Marius Mosbach

I am a postdoctoral fellow at McGill University and a postdoctoral researcher at Mila - Quebec AI Institute in Montréal :canada:.

Before that, I was a postdoctoral researcher at the Language Science and Technology Department of Saarland University :de:, where I also did my PhD in Computer Science. My PhD advisor was Dietrich Klakow.

The overarching goal of my research is to build natural language processing (NLP) systems that are well understood, robust, and adaptable; we should have a rigorous understanding of when and why NLP systems work or fail, they should be robust in a variety of real-world scenarios with guarantees about their performance and generalization, and they should be easy to adapt to new domains and problems.

For our work, my collaborators and I have received a Best Paper Award at COLING 2022, the Best Theme Paper Award at ACL 2023, and the Most Interesting Paper Award at the BabyLM Challenge 2023 :trophy:.

In my free time, I enjoy CrossFit :weight_lifting_man:, playing soccer :soccer:, and baking :cake:.

Feel free to get in touch:

news

Oct 20, 2024 Our paper From Insights to Actions: The Impact of Interpretability and Analysis Research on NLP was accepted to EMNLP 2024. See you in Miami. :us: :beach_umbrella:
Oct 01, 2024 I will attend COLM 2024 to present our paper LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders. See you in Philly :us: :boxing_glove:
May 22, 2024 Our papers on the The Hidden Space of Transformer Language Adapters and The Impact of Demonstrations on Multilingual In-Context Learning: A Multidimensional Analysis got accepted to ACL 2024. :boom:
Apr 23, 2024 I will give an invited talk at the Workshop on Insights from Negative Results in NLP @ NAACL 2024 about our new paper analyzing the impact of interpretability and analysis research. See you in Mexico. :mexico:
Feb 19, 2024 I’m excited to announce that I will join Siva Reddy’s group at MILA & McGill University for a postdoc! :canada:
Jan 19, 2024 I successfully defended my PhD thesis on Analyzing Pre-trained and Fine-tuned Language Models. :confetti_ball: :mortar_board: