About
Table of contents
Reading reflections
Every class will include assigned readings. For each class, I expect you to write a reading response to the assigned readings, with the exception of classes where we will be presenting or discussing project updates (marked in the schedule).
Reading responses are due by 7 pm the day before each class. This is to give the discussion leads time to read the responses and prepare discussion questions for class. There are a total of 16 reading responses, to accommodate conflicting deadlines, emergencies, or anything else that might come up during the semester, you are allowed to skip 3 reading responses, no questions asked. You also do not need to submit a reading response for the papers you are a discussion lead on.
Submit responses via Canvas discussion. If you are leading a discussion, you do not need to prepare a discussion response (see below for what to prepare if you are leading a discussion). The Canvas discussion will be closed after 7 pm the day before class. If your reading response is not in by then, it will not be graded.
Reading reflections format
The reading reflection will be about 2 paragraphs (5-10 sentences), focused on what you learned from the readings, connections to other readings, and critical/thought-provoking questions you had. Please do NOT just provide a summary of the papers. Remember that everyone else will also be reading the same papers. To give you an idea about what a reading response could look like, here are some (non-exhaustive) options:
- Option 1: One paragraph per paper summarizing the contributions of the work(s). Write about potential improvements and how someone could build on the existing work.
- Option 2: A single paragraph for all papers that fits the papers into an overarching theme. This is not a summary of the papers, but should instead answer questions like: what is the overall topic of the readings? Why are we reading them all together?
- Option 3: A single paragraph for all papers in the reading list for that day that asks 2-3 thoughtful discussion questions.
- Option 4: A single paragraph in which you respond to someone else’s reading response (e.g., answer or expand on questions they pose) from the same day.
- Option 5: Find a new paper that you think builds on one of the readings. Briefly summarize the paper and explain what new ideas it introduces to the current reading.
Also as a general rule, the reading response is mainly there to help you remember the paper and to help the discussion leader summarize different views. When in doubt, please ask me! Send me an email and I will look at the response.
Reading response grading
Reading responses will be graded on 0-3 point scale. As noted above, you may pass on 3 of the reading responses. Any additional missed responses will receive a 0.
3 points = the reading response satisfies one of the options above, is 5-10 sentences in length, well-written, and thought-provoking (i.e., you got some ideas from it).
2 points = the reading response is too short/long, or does not show that you have engaged with the paper.
1 point = the reading response is lacking two or more of the above requirements.
0 points = no submission
The options for reading responses mentioned above will be equally graded.
Leading a class discussion
Each week, a pair of students will lead the class in a ~1 hour discussion of the assigned papers for that week. You are expected to sign up to be a discussion lead once during the semester. Your responsibilities for a discussion lead are:
- Read and understand the assigned papers that week along with any supporting material that you think would provide helpful background. Please also read the optional readings for that day, if any.
- Present a 10-12 minute talk on the papers. Summarize the papers’ key ideas, research questions, methodology, and findings.
- Lead a class activity based on the papers, followed by a discussion of the reading responses that week. To keep the discussion going, read all reading responses by the other students that week and integrate them into the discussion. You should prepare 8-10 discussion questions about the papers. These should be creative questions that do not have easy answers. The questions should be the result of your attempts to understand the material, its implications, and its connections with other work.
- Double-check with the instructor at least 24 hours before class regarding your proposed group activity and the contents of your slides. This does not need to include your responses to reading responses. We will be able to provide feedback if you send us the materials the day before class.
Below are ideas for group activities and general questions from past classes. Here are some group activities that worked well in prior classes:
- Role-playing activities: have students within groups analyze the same problem/question by assuming different roles as stakeholders: e.g., writers, news reporters, artists, patients or caregivers, moderators, developers, researchers, policy makers.
- Design activities: have students come up with new design ideas for additions/modifications to a tool or solution. How do we redesign existing applications, or design completely new ones, to implement ideas or address certain key challenges identified in the paper?
- Extending current research: have students collectively brainstorm how they would extend or modify the paper, and why. In addition, consider discussing the strengths and weaknesses of the current paper’s approach, and their favorite parts in the paper.
- Class surveys based on the paper (e.g., google forms): This has served as a great way to understand overall patterns and variations in class opinions regarding design and challenges in intelligent communication technologies.
Tips
Break students up into small groups (3-4 people max) Initiate discussion with example answers to the questions you pose. Introduce hypothetical or real-world examples (e.g. products or recent events) to frame the discussion Tie your questions to research that you or other folk may be conducting. Keep a timer.
Contribution statement
We expect both discussion leads to participate equally in the presentation, activity, and project. To help ensure equal participation. Each discussion lead should include 1-2 sentences documenting their specific contribution to preparing and leading the class discussion.
Class participation
You are expected to attend and actively participate in every class session. To make the most of this class and not disrupt others, I expect you to be on time and ready to discuss the readings. If you can’t make it because of sickness or a family emergency or similar, please let me know before class.
Class Project
A major goal of the course is to help students critically reflect and build upon the ideas introduced in class. You will work on a group research project related to language, interfaces and communication throughout the semester. We have various milestones for the project below.
Project pitches
Pick and develop a research idea or question that interests you, develop your own views, and come up with pitches for two new research projects. The projects can fall into two broad categories:
A data project analyzing language use in some communication context (e.g., conversations on reddit, the wording of security and privacy instructions). Note that the communication does not have to be two-way, social or synchronous: it could be conversations on reddit (two-way, social communication), or newspaper articles (one-way).
A design project proposing a new system (or modifying an existing one) to enhance communication through language.
More details on the two directions below. Use the research pitch session to get early feedback from the instructor and your peers on topic selection and potential resources for literature survey. Make sure to include research questions (RQs), why the problem is important (why), and how you plan to solve it in each research pitch (how). There are some great resources online on how to approach research and selecting what problems to work on (or not):
- Read this section on “Selecting Problems” on Page 6 in the “Syllabus for Eric’s PhD students”.
- First two sections of Daniel Fried’s lecture on conducting NLP research
- Two more sets of suggestions from other fields:
Time limits: Try to keep your pitch/presentation to 6-8 minutes in total. This will be followed by 3-4 minutes of Q&A from the instructors and the rest of class. Please practice beforehand to not run out of time!
Guidelines for project pitches
For each project pitch, submit 2-3 paragraphs that covers:
- Clear description of a new problem statement and the current gap that you aim to address through your research idea (examples).
- An argument, with citations to relevant literature, that these are important research questions that we likely don’t have answers for yet.
- Initial thoughts on research questions and methods you would employ to answer these.
- Clearly articulate RQs/Why/How in pitches.
Project resources
We have compiled some databases, codebases, and compute credits that might be helpful for planning and executing your project. Here is a list of potential resources. Let us know if you have questions about any of it.
Project proposal talk
After selecting one research pitch, come up with proposal for a new research study or system design. Incorporate feedback from the class obtained during the pitches, and remind us how you propose to address any challenges during your proposal.
The proposal will include a written document (the project proposal report, covered below) and talk. For the talk, be sure to cover:
- Research questions that can be answered with empirical results (RQ/Why/How).
- A complete research study or system design. For example, coming up with either a quantitative study to answer clearly-defined research questions or building a new system, chrome plugin, or ML framework to solve some current challenge. You’re free to choose your favorite methods that are appropriate to answering your research questions as long as you justify your methods.
- For a data project: a plan for how you’re going to collect data and drawbacks to your approach.
- For a design project: the feasibility of your system design and how you would address any challenges that may arise.
- Timeline for the proposed project, clearly identifying next steps and possible roadblocks or risks.
Project proposal report
We will be using the CHI paper format for both the project proposal and final report. Your project proposal report will be a first draft of your final paper that you will continue to fill in over the course of the semester. The project proposal report should reflect what you presented in the proposal presentation. It should include an introduction, related work, proposed study design and methodologies, timeline, distribution of work among students, and expected project outcomes and contributions. For the project proposal (not the final report), try to keep the PDF to under 7 pages.
Note: A single-paragraph collaboration plan will be required in each student’s project proposal to receive full credit. Eventually, a single-paragraph contribution statement summarizing each group member’s contribution to the project will be required in each student’s final paper.
What do you need to provide in the collaboration plan?
Each group member should include a single paragraph (ideally at the end of their proposal) documenting their group’s plan to complete the project collaboratively and specifically highlight their specific role as a collaborator. In other words, what is/will be your contribution to the group’s successful completion of the 1, project and ii) proposal document + final paper.
Think of this as an informal contract among group members.
Final Report
Every team will hand-in a 10 page report summarizing their findings. When writing up your findings, look at the research papers we have read in class. My expectation for the final project is that it significantly extends previous findings in an area within CS. Your research project should be a plausible attempt at contributing a novel result. The final write-up should have the quality of a publishable paper, with the contribution being roughly that of a short paper (e.g., a workshop paper). The papers should be a maximum of 10 pages.
For guidance on writing research papers, please refer to papers you enjoyed reading in class, here are also two links for writing HCI and NLP research papers.
Final Talk
You will also present your project in a final talk. The talk should be a total of 10-12 minutes. It should include:
- Problem statement. Your presentation should open with a short (i.e., < one minute) description of the problem you are trying to solve. This should be easy if you pick a pitch you’ve already made. You simply need to remind us of your problem. If you choose something you haven’t pitched in the past, then you can spend up to two minutes on the problem statement.
- Implementation (e.g., RQs, data, methods). Tell us about your overarching research design, research questions, data collection, how you answered these questions.
- Key takeaways and findings. You are free to reuse any of your presentation materials from earlier in the class.
Grading criteria
- Identifying an interesting problem.
- Quality of problem statement.
- Addressing problem statement with social data or a proposed system.
- Quality of presenting solutions in report.
- Thoughtful use of course readings in presentation.
- Satisfactory participation as a contributing member.
What do you need to provide in the contribution statement?
Each group member should include a single paragraph (ideally at the end of their final paper) documenting their specific contribution to the group’s successful completion of the: 1, project, and 2, proposal document + final paper. Include details from each stage of the project: ideation, formulation, execution, presentation.
I will be asking all group members to individually provide member evaluations to me along with individual contribution statements in the final papers, and allocating points based on satisfactory participation in group projects.