Sondre Wold

Sondre Wold is a Doctoral Research Fellow and PhD student at the University of Oslo's Department of Informatics, affiliated with the Language Technology Group (LTG) under the Faculty of Mathematics and Natural Sciences. His research focuses on compositional generalization in language models, particularly exploring how sequence models generalize to novel compositional structures using knowledge graphs and synthetic languages derived from context-free grammars (CFG).

He teaches courses IN1140 (Introduction to Language Technology) and IN5550 (Neural Methods in Natural Language Processing), covering topics like N-gram models, semantics, syntax, encoder-decoders, attention mechanisms, and graph neural networks. His work bridges theoretical NLP concepts with practical applications in model interpretability and retrieval systems.

Wold's recent publications (2023-2024) address compositional generalization, lexical complexity estimation, and Norwegian language resources like NorQuAD and BRENT. He collaborates with researchers such as Lucas Charpentier and Erik Velldal, advancing methods for knowledge graph alignment and retrieval-enhanced language models.