Lab News

• 10/3: 5 papers (3 long, 2 short) accepted to NAACL 2021
  
  • A Non-Linear Structural Probe
    
  • What About the Precedent: An Information-Theoretic Analysis of Common Law
    
  • Do Syntactic Probes Probe Syntax? Experiments with Jabberwocky Probing
    
  • Finding Concept-specific Biases in Form–Meaning Associations
    
  • How (Non-)Optimal is the Lexicon?
    

• 25/2: Ryan is giving an invited talk at the ELLIS NLP Workshop on joint work with Clara and Tim

• 28/1: 3 papers accepted to EACL 2021

• 15/12: Ryan is giving an invited talk at the University of Geneva.

• 11/11: Ryan is giving an invited talk at Moscow State University.

• 30/9: Paula Czarnowska’s paper Morphologically Aware Word-Level Translation was accepted to COLING 2020. Joint work with Ryan, Sebastian Ruder, and Ann Copestake. We are super excited for the release of Paula’s follow-up to her well received EMNLP 2019 paper Don’t Forget the Long Tail! A Comprehensive Analysis of Morphological Generalization in Bilingual Lexicon Induction.

• 25/9: Through ETH Zurich, we are officially part of the ELLIS PhD program, which supports PhDs and Post-Docs by providing access to leading researchers through boot camps, summer schools and workshops of the ELLIS programs.

• 17/9: Ryan is publication chair of NAACL 2021.

• 15/9: 8 papers (7 long, 1 short) accepted to EMNLP 2020

  • If Beam Search is the Answer, What was the Question?
    
  • Pareto Probing: Trading Off Accuracy and Simplicity
    
  • Intrinsic Probing through Dimension Selection
    
  • Investigating Cross-Linguistic Adjective Ordering Tendencies with a Latent-Variable Model
    
  • Speakers Fill Semantic Gaps with Context
    
  • Please Mind the Root: Decoding Arborescences for Dependency Parsing
    
  • Exploring the Linear Subspace Hypothesis in Gender Bias Mitigation
    
  • Measuring the Similarity of Grammatical Gender Systems by Comparing Partitions
    
    

• 14/9: Parameter Space Factorization for Zero-Shot Learning across Tasks and Languages to appear in TACL 2020.

• 9/2: Eleftheria Tsipidi will start her PhD at ETH Zürich. She will be co-advised by Severin Klinger and Ryan and will be associated with ETH Zurich’s Media Technology Center.

• 29/8: Efficient Computation of Expectations under Spanning Tree Distributions to appear in TACL 2020.

• 7/7: Best-First Beam Search to appear in TACL 2020.

Thesis Projects

If you are a BSc or MSc student at ETHZ interested in writing your thesis with us, we would be delighted to hear from you! We accept students who are seeking an independent research project as well. Our research broadly revolves around theoretical and applied problems in Natural Language Processing, Computational Linguistics, Machine Learning and Statistics. To obtain a better understanding of what currently interests us, we invite you to check our recent publications. However, feel free to express interest in any topic you think our group might be well suited to advise you on: Just because we have not yet looked into a topic does not mean we are not interested in it or willing to become interested in the topic.
 
Please send an email to ryan.cotterell@inf.ethz.ch with CC to clara.meister@inf.ethz.ch and niklas.stoehr@inf.ethz.ch and state either [bachelor’s thesis] or [master’s thesis] at the start of the subject. For us to get to know you a little, please write a paragraph introducing yourself and why you are interested in working with us. It would help us a lot if you also provided a list of four or five more concrete topics that you are interested in. We will try our best to find a project that suits your interests. We are looking forward to receiving your inquiry!

Project Announcements

Projects on “NLP against Armed Conflict” – Niklas Stoehr, Ryan Cotterell

ETHZ++ NLP Reading Group

SEE FULL READING LIST HISTORY

Our lab, along with Mrinmaya Sachan’s lab, meets weekly on Fridays at 15:00 CEST. Here’s what we’re reading!

04/12/19: Semantic categories of artifacts and animals reflect efficient coding
11/12/19: SpanBERT: Improving Pre-training by Representing and Predicting Spans
15/01/20: Gender-Aware Reinflection using Linguistically Enhanced Neural Models

Joining Our Lab

Thank you very much for your interest in joining our group – we would be delighted to hear from you!

If you are interested in working with us as a Master’s student, please see here. Ryan has previously co-advised Master’s students on NLP topics with Mrinmaya Sachan and others, if co-advising is an option you would like to pursue. At Cambridge, Ryan co-advises MPhil students with Simone Teufel. We are looking forward to receiving your inquiry!

If you are interested in joining us as a PhD student: we do get a lot of emails about joining our lab from super interesting, well-qualified applicants who we would be very happy to have as research colleagues. However, there has been unbridled exuberance over joining our group and we grew quite quickly. Thus, for the foreseeable future, we can only accept new PhD students under one of the following circumstances:

• Consider that ETHZ is part of the ELLIS PhD program, which supports PhDs and postdoctoral fellows by providing access to leading researchers through boot camps, summer schools and workshops of the ELLIS programs. You can apply for admission through the centralized application portal; the deadline is the 1st of December 2020. Our lab works closely with CopeNLU headed by Isabelle Augenstein at the University of Copenhagen; applicants who apply through ELLIS are encouraged to tag both Ryan and Isabelle as they are both part of the ELLIS network

• Ryan is also core faculty member of the newly-established ETH AI Center which offers fully-funded PhD and Post-doctoral fellowships. Please apply here and describe how you intend to take an interdisciplinary perspective on NLP.

• If you are interested in NLP at ETHZ and do not yet have a Master’s degree, please consider applying to ETHZ’s direct doctorate program. You should tag Mrinmaya Sachan and Ryan and that you are interested in NLP.

• If you have a previous working relationship with members of our group or there is a co-advisor who has a previous working relationship with our lab.

Note on our research focus and collaboration style: we don’t do leaderboard science or chase state-of-the-art numbers on ephemeral datasets. If you’re okay with that, awesome! You’ll probably fit right in. I spend a lot of time with those I collaborate with. I am a hands-on advisor who writes large swathes of our papers. I am also known to have a strong opinion on modeling aesthetics, code style and experimental design. That may not be what you are looking for during your PhD, which is perfectly fine. This style of advising takes up a lot of my time, which is already stretched quite thin as it is, and, hence, the need to limit admissions. In many ways, I am more of a collaborator than a traditional advisor.