PhD Course - Communicating State-of-the-art NLP Research to a Broader Audience
Lecturers:
- Barbara Plank (Professor, ITU)
- Veronika Cheplygina (Associate Professor, ITU)
- Christian Hardmeier (Associate Professor, ITU)
- Carolin Lawrence, PhD (Senior Machine Learning Research Scientist at NEC; invited speaker on scientific communication)
Organizers:
- Barbara Plank (Professor, ITU)
- Veronika Cheplygina (Associate Professor, ITU)
- Christian Hardmeier (Associate Professor, ITU)
- Elisa Bassignana (PhD Student, ITU)
- Max Müller-Eberstein (PhD Student, ITU)
- Mike Zhang (PhD Student, ITU)
Course advertisement:
The annual Conference on Empirical Methods in Natural Language Processing (EMNLP) is a valuable opportunity for researchers and students to learn about and discuss state-of-the-art work in many subdisciplines of NLP. While the virtual format of conferences in the past year has allowed for much broader participation, it also falls short of the scientific interactions which are especially relevant for students.
Through concentrated discussions in smaller, interest-specific reading groups, this course helps students navigate the large number of newly published papers while also encouraging pro-active interactions at the conference itself by jointly discussing and preparing questions beforehand. This will enable attending the conference’s Q&A sessions with the papers’ authors well prepared and with confidence.
Groups will also present their most interesting papers in plenum and can thereby get a broader overview beyond their own subdiscipline and field of research.
A crucial goal of the course is to reflect upon larger trends within NLP research and to communicate these findings concisely to a broader audience. Since this can be a difficult task, the course will feature a half-day workshop on scientific communication after the conference. Based on her extensive experience in NLP outreach, Carolin Lawrence (Senior ML Research Scientist at NEC) will work together with students in order to find the appropriate writing style for the relevant audience.
Each student will prepare a final blog-style report of their favourite papers from the conference with reflections regarding their implications for the field and the general public. These reports should be accessible to and foster scientific communication with a broader audience.
Dates of the course:
- Tue, 2.11.21 Kick-off Meeting
- Tue, 2.11.21 – Thu, 4.11.21 Independent Discussion Meetings in Groups Fri, 5.11.21 Pre-conference Check-in and Presentations
- Tue, 9.11.21 Mid-conference Check-in
- Thu, 25.11.21 Writing Workshop with Carolin Lawrence and Closing Remarks
Time:
Multiple Timeslots (see programme for details)
Room:
Hybrid or Online (Zoom)
Course description:
A student who has met the objectives of the course will be able to:
- Read and discuss with peers EMNLP 2021 papers on natural language processing
- Explain the scientific content of the EMNLP 2021 papers to other researchers and PhD students in the field (in short oral presentations)
- Communicate the scientific contents of multiple EMNLP 2021 papers and overall trends to a general audience
- Compare paper discussions with publicly available peer reviews
- Reflect upon future impacts of EMNLP 2021 papers
- Explain key points of EMNLP 2021 tutorials and workshops to peers
- Attend Q&A sessions with EMNLP 2021 paper authors and discuss relevant issues
The reading list will be based on the proceedings of the EMNLP conference. These are typically published one week before the start of the conference (see also
last year’s papers).
Students will be asked to prepare papers of interest to their track (at least 5) to discuss in the interest groups and in plenum.
The interest groups as well as the topics covered in the papers will include (but not be limited to):
- Computational Social Science and Cultural Analytic
- Dialogue and Interactive Systems
- Discourse and Pragmatics
- Efficient Methods for NLP
- Ethics and NLP
- Generation
- Information Extraction
- Information Retrieval and Text Mining
- Interpretability and Analysis of Models for NLP
- Linguistic Theories, Cognitive Modeling and Psycholinguistics
- Machine Learning for NLP
- Machine Translation and Multilinguality
- NLP Applications
- Phonology, Morphology and Word Segmentation
- Question Answering
- Resources and Evaluation
- Semantics: Lexical, Sentence level, Textual Inference and Other areas
- Sentiment Analysis, Stylistic Analysis, and Argument Mining
- Speech, Vision, Robotics, Multimodal Grounding
- Summarization
- Syntax: Tagging, Chunking and Parsing
See EMNLP 2021’s call for papers for details.
Programme (preliminary) Tue, 2.11.21 at 9:00 – 11:00 — Kick-off Session (overview, group formation)
- Organizers provide an overview of the course
- Groups of ~4 are formed based on fields of interest
Tue, 2.11.21 – Thu, 4.11.21 (~4 hours) — Independent Group Discussions
- Groups should meet independently to gather and discuss relevant papers from the upcoming conference
- Until Friday, each group should have a list of relevant papers as well as questions for the authors to ask during EMNLP sessions
Fri, 5.11.21 at 10:00 – 12:00 — Pre-conference Check-in
- The full course meets in order to get an overview of each field of interest
- Each group presents their list of relevant papers
Tue, 9.11.21 at 14:00 – 16:00 (CEST), 8:00 – 10:00 (AST) — Intermediate Report
- During the conference, there will be an opportunity for students to exchange their experience so far
Thu, 25.11.21 at 9:00 – 15:00 — Writing Workshop by Carolin Lawrence and Closing Remarks
- In preparation for writing their blog post report, students are asked to select their favourite EMNLP papers and prepare a rough draft
- During the lectures and exercises of the workshop, students will learn to communicate in-depth NLP knowledge to audiences ranging from experts to the general public
- Final remarks regarding the course and the possibility for students to voice their feedback
Prerequisites:
- Participants must be registered to attend the 2021 EMNLP conference
- A strong background in NLP is recommended
Exam:
- Students are evaluated based on individual reports (blog post).
- Grading: pass / not pass, internal examiner
Credits:
3 ECTS (see next section for details)
Amount of hours the student is expected to use on the course:
Participation: 56 hours
- 3x 2 hrs for 3 meetings
- 2x 2hrs for independent group discussions
- 5x 8 hrs for conference attendance (see appendix and last year’s programme)
- 6 hrs writing workshop
Preparation: 32 hours
- 24 hrs for paper preparation before the course 8 hrs for blog post writing
Total: 88 hours
Participants:
- Maximum Participants: 30
- Expected Participants: 20
How to sign up:
Students can sign up via this form.