group (3)
• Do a practice (1) and main (1) in-class presentation (individual or in pairs) for papers that you have read
• Attend in-class presentations by your colleagues, and ask a question for each presentation (9 to 11)
• Write a paper review (1) and a literature survey (2) according to a theme of your choosing
• Write a constructive review of a capstone report (1) and of a capstone presentation (1)

Main assessments

Presentations

A main goal of the class is to learn how to clearly and effectively present research/papers to others, which is a core component of being a data scientist. There will be two opportunities to present papers in class:

• Practice presentations: Each student will present a paper to a smaller audience over Zoom, and will receive feedback from one of the TAs.
• Main presentations: Each student (individually or in pair) will present a paper to the entire class.

Presentation length: There will be 4 presentations per class, which means each presentation must be at most 15 minutes long, with 5 minutes for questions from the audience.

Audience questions: see below.

Selection of papers: For the practice presentations, you will be assigned the paper that you have to present. For the main presentations, we will ask for paper suggestions from all students, and then assign presentation papers taking into account who nominated what paper.

Presentation format and visuals: Please refer to this guide for tips on how to present your paper.

Audience participation (main presentations)

During the main presentation phase, those who are not presenting that week must fill out an audience form with a question for each of the papers. You do not need to ask a question on days where you are presenting yourself.

The instructor will randomly select a 1-3 students to ask their question out loud for each paper. You are not allowed to ask ChatGPT or other GenAI tools to produce a question for you (see GenAI policy below).

Questions will be both graded as your attendance and for question quality. The reason for the latter is that it’s important to be able to ask good specific questions for new research.

Rubric for questions:

• 0 points: no question
• 1 point: low-effort or low-hanging fruit question (e.g., How does [newest OpenAI model] do on this? How would this work on other languages? Is there a scenario where [method/system] fails? How fast is the system? Can the system design be extended to other networks?)
• 2 points: good question

Summaries and Discussions

Another main goal of the class is to learn how to critically read a paper, analyze it through different perspectives, and have productive discussions about research with others. There will be a total of three discussion lectures, one for each of the three data science areas covered in the course.

For each discussion:

• You will be assigned a paper to read before class.
• You should print the paper and are encouraged to annotate it as you read.
• In class, you will be assigned a discussion role, which provides a particular lens through which to analyze the paper.
• During the first 20 minutes of class, you will write two short summaries by hand, without laptops or internet access: (1) a 3–5 sentence general summary of the paper and (2) a 7–8 sentence analysis through the lens of your assigned role.
• You will then join a small group and spend the remainder of the class discussing the paper from your different assigned perspectives.

The five possible discussion roles are:

• 🔬 Researcher: identify a meaningful follow-up study or research direction.
• 🏭 Industry Expert: consider how the work could be applied and deployed in the real world.
• ⚖️ Social Impact Assessor: examine potential societal benefits, harms, and unintended consequences.
• 🧐 Skeptic: identify the strongest reason to be cautious about the paper’s conclusions.
• 🔗 Connector: analyze the paper from the perspective of another area of data science (Human-Centered, Analytics, or Systems).

See the Paper Discussions assignment on Canvas for detailed instructions and descriptions of each role.

Review & literature survey

To teach you how to distill open questions in specific sub-areas of data science, you will learn how to write a literature survey of at least 8 papers. You will do this in the topic area of your choice, in a group of at most 4 students. This will encompass four steps:

• Team and topic choice: you will have to come up with a topic area, 4 relevant papers, and find your teammates.
• Paper review: you will each write a review of a paper, focusing on the remaining broad questions that the paper leaves open. Each teammate will get one of the papers you submitted. Write at most 1.5 pages of content.
• Draft literature survey: focusing on your task and topic area, summarize and compare your 4 related papers’ similarities and differences, as well as the remaining questions that the papers leave open. Use GenAI to make a first draft, critique that draft, and improve upon it.
• Literature survey: staying in the same topic area and task, you will choose an additional 4 papers with the help of GenAI, and incorporate them into the draft literature survey (8 total) using a similar GenAI-critique-improve sequence.

For more details, all assignments from this sequence can be found on Canvas.

Capstone project review

Finally, during the last two weeks of class, students will attend at least one final second-year capstone presentation and review one draft capstone report written by a second-year MCDS student team.

Extra credit assignments

LLM-review critique

Extra credit: Due at the same time as your paper review, you can optionally turn in a comparative critique of how a GenAI system would review the paper you were assigned. More details in the assignment on Canvas.

Science communication

Extra credit: Create a short piece of social media content (e.g., TikTok/Instagram Reels/YouTube shorts video, henceforth: content piece) on a concept related to data science. Along with your video, you will submit a short write-up that describes your process (e.g., daily journal log, production diary, challenges), contributions of each team member, and answers to the rubric questions (see requirements below). More details in the assignment on Canvas.

Attendance Policy

This course will be held in person. You are responsible for completing the work assigned and seeking clarification as needed. Late work is generally not acceptable.

• Attendance is required for:
  • Lectures and guest lectures
  • Discussions (absence will incur a penalty)
  • Assigned presentation slots (practice and real)
• For presentation phase, on days you are not presenting:
  • You get a total of 3 absences. If you know that you have a conflict (e.g., internship interview, medical procedure), please do let the instructors know that you will be absent. But this is not strictly necessary.
  • For any further absences during presentation phase, you will not get attendance points for that lecture.
  • If you get caught posting questions without being in class, you will get 3 absences worth of penalty.

Late submissions

We do not allow late submissions.

• For presentations, if you do not upload your slides by 5pm day of, you will lose points. If you are late to class, you will also lose points.
• For paper discussions, if you are late to class, you will also lose points and will have less time to write your summary.
• For other assignments (e.g., paper review, literature survey assignments, capstone review, etc.), for every day that you submit late, you will receive a deduction proportional to the lateness (10% per late day; if you’re 2 days late, you will lose 20% of your grade). You have a 30-minute grace period before the lateness deduction will be applied (i.e., you can submit up to 30 minutes after the deadline without any penalty, but any minute after will result in 10% deduction).

Issues with presentation schedule

For main presentations only, you are allowed to swap dates with another student exactly once. If you chose to swap main presentation slots (e.g., due to a job interview), you and your swapping partner must let the instructors know via email at least 3 days in advance (otherwise you will get a zero on the presentation assignment).

Assessment

AIV, Plagiarism, and GenAI Policy

Collaboration policy: For preparing each presentation and literature survey, you must only share work with your assigned teammates and no other students. Paper summaries, reviews, and capstone reviews are individual assignments. This course is intended to give you experience in autonomous research, so trying to delegate or shortcut preparation is a wasted learning opportunity. Acting against this rule will be considered an academic integrity violation and lead to reprimands, including possible dismissal from the program (see the MCDS Handbook).

Plagiarism and AIV policy: The presentation and related work survey emphasize a literature search and compare/contrast to other material. All material you find and use in any of the course deliverables must be explicitly and correctly referenced/cited. Notes:

• Directly copying text from the paper being summarized, and/or from author websites or other sources, without using “quotation marks” around everything that is a direct quote, followed by a reference to the source being quoted, is plagiarism.
• Text and/or slides copied directly from other sources without attribution in presentations is also considered plagiarism.

Here are some resources for learning what is and isn’t plagiarism:

• Paraphrasing and Direct Quotations
• Paraphrasing and Plagiarism
• Paraphrasing Spectrum

GenAI use policy: The goal of this course is for you to learn how to read, present, and discuss papers in various data science tracks, as well as distill open questions in a subtopic of your choice. We know GenAI can do this too to some degree, so instead of banning GenAI, we are actively encouraging you to think critically about GenAI outputs. But you are still responsible for all the work you turn in (you’re the one getting a grade), so you must ensure correctness and matching of the grading criteria.

It is very easy to tell when a student is not actually familiar or has not actually understood the material. If we suspect or confirm that you turned in something AI generated or relied too heavily on AI for your assignments, we reserve the right to ask you to justify your turned-in assignment, waive all your grades in that homework category, or even report an academic integrity violation (AIV).