معرفی
Sebastian Schuster is a Lecturer in Computational Linguistics at University College London (UCL), effective January 2024. He will transition to the University of Vienna in mid-2025 to lead a WWTF-funded research group. Previously, he held postdoctoral positions at Saarland University and NYU (via the 2020 Computing Innovation Fellowship), and completed his PhD in Linguistics at Stanford University, affiliated with the ALPS Lab and Stanford NLP Group.
His research focuses on evaluating large language models (LLMs), computational semantics, and pragmatics. Key themes include improving NLU systems through entity tracking analysis, understanding scope ambiguities, and modeling pragmatic inferences. Recent work explores how pretraining affects LLM capabilities and how context influences scalar implicature processing.
Notable contributions include datasets like SIGA for scalar implicatures and SpreadNaLa for code generation evaluation. He has presented at ETH Zurich, Oxford, and served as Area Chair for *SEM2023's psycholinguistics track. His fellowship and invited talks reflect recognition of his interdisciplinary work bridging computational and experimental linguistics.
Professional activities include contributions to Universal Dependencies projects, software tools like Open Linguistics, and collaborations on child language development. His work integrates methodologies from NLP, experimental psychology, and cognitive science to advance both theoretical and applied linguistic research.
