I plan to maintain an accurate schedule for at least two weeks into the future and to give assignments at least one week before they are due. On the flip side, nothing on this schedule that is further than two weeks in the future can be trusted. This schedule goes from best friend to schoolhouse schemer at the two week mark. Plan accordingly.
We will take a 5-minute break near 12pm on Fridays. If I forget, remind me.
9/4 Wed. 1pm–2pm Lecture: Introduction to the course
9/6 Fri. 11am–1pm Guest lecture by Nick Gillian (MIT): a review of supervised learning with some hands-on interactive ML, including use of his Gesture Recognition Toolkit
Due Wednesday 9/11: One response to two readings, A Few Useful Things to Know about Machine Learning by Domingos and Machine Learning that Matters by Wagstaff; Piazza folder for responses: reading_response_1
9/11 Wed. 1–2 Discussion on assigned reading
9/13 Fri. 11–1 Guest lecture by Rebecca Fiebrink (Princeton): research on Wekinator and interactive tutorial in its use; preparation: have Wekinator installed on your computer or pair up with someone who does (instructions given in email)
Due Tuesday 9/17 by 1pm: Reading response to draft Power to the People: The Role of Humans in Interactive Machine Learning (available on the Course Page / Resources tab on Piazza)
9/18 Wed. 1–2 Toolkit presentation example by Brad Knox; Discussion on assigned reading
9/20 Fri. No class; student holiday
Due Wednesday 9/25 before class: Create an ML system with hard data input (instructions on assignment page)
9/25 Wed. 1–2 First toolkit presentations (Ali)
Due Thursday 9/26 by 1pm: Reading response to Designing Interactions for Robot Active Learners by Cakmak et al.
9/27 Fri. 11–1 Lecture: Research methods for human-computer interaction and possible discussion on assigned reading
Due Wednesday 10/2 before class: Add human input to your simple ML system (instructions on assignment page)
10/2 Wed. 1–2 Guest lecture – Kayur Patel (Google/Columbia) on machine learning in practice and viewing development as an interactive machine learning process
Due Thursday 10/2 at 1pm: Special reading response to Principles of mixed-initiative user interfaces by Horvitz (see email or Piazza announcement for instructions)
10/4 Fri. 11–1 Toolkit presentation (Oren), demonstrations of simple ML system in progress (3-5 minutes each), and discussion on assigned reading.
Due Tuesday 10/8 by 1pm: Reading response to Overview-Based Example Selection in End-User Interactive Concept Learning by Amershi et al. (no discussion point needed in response; may have a slightly different title with “Mixed-Initiative” in some places)
10/9 Wed. 1–2 Lecture: Research methods for machine learning (part 1)
Due Friday 10/11 before class: Add transparency to or feedback from your simple ML system (instructions on assignment page)
10/11 Fri. 11–1 Guest lecture – Saleema Amershi (Microsoft Research) presenting the talk Designing for Effective End-User Interaction with Machine Learning
Due Tuesday 10/8 at 1pm: Reading response to Interacting Meaningfully with Machine Learning Systems: Three Experiments by Stumpf et al.
10/16 Wed. 1–2 Guest lecture – Simone Stumpf (City University London)
10/18 Fri. 11–1 Discuss final projects, Toolkit presentation (Steve), and peer evaluation of students’ simple IML systems
10/23 Wed. 1–2 No class; Media Lab Members Week
Due Friday 10/25 before class: Dissect an ML algorithm for potential forms of interactivity (instructions on assignment page)
10/25 Fri. 11–1 Guest lecture – Henry Lieberman (MIT); Brad Knox will be at a conference.
10/30 Wed. 1–2 Guest lecture – Krzysztof Gajos (Harvard); Brad Knox will be at a conference.
Due Thursday 10/31 at 1pm: Reading response to A survey of robot learning from demonstration by Argall et al. (no discussion point needed)
Due Friday 11/1 before class: Final project proposal (instructions on assignments page)
11/1 Fri. 11–1 Lecture and in-class activity: Learning from demonstration (Sequential decision-making I)
Due Wednesday 11/6 at 4pm: Create and submit at COUHES application for your final project (instructions posted on assignments page)
11/6 Wed. 1–2 Toolkit presentations (Jennifer and Artem)
11/8 Fri. 11–1 Guest lecture – Joe Konstan (Univ. of Minnesota) on HCI for recommender systems
Due Tuesday 11/13 at 1pm: Reading response to Chapter 3 of Sutton and Barto’s Reinforcement Learning: An Introduction. HTML version available here.
11/13 Wed. 1–2 Discussion – Reinforcement learning (Sequential decision-making II)
11/15 Fri. 11–1 “Guest” lecture – Brad Knox (MIT) on learning from human-generated reward
11/20 Wed. 1–2 Guest lecture – Kshitij Judah (Oregon State) on New Modes of Human-Assisted Policy Learning
11/22 Fri. 11–1 Guest lecture – Tessa Lau (Savioke); discuss each final project, getting feedback from other students
11/27 Wed. 1–2 No class meeting; work on your projects
11/29 Fri. 11–1 No class; Thanksgiving holiday
12/4 Wed. 1–2 Toolkit presentations (Been and Greg)
12/6 Fri. 11–1 No class; moved to the following week
12/11 Wed. 1–2 Toolkit presentation (Minshu on Orange)
12/13 Fri. 11-1:30 Final project presentations