Matthias Lindemann defends his PhD

May 2025

Matthias Lindemann successfully defended his PhD. Congratulations, Dr. Lindemann!

Matthias worked on methods for improving the inductive bias of neural seq2seq models in order to promote the correct generalizations. His work was motivated by the task of semantic parsing, but some of his results included finite-state and syntax-based generalization. His paper on Compositional Generalization without Trees using Multiset Tagging and Latent Permutations won an Outstanding Paper Award at ACL 2023.

Matthias was co-supervised by Ivan Titov at the University of Edinburgh (where he obtained his PhD) and Alexander Koller. He will join André Martins’ lab in Lisbon as a postdoc.

More information is available on his homepage. We wish Matthias all the best for his further career!