CSE 290: ML Methods for NLP

CSE 290 : Machine Learning Methods for NLP

Term: Winter Qtr 2019
Instructor: Ndapa Nakashole, CSE 4108
Lecture: Wednesdays 3:30pm-4:50pm, CSE 4258
Credits: 1

Announcements

Jan 2: Enrollement questionnaire posted.

Course Description

The course involves reading and discussing current research papers, once a week. This course assumes background in basic machine learning. Prior NLP experience is helpful, but not required.

Students will be assigned to papers by the organizer. Each student will present one paper and lead a discussion about the paper.
Papers will be assigned with enough time in advance.
There will be 9 seminars in total this quarter.

Theme

The theme for this quarter is generalization.
Papers will fall under topics such as:

Transfer learning
Domain adaptation
Multi-task learning
Meta-learning

Enrollment

Class size will be capped due to the format of the class.
Enrollment decisions will be made by the instructor by the end of first week of classes.
Please respond to this questionnaire. The last day to fill it out is January 9. First discussion is on Jan 16.

Grading

This is a 1-credit pass-fail seminar. Grades will be based on your paper presentation and class participation

Syllabus

Date	Topic/Readings	Discussion Leader
Jan 9	Paper 0
Jan 9	Yosinski, NIPS 2014 How transferable are features in deep neural networks?	All
Jan 16	Paper 1
Jan 16	S. Thrun, NIPS 1995 Is learning the n-th thing any easier than learning the first?	Ndapa Nakashole
Jan 23	Paper 2
Jan 23	Finn et al. ICML 2017 Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks	Bodhisattwa Majumder
Jan 30	Paper 3
Jan 30	Ganin et al. JMLR 2016 Domain-Adversarial Training of Neural Networks	Khalil Mrini & Yutong Shao
Feb 6	Paper 4
Feb 6	Luo et al. NIPS 2017 Label Efficient Learning of Transferable Representations across Domains and Tasks	Aishma Raghu
Feb 13	Paper 5
Feb 13	Zhang et al. TACL 2017 Aspect-augmented Adversarial Networks for Domain Adaptation	Yuguang Lin
Feb 20	Paper 6
Feb 20	Collobert et al. JMLR 2011 Natural Language Processing (Almost) from Scratch	Aman, Achpal
Feb 27	Paper 7
Feb 27	Daume ACL 2007 Frustratingly Easy Domain Adaptation	Aashi Jain
Mar 06	Paper 8
Mar 06	Pan et al. TKDE 2010 Survey on Transfer Learning	Yifan Zhou
Mar 12	Paper 9
Mar 12	Stewart & Ermon. AAAI 2017 Label-Free Supervision of Neural Networks with Physics and Domain Knowledge	Hao Liu & Jingwu Xu