Search papers, labs, and topics across Lattice.
This paper explores the intersection of computational semantics and transformer-based language models, elucidating how these models learn and represent linguistic meaning. It critically examines three primary semantic theories鈥攆ormal, grounded, and distributional semantics鈥攁nd highlights the distinctions between human language acquisition and the mechanisms employed by these AI systems. The findings underscore the importance of understanding these semantic frameworks to better leverage and improve AI language models in practical applications.
Transformer models may grasp semantics differently than humans, revealing critical gaps in their understanding of language meaning.
As people adopt transformer-based language models (e.g., ChatGPT and Gemini) for an increasing number of use-cases, it is important to know how such models learn and represent the meaning of the language, and to be more informed about what language is. This document is an attempt to help the reader understand how linguistic meaning (i.e., semantics) is approached from different fields of scientific and philosophical examination. I also explain three primary semantic theories: formal semantics, grounded semantics, and distributional semantics then compare how transformer-based language models differ from how humans learn language.
