NLP Colloqium Talk with Nava Haghighi: Ontologies in Design: How Imagining a Tree Reveals Possibilities and Assumptions in Large Language Models

As part of the NLP Colloquium series hosted by Lamarr NLP Area Chair and Principal Investigator Prof. Dr. Lucie Flek, we are pleased to welcome Nava Haghighi from Stanford University for a guest talk titled “Ontologies in Design: How Imagining a Tree Reveals Possibilities and Assumptions in Large Language Models”. The talk will take place on August 6, 2025, from 10:00 to 11:00 CEST.

What to expect
While value-based analysis is important, this guest talk will highlight that ontologies are just as relevant, however, often under-recognized. Therefore, a practice-based engagement with ontologies is being proposed. Thereby pluralism, groundedness, liveliness, and enactment are the four orientations for considering ontologies in design.

Why attend?
This Talk will showcase examples of potentialities that are opened up through these orientations across the entire LLM development pipeline by conducting two ontological analyses. The first one consisting of examining the responses of four LLM-based chatbots in a promptingexercise, and the second one of analyzing the architecture of an LLM-based agent simulation. Lastly, the talk will bring forth various opportunities and limitations which are involved when working with ontologies in the design and development of sociotechnical systems.

About the Speaker
Nava Haghighi is PhD candidate in computer science at Stanford University, advised by Professor James Landay. Her research examines the ontological assumptions—the boundaries of what we allow ourselves to imagine—embedded in the design of sociotechnical systems. Nava holds a bachelor of architecture from California Polytechnic State University, San Luis Obispo, and a dual master of science in computer science and integrated design and management from MIT.

If you’d like to join, please subscribe to our mailing list here.
You’ll receive all details about the talk, including the Zoom link, via email.