Lab rotations are a regular feature of work in my research groups. Students join the lab and figure out a project they want to work on themselves. While this typically results in at least a serviceable final report and good learning outcomes for the student, occassionally the work done lends itself to wider dissemination. If a student feels motivated and we can support them, we may try to move towards publication. In the past few years we’ve produced several such student-led papers. What we try to do in such cases is have them take the lead on a publication — often their first first-author publication.
Sometimes things can move fast. For instance, when Jacqueline van Arkel neared the end of her internship, co-supervisor Marieke Woensdregt eyed a highly relevant conference and the work was published at CoNLL within a few months. Same thing with Ada Lopez’ 2022 work on ASR evaluation; around the conclusion of the lab rotation, we spotted a deadline and worked to a tight schedule to turn the findings into a first paper at KONVENS, Germany’s premier computational linguistics conference.
Sometimes things take longer. Our most recent paper, The Anatomy of Iconicity, is an example. Led by Stella Punselie, it builds on an idea I had first implemented in 2017 and adds data from Stella’s lab rotation and excellent BA thesis, both done in 2020. Pandemic curfews and other projects intervened as Stella went to work in industry. We enlisted the help of Bonnie McLean in Uppsala who works on similar topics for her PhD. Recently we found some time to finish the paper (see blog here) and we’re very happy with how it’s turned out.
Probably the most extreme example is a paper with Hannah Lutzenberger, Rehana Omardeen and Lierin DeWael that also came out this year. Hannah and I first discussed this in 2016 as a possible spin-off of her MA thesis. That is a good eight years ago! A full PhD trajectory intervened, which took things in a different direction for Hannah. Nevertheless, we never forgot about the project, and over the years Hannah managed to add some excellent collaborators and a novel dataset. In those years, between the four of us we changed countries and affiliations at least 8 times. The end result was worth the wait: a first-of-its-kind direct comparison of interactional structure in British Sign Language and Spoken British English. The paper is now out in Sign Language Studies.
Student projects open up opportunities for serendipity
I strongly believe that serendity is one of the most valuable things in science: the je ne sais quoi without which interesting things rarely happen. I’ve learned that we can never plan for serendipity —that is pretty much the point— but that we can create opportunities for it to arise. And student projects are one way to do that.
With undergraduate student projects you never quite know what direction things will take. Jacqueline van Arkel’s work gave us an opportunity to connect Marieke Woensdregt’s computational modelling work in our Elementary Particles of Conversation project to work on computational complexity with Mark Blokpoel from the Computational Cognitive Science group. It was not something we’d planned in advance, and it made for a very fruitful and fun collaboration.
When Ada Lopez joined our research group, we were not doing much work on speech recognition, but her interest and input led to several papers (most recently a SIGDial 2023 paper that made it to one of the two best paper nominations). For the paper led by Hannah Lutzenberger, it may have taken us nearly 8 years to move from first idea to publication, but the intervening meanderings meant that the final product was much the richer for it, and could be informed by some hypotheses and computational modelling work that had been realized in the meantime.
Here are four of the student-led papers we’ve realized in the past years:
- Arkel, Jacqueline van, Marieke Woensdregt, Mark Dingemanse, and Mark Blokpoel. 2020. ‘A Simple Repair Mechanism Can Alleviate Computational Demands of Pragmatic Reasoning: Simulations and Complexity Analysis’. In Proceedings of the 24th Conference on Computational Natural Language Learning, 177–94. Online: Association for Computational Linguistics. doi:10.18653/v1/2020.conll-1.14.
- Lopez, Alianda, Andreas Liesenfeld, and Mark Dingemanse. 2022. ‘Evaluation of Automatic Speech Recognition for Conversational Speech in Dutch, English and German: What Goes Missing?’ In Proceedings of the 18th Conference on Natural Language Processing (KONVENS 2022), 135–43. Potsdam. online
- Lutzenberger, Hannah, Lierin de Wael, Rehana Omardeen, and Mark Dingemanse. 2024. ‘Interactional Infrastructure across Modalities: A Comparison of Repair Initiators and Continuers in British Sign Language and British English’. Sign Language Studies 24 (3). Gallaudet University Press: 548–81. doi:10.1353/sls.2024.a928056.
- Punselie, Stella, Bonnie McLean, and Mark Dingemanse. 2024. ‘The Anatomy of Iconicity: Cumulative Structural Analogies Underlie Objective and Subjective Measures of Iconicity’. Open Mind 8 (September): 1191–1212. doi:10.1162/opmi_a_00162
Thanks for reading!