| Instructor: | Tatjana Scheffler |
| Time: | Mondays, 4-6pm |
| Place: | Golm, building 14, room 032 |
| Moodle: | Please register on the course's Moodle site. |
Projects on processing natural language questions. Possible areas:
We will meet weekly in the first half of the course, then carry out the projects and meet in smaller groups or only as needed.
| Date | Topic | Readings | Dataset |
|---|---|---|---|
| M 10/22 | Introduction | -- | |
| M 10/29 | Question detection | Faruqui & Das, 2018, Identifying well-formed natural language questions,
Proc. of EMNLP (Huddleston, 1994) (Huddleston & Pullum, 2002) |
q wellformedness corpus |
| M 11/5 | Question types | (Huddleston & Pullum, 2002) Li et al., 2011, Question Identification on Twitter Ranganath et al., 2016, Identifying rhetorical questions on social media |
SWDA |
| M 11/12 | Question typology | Asking too much (Anja) | convotoolkit |
| M 11/19 | Indirect answers | Marneffe et al. (Tunc) Stevens et al. |
iqap |
| M 11/26 | Question generation | Heilman (rule based) Du et al. 2017 (Neural QG) (Yulia K.) |
LearningQ (Sonu) |
| M 12/3 | Question answering | Petrochuk/Zettlemoyer (Jacob) | SQuAD (Sören) |
| M 12/10 | Questions in dialog (chatbots) | Hawkins/Goodman Ono et al. (Alejandra) |
Rasa dialog engine QuAC (Ali) |
| M 12/17 | Project brainstorming | all | |
| 12/24-30 | no class (winter break) | ||
| 12/31-1/6 | no class (winter break) | ||
| M 1/7 | (project work, meetings as needed) | ||
| M 1/14 | (project work, meetings as needed) | ||
| M 1/21 | (project work, meetings as needed) | ||
| M 1/28 | Project presentations | ||
| M 2/4 | (writing time) |
dialog act datasets (selection):