Assignments

Weekly assignments will (generally) be due 24 hours before class each week (Tuesdays at 11am) but sometimes 48 hours before (Mondays at 11am)

Piazza

We will be using Piazza for informal questions about homeworks or other class discussion. The system is highly catered to getting you help fast and efficiently from classmates and the instructors. Rather than emailing questions to the teaching staff, please be encouraged to post your questions on Piazza. If you have any problems or feedback for the developers, email team@piazza.com. https://piazza.com/mit/fall2018/mas630/


HWK11: Project work week and (optional) reading

There is no homework to hand in on Thanksgiving week, although we will have class as usual.

You might enjoy skimming or reading the below articles before the presentations next week, also in case you have questions you might want to ask the presenters.

Since this is a 12-hour/week class, we STRONGLY recommend that you put in 10 hours of work outside of the 2 hours in class this week to move your project forward. Please put in these hours now and be more likely to stay healthy through the end of the term.

Your final project will consist of a presentation (about ten minutes long per person) that you will give to the class either on Dec 5 or on Dec 12. That is coming SOON so please let us know how we can help you. Those teams or soloists who present on the 5th will not be expected to be as far along as those who present on the 12th. You will present as much as you have done and it is expected that some projects will be further along than others. Final scheduling of presentations will be announced by Nov 29.

Everybody will be expected to hand in a final paper due by midnight on the last day of class. Here is a MIT's rule: MIT allows an instructor to give an extension to an individual student for an assignment, but blanket extensions are not allowed to be given to the entire class. If you find on the last day that you need an extension, then please just ask.

The format of the paper is up to you. Please be encouraged to find a conference that interests you and write up your project using that conference's paper format. If you don't have another one in mind, then I recommend the ACII format - see for example their guidelines at http://acii2017.org/content/camera-ready-instructions (there will soon be a CFP for the 2019 ACII in Cambridge, UK). A typical soloist paper might be 4 pages, perhaps with additional figures or plots of data as supplemental information. Group projects might be up to 8 or even 10 pages. The page limit is not strict, but it is preferred that you write concisely and clearly about what you've done, include a descriptive background reference section, and include descriptions of your motivation for your work, experimental design, participants, data analysis, results, or whatever best fits the protocol of the place where your work would be most likely to get published. We'll be happy to discuss this more with you and feel free to arrange appointments with Roz and/or Oliver for any help.

Optional Readings Sentiment:

Optional Readings Emotion Context:

Have a Happy Thanksgiving!



HWK10: Presentation in class

DUE Wed Nov 14 by 10am: Upload your slides to Stellar; also please bring them to class at 11am on a laptop (ideally w/HDMI port or dongle)

Prepare a presentation with the following content (6 mins soloists; 12 mins duets; 18 mins trios)

Opening slide: Title of project, your names, "Twitter-sized description" of what the project is about, status of any COUHES/IRB (approved, submitted & iterating, exempt, etc.)

Next slides: "Work done": Describe key work you've done on your project so far. You could show your completed study design and a nice plot of the flow of the participant's planned experience. You could show examples of playing with the technology you're planning to use and how it works. You could show examples of how you are practicing collecting data, learning what data are good/not good, or planning to analyze your data once you get it. Share with us something you've learned in a way that others in the class will learn from your efforts.

Next slide(s): "Work to do" Describe what remains to be done and your timetable to do it. Feel free to add "HELP NEEDED" lines.

Last slide(s): For each person on the team, list four references related to your project -- for each one, give a couple sentences of its main contribution or otherwise make clear how it is related to your project. Provide also the full citation for each reference (as you would publish in a journal - click on the quote mark listed by the article in scholar.google.com for citations (in various formats) if you wish to see examples.



HWK9

DUE Tuesday Nov 6 by 11am:

1) Download Wavesurfer (https://sourceforge.net/projects/wavesurfer/) and figure out how to use it

2) Audio annotation: For each of the following three clips, examine it using wavesurfer (specifically, look at the pitch contour and the spectrogram) and use the annotation panel to find the onset (the start time of the vocal emotion indicated below) and the offset (the end time of the vocal emotion indicated below)

a) Find one audio clip where a person sounds angry or record yourself sounding angry

b) Find one audio clip where a person sounds frustrated or record yourself sounding frustrated

c) Find one audio clip of where a person is laughing/giggling/chuckling or record yourself doing the same.

For each clip, please take a screenshot of Wavesurfer, and write a few sentences detailing what you noticed about the audio characteristics relative to the other clips. Put the screenshots and your reflections in a PDF and submit via Stellar.

3) Read this paper: https://www.researchgate.net/profile/Christos-Nikolaos_Anagnostopoulos/publication/257513035_Features_and_classifiers_for_emotion_recognition_from_speech_a_survey_from_2000_to_2011/links/548cba940cf225bf66a29b71/Features-and-classifiers-for-emotion-recognition-from-speech-a-survey-from-2000-to-2011.pdf


HWK8

DUE Tuesday Oct 30 by 11am:

READ/WATCH & PREPARE:

Picard, Affective Computing Chapter 4 "Potential Concerns"

Cowie, Oxford Handbook of Affective Computing, Chapter 24 "Ethical Issues in Affective Computing" (Stellar materials)

Watch Kate Darling's talk https://www.ted.com/talks/kate_darling_why_we_have_an_emotional_connection_to_robots#t-93180

(1) Answer: What is a potential concern of AC technology that most worries you? Give one or two lines saying why.

(2) Answer: What do you think is the best approach that we (the research community) could take to mitigate or possibly prevent this concern?

(3) Upload on Stellar your 1-2 slides with bullets on your views on the debate issue assigned to you, naming them YOURLASTNAME.pptx. Do not write out your arguments on the slides, just put prompts, keywords, or images to help you remember your main points. Please coordinate among your debate team so you each make different points on the slides you present.



HWK 7

THERE IS NO CLASS NEXT WEEK, OCT 24, because of Media Lab Members' meetings. You do not have to upload any homework this week. However, please make sure you do (1) and (2) by the dates/times below:

(1) BEFORE 11am Monday OCT 22, Fill out your ethical discussion preferences on the form at https://docs.google.com/forms/d/1MMnLI9qkd_Hccp6DyIDfrIcegwGYTuwg4D28LyhlEbE/edit

What you put on this form will be used to assign debate groups, which we'll mail to you by Wed Oct 24. You will then need to plan to meet with your group before class Oct 31, and then to present in class on Halloween, Oct 31, your arguments for your position.

(2) Make sure your COUHES form is completed and signed by your PI and by your Dept Head by OCT 24, and delivered to COUHES by OCT 25 (note: the last date you can get Prof. Picard to sign off is Fri OCT 19 unless you arranged something else with her).



HWK 6

Due 11am, MONDAY, Oct 15th, 2018.

READ:

Ghandeharioun, A., Picard, R. "BrightBeat: Effortlessly Influencing Breathing for Cultivating Calmness and Focus," Proceedings of the 2017 CHI Conference Extended Abstracts on Human Factors in Computing Systems. ACM, Denver, May 2017. (on stellar)

Jean Costa, Alexander T Adams, Malte F Jung, François Guimbetière, and Tanzeem Choudhury. 2016. EmotionCheck: leveraging bodily signals and false feedback to regulate our emotions. In Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing. ACM, 758–769. (on stellar)

Jaques & Taylor et al. "Multi-task Learning for Predicting Health, Stress, and Happiness," NIPS Workshop on Machine Learning for Healthcare, December 2016, Barcelona, Spain.


REMINDER: (Sorry if this reminds you of Javier's "stress-eliciting" example!) COUHES forms are due to MIT COUHES by Oct 25 and must be signed by your PI and by your department (ink on paper) before COUHES will take them. This means your application, consents, recruiting materials, EVERYTHING must be signed off before Oct 25. Ideally your PI=YourThesisAdvisor. If that can't work and you need Picard to be your PI, then you must get Picard **paper versions** of all of your COUHES materials by Oct 18 1pm latest. Please deliver them to Picard's office office (E14-348A) by 1pm Thursday Oct 18.


WRITE & SUBMIT on STELLAR by MONDAY Oct 15, 11am:

(1) Who is the PI on your COUHES form (or equivalent?)

(2) Have you confirmed with them that they have agreed to be PI on this and have you checked that they and your department head (or equivalent) will have time to sign off in time?

(3) Finish a "good first draft" of your experiment design by Monday 11am and upload it as TEAMNAMES-exp for Oliver and Roz to read. We would like to help you make the design and will plan to give you feedback ASAP. The level of design description should be the same as that in the example COUHES forms that you already read. (a) Make a diagram of the conditions in your experiment - for example, see Fig 3. in the paper by Ghandeharioun. (b) Make a wild guess and sketch a graph like Fig 5. in the Costa et al paper showing "what you wish your study could show" in terms of a hypothesized outcome.

(4) Both Oliver and Roz will post extra office hours in case you'd like to meet to get help on your study design - please book with us by Monday 11am if you want help - using our appointment calendars or emailing us with MAS.630 in the subject line. Write here that you've either booked a slot or you don't need to right now.

Please be ready by class Wed to ask the authors Ghandeharioun and Taylor a question based on their papers this week. As you read the Ghandeharioun and Costa papers, think about their experimental designs, which may help you with your project experiment design. As you read Jaques & Taylor, think about what a system would have to measure in your life to accurately forecast your stress, health and happiness.



HWK 5

Due 11am, Tuesday, Oct 9th, 2018.

SUBMIT YOUR HOMEWORK FILES DIRECTLY TO STELLAR

Homework answers should be uploaded by Tuesday 11am with filename YOURLASTNAME-HWK5.PDF

READ:

  1. Barrett, L. F. (2013). Psychological construction: The Darwinian approach to the science of emotion. Emotion Review, 5(4), 379-389. https://www.affective-science.org/pubs/2013/psych-construction-darwinian.pdf
  2. Hernandez, J., McDuff, D., Benavides, X., Amores, J., Maes, P., & Picard, R. W. (2014). AutoEmotive: bringing empathy to the driving experience to manage stress. In Designing Interactive Systems (pp. 53 - 56). (on stellar)

WRITE:

  1. Pick an emotion and describe what is unique about it - how do you know you feel a particular emotion?
  2. Have you ever felt unsure what emotion you were experiencing? Using the theory from the Barrett reading, explain how such an experience might occur. Could the "natural kinds" theory of emotion also accommodate such an experience?
  3. Describe one way that your project is informed (explicitly or implicitly) by a theoretical model of emotions (either one described in the reading, or another model).
  4. Describe a situation in which high stress levels may be positive.
  5. In the AutoEmotive work, what stress interventions do you think are more likely to work and which are less likely to work? Propose one additional intervention in the context of stress management when driving that might work for you.



HWK 4

Due 11am, Monday, Oct 1st, 2018.

SUBMIT YOUR HOMEWORK FILES DIRECTLY TO STELLAR

Homework answers should be uploaded by Monday 11am with filename YOURLASTNAME-HWK4.PDF

Project proposal slides should be uploaded by Wednesday 10am. YOU WILL PRESENT THEM IN CLASS ON WED Oct 3. If you are collaborating, then you only need to upload the file to one team member's stellar account. Please name the file LASTNAME(s)-PROJ.PDF


READ & WRITE:

1. Read Fredrickson, B. L. (2016). Love: positivity resonance as a fresh, evidence-based perspective on an age-old topic. Handbook of emotions, 847-858. Write a question you would like to ask Dr. Fredrickson about her work, showing that you've read the paper above or related work of hers.

2. Read the three COUHES application examples on Stellar (under Materials/COUHES) - including the consent forms, examples of advertisements, and more. If you have any questions about the parts of these feel free to ask them here; else write "I understand the examples."

3. Find and read a peer-review published article, of your choosing, which is somewhat similar to what you are proposing for your project (please do some online searching before you propose your project). Ideally this is a recent journal paper or conference paper. Give the full citation and abstract for the paper here, and a link to its pdf. (If there are two people on your project, you should each read a different paper).

4. (due Wed an hour before class) Prepare a SHORT presentation of your proposed project to give in class (5 minutes for soloists, 8 minutes for 2-person team). Put all teammates name(s) on every one of your slides. One slide should clearly state: what is the problem (or question) you want to address? Another should state: What do you plan to test or evaluate or build? Finally, does your team have the skills you need to do this and if not, what skills do you wish you had on your team? Please feel free to interact on Piazza about your idea(s) or mail others if you are looking for a teammate and haven't found one. We will use these ideas presentations to help pair up more of you (if that seems good) and/or help you shape your projects to all be successful.


ATTEND IF YOU CAN & PLEASE HELP US PROMOTE THIS TALK ON YOUR SOCIAL NETWORK:

  • Barbara Fredrickson, Monday Oct 1, 2018, 2:00-3:00pm, E25-111 (near MIT Medical and the Kendall Square T)

Positivity Resonates: Effects of face-to-face social connection on human well-being

Are the moments of positive interpersonal connection we experience positive health behaviors? Positivity resonance is a momentary affective state co-experienced by two or more people simultaneously, marked by the amplifying trio of: shared positive affect; mutual care and concern; and cross-person synchrony in behavioral and biological dynamics. Dr. Fredrickson will share her most recent empirical evidence that positivity resonance – whether indexed by self-report, behavior, or biology – predicts mental and physical health, measured concurrently or prospectively.

Barbara Fredrickson is a Kenan Distinguished Professor and Director of Positive Emotions and Psychopsychology Laboratory at the University of North Carolina at Chapel Hill and the recipient of numerous honors for her research in positive emotions, including the American Psychological Association's inaugural Templeton Prize in positive psychology and the Society for Experimental Social Psychology's Career Trajectory Award. A pioneer in the field of positive psychology, Dr. Fredrickson's research illuminates the vital impact of positive social connection on psychological and physical well-being, and offers important implications for the fields of psychology, technology, and business as well as our personal physical and mental health. She is also the author of Love 2.0 (2013).





HWK 3

Due 11am, Tuesday, Sep 25th, 2018.

SUBMIT YOUR HOMEWORK FILES DIRECTLY TO STELLAR

Attach answers with filename YOURLASTNAME-HWK3.PDF


READ:

http://nymag.com/scienceofus/2017/01/how-interpersonal-synchrony-works.html

and your choice of either of these two overviews (A:Face or B:Interpersonal):

A: De la Torre, F., & Cohn, J. F. (2011). Facial expression analysis. In Visual analysis of humans (pp. 377-409). Springer, London.

B: Palumbo, R. V., Marraccini, M. E., Weyandt, L. L., Wilder-Smith, O., McGee, H. A., Liu, S., & Goodwin, M. S. (2017). Interpersonal autonomic physiology: A systematic review of the literature. Personality and Social Psychology Review, 21(2), 99-141. (pdf on Stellar)


Optional reading:


WRITE:

  1. What progress have you made toward defining your project this week? Please try to find a partner to collaborate and/or share the load (this is wise for helping with a study you carry out - e,g, work like recruiting people, getting data, etc can be shared, even if you have participants do two separate tasks once they're with you, either to contribute to two projects or to one combined project.)
  2. Have you met with Roz or Oliver to discuss project ideas and study designs? If not, please plan to do so (you can schedule an appointment slot with Roz or mail her & Mary or Oliver - use MAS.630 in the subject)
  3. Which reading did you pick: the De la Torre + Cohn or the Palumbo et al.? For this reading, answer:

(a) What is something you found surprising in this article? Jot a couple lines about what you learned that was surprising. If nothing surprised you, then perhaps you know a lot already about this area; please tell us what you think is an even better reading for learning about the most cutting-edge work on these topics.

(b) Craft a question for the lecturer that shows you have read these papers - and that you would like to hear them answer.

4. Download the AffdexMe app (free) on an iOS or Android device and play with it.

(a) Try to use it near a window on a bright day. Which way does it work best: when you are facing the window, or when your back is to the window? Why?

(b) Facial exercises! Click into the app's settings and change the metrics that are used, trying each one. Could you get them all to have a high score when you made the particular facial movements? Comment on which were particularly hard for you to achieve.

(c) Consider the ethical implications of emotion measurement of faces on large-scales: Give an example of how it might help improve your life.

(d) Give an example how it might make your life worse.

(e) What are some ways that the troubling uses could be addressed/mitigated?



HWK 2

Due 11am, Monday, Sep 17th, 2018. (Note: this is earlier than usual to allow everyone to review all project ideas, please submit no later than 48 hours before start of class)

SUBMIT YOUR HOMEWORK FILES DIRECTLY TO STELLAR

Attach answers with filename YOURLASTNAME-HWK2.PDF

Attach 1-3 slides (for project ideas) naming them YOURLASTNAME-ProjIdeas.pdf

READ:

Affective Computing: Chapter 5

Oxford Handbook of Affective Computing: Chapter 14, Physiological Sensing of Emotion by Jennifer Healey

R.W. Picard, E. Vyzas, and J. Healey (2001), "Toward Machine Emotional Intelligence: Analysis of Affective Physiological State,"IEEE Transactions Pattern Analysis and Machine Intelligence, Volume 23, No. 10, pp. 1175-1191, October 2001. (Just read sections 1,2, and 5 unless you're also interested in the machine learning details).

OPTIONAL: R.W. Picard, S.B.Daily (2005), "Evaluating affective Interactions; Alternatives to asking what users feel" CHI Workshop on Evaluating Affective Interfaces: Innovative Approaches, April 2005, Portland, OR

OPTIONAL: M. Zhao, F. Adib, and D. Katabi (2016), "Emotion Recognition using Wireless Signals" MobiCom '16, Oct 3-7, New York City, NY, USA.

WRITE/Prepare:

1. What is a question you have based on reading the chapters/article above? Submit this as YOURLASTNAME-HWK2.pdf

2. Prepare 1-3 slides of project ideas you have. On EACH slide put your name, and ONE of the ideas, and number the ideas "1 out of 2, "2 out of 2". For example if you have three ideas, the first slide should say "YOURNAME, Idea 1 of 3" and then give the following two points, being clear for each idea: (1) What would you like to build/test/study/learn? (2) What skills would you need to carry this out that you don't have? We will use this to seed discussions of projects and to help make project teams. Submit the .pdf of your 2-3 slides with the filename YOURLASTNAME-ProjectIdeas.pdf on Stellar.

3. By 11am Wednesday, September 19th review everyone else's project ideas (to be posted Monday evening), and select two other projects of high interest to you and contact them with a question about their project on Piazza.

HWK1

Upload to Stellar by Tuesday Sep 11 at 11am (Please note Policy on Lateness: -1 point out of 10 total, for each hour late)

If you can't upload on Stellar yet, then email your answers to mas630-staff@media.mit.edu and use the subject: MAS.630 HWK1.

READ:

Affective Computing: Introduction, Chapter1, Chapter2 (especially pp. 75-82 on emotional intelligence), and Chapter3.

WRITE:

1. Just like we constructed a scenario in class (See the class slides on Stellar if you missed this) construct your own scenario, showing a specific example of a human-human interaction that clearly involves emotion for both actors in the scenario. Pick a scenario where the Media Equation holds, e.g. where you take out the word "human", put in the word "computer" and the exactly worded scenario seems reasonable. (a) Write the interaction scenario twice: the first time using the word "person" and the second time replacing on that word with the word "computer" (or the name of the computer agent/program/robot) but otherwise the descriptions should be the same. For example, in class we spelled out each step of the interaction replacing the word "person" with the name "Clippy." (b) What is the emotion, explicitly, that is involved in your interaction for the computer, and for the person? For example, the class interaction involved "frustrated" for the person and "cheerful" for the computer/agent.

2. Repeat #1 for another interaction where the media equation does not seem to hold when you swap the words "computer" and "person." Make sure you follow the other rules as above-- pointing out the affective state for both actors in the scenario.

** Bonus points are given if your case(s) are humorous and elicit laughter from Picard. **

3. (a) Give an example where you have seen a smartphone, smartphone app, or other consumer device demonstrate use of emotional intelligence. If you can't think of one, then pick some kind of technology that you use in everyday life, and describe where you'd like to see it use emotional intelligence. In the latter case, describe what the interaction looks like now, and how it would look differently if it were emotionally intelligent. Show each turn in the interaction or dialogue as if it were in a script of a play or movie. (b) What makes it emotionally intelligent?

4. Pick one of your least favorite applications from Chapter 3 and critique it (pros and cons) based on your own personal and unique research perspective. I wrote these over twenty years ago, and while some things have not changed much, others have changed dramatically. What context/situation would make it most likely to succeed? What would make it most likely to fail? Explain your thinking.

5. Repeat #4 for one of your most favorite applications from Chapter 3.

6. What two affective computing topics do you most want to see us cover during this term?

7. What is the status of your CITI certification? Please make sure you have CITI-certification that is valid through another institution or through MIT for the duration of the course. At MIT you can get it through MIT COUHES Training. Please confirm "YES" that you have completed this training (required for any class project that involves people.) Note, this can take many hours to complete, but it can be done in several sessions.