Steam-powered Turing Machine University of Washington Department of Computer Science & Engineering
CSE 481 N - Natural Language Processing Capstone - Spring 2018
Lecture: TTh 10:00 - 11:20am in OUG 136
CSE Home

Teaching Crew

Personnel Contact Office Hours
Instructor: Yejin Choi yejin at cs dot washington dot edu Wed 4:30 - 5:30pm @ CSE 578
TA: Ari Holtzman ahai at cs dot washington dot edu
Mon 10 - 11am @ CSE 203
TA: Nelson Liu nfliu at cs dot washington dot edu
Fri 4-5pm @ CSE 220

Class Objectives

This class will provide students with an intensive 10-week experience in successfully completing a challenging, but well-scoped research project.

Participants will work in small groups (2-3 people in each group) to hone their technical skills to quickly absorb and adapt new technical knowledge, gain experience in complex programming, perform thorough experiments and analysis, and learn how to find a path when faced with negative results.

Additional objectives of this class include:
(1) technical communication skills to produce high quality interim technical reports that inspire insightful discussion across project groups,
(2) advisory project experience to provide technical advice and constructive feedback on others, and
(3) project management skills to prioritize work items to maximize the chance for successful outcome.

Mini Lectures, Project Discussion & Presentation Schedule

Week Dates Topic Supplementary Material
1 Mar 27, 29 Course overview & project pitch from xlab (Tue)
AI2's guest lecture on Semantic Scholar by Waleed Ammar and Islam Beltagy (Thu)
2 Apr 3, 5 AI2's guest lecture on AllenNLP by Mark Neumann and Joel Grus (Tue)
Yejin's research overview talk (Thu)
3 Apr 10, 12 Project proposal presentations
4 Apr 17, 19 Guest lecture (Tue)
Paper discussion I (Thu)
5 Apr 24, 26 Project update presentations
6 May 1, 3 Project update presentations (Tue)
Paper discussion II (Tue, Thu)
7 May 8, 10 Project update presentations and demo
8 May 15, 17 Project update presentations and demo (Tue)
Paper discussion III (Tue, Thu)
9 May 22, 24 Project update presentations and demo
10 May 29, 31 Project update presentations and demo (Tue)
Paper discussion IV (Tue, Thu)
11 TBD Finale! - final poster presentation & demo @ CSE atrium TBD

Milestones, Artifacts & Objectives

Each week, you will make a blog post that describes exciting progress (new trials, results, failure, insights) your team has made. You will make a submission to Canvas the URL to your new blog post each week. Keep in mind the overall milestones when planning out your project activities.
Week Due Milestones Artifacts Objectives
2 Apr 3 Tue First blog post! blog post #1 - Form a project group, make a team name. (the cooler the better!)
- List top 3 project ideas your team is the most excited about, briefly outline the minimal viable action plan with stretch goals.
- Start a github project and share the URL in your blog post.
- Decide which project mode you want to be in: start-up mode (heavier focus on the coolness of the project and system development) or research mode (heavier focus on novel models, analysis, and insights).
Apr 5 Thu Warm-up blog post #2 If you're (considering to be) on a deep learning track, install pytorch or tensorflow, get familiarized with pytorch/tensorflow APIs, download and run any existing codebase that can support your potential project, and tell us about your experience. If you don't want deep learning, that's totally cool too! Run any software that you either downloaded or wrote yourself that can be potentially useful for your project and tell us what you've been up to.
3 Apr 10 Tue Project Proposal blog post #3 Time to make a formal proposal! A good proposal should sketch out both the minimal viable action plan as well as stretch goals. Clearly state your project objectives, proposed methodologies, available resources, and the evaluation plan. Don't forget to include literature survey!
4 Apr 17 Tue Strawman I blog post #4 Complete at least one strawman / baseline approach, run experiments, and set up the evaluation framework.
5 Apr 24 Tue Strawman II blog post #5 Complete *multiple* strawman / baseline approaches, record their performance, plus perform *error analysis*.
6 May 1 Tue Advanced model attempt #1 blog post #6 Time to make an advanced model attempt #1! Tell us what you tried, share any exciting results. If no good news, at a minimum tell us how you'd characterize the failure modes and what your team will investigate next.
7 May 8 Tue Advanced model attempt #1 blog post #7 Continue making the advanced model attempt #1, run more experiments, do more error analysis, and sketch out the next action plan.
8 May 15 Tue Advanced model attempt #2 blog post #8 Let's make an advanced model attempt #2! Tell us what you tried, how it went, and why it didn't work if it didn't work.
9 May 22 Tue Advanced model attempt #2 blog post #9 Continue making an advanced model attempt #2, run more experiments, do more error analysis, and sketch out the final action plan.
11 TBD Final Poster Presentation & Demo Poster Final poster presentation & demo @ CSE Atrium, TBD.
11 Jun 7 Thu 5pm Final Project Report 10 page report in pdf

Class Activities & Grading

Class activities consist of the following components (% for final grade):

Project (75%) -- weekly blog posts (40%)
-- in-class presentations, github activities (10%)
-- final poster presentation, demo, and report (25%)
Paper Discussion (25%) -- Presentation (15%) -- Participation (10%)

Discussion Board

Available at Canvas

Optional Textbooks