Exploring the intersection of formal linguistics, machine learning, and cognitive science to build language systems that truly understand meaning.
View Publications →
Developing neural architectures that map natural language to formal meaning representations, enabling machines to reason about language with logical precision.
Modeling how humans infer meaning beyond literal words — implicature, presupposition, and discourse context — using probabilistic frameworks.
Building language-agnostic representations that transfer linguistic knowledge across 100+ languages, with focus on low-resource languages.
Studying failure modes of large language models — hallucination, bias propagation, and alignment — with formal verification approaches.
Graduate seminar on neural approaches to language understanding, generation, and reasoning.
Fall 2024Formal methods in natural language processing: grammars, parsing, semantics, and pragmatics.
Spring 2024New course on safety, alignment, and interpretability in large language models.
Spring 2025Interested in joining the lab, collaborating on research, or inviting me to speak? I'd love to hear from you.
slin@mit.edu →