Cameron Jones is a PhD Student in Cognitive Science at UC San Diego. His work focuses on how human beings use their embodied experience with the world to understand language and whether artificial models could understand language in a human-like way. He runs experiments that compare the way in which humans and language models respond to sentences that are theorised to require embodied experience to understand: for example, sentences about physical events, affordances, or social interactions. To the extent that human responses can't be predicted by text-only language models, these experiments suggests that human comprehenders are drawing on experience that goes beyond language alone.
Apr 10, 2023 16:15 pm UTC| Technology
When we asked GPT-3, an extremely powerful and popular artificial intelligence language system, whether youd be more likely to use a paper map or a stone to fan life into coals for a barbecue, it preferred the stone. To...