This week on the NLP deep dive we'll be analyzing some survey data together! I've been thinking a lot lately about how people's beliefs about whether large language models (like BERT or GPT-3, or even as part of multi-modal systems like Stable Diffusion) have some meaningful type of language understanding pattern with their beliefs about computers in general and other types of language technology. Basically: if you are inclined to talk or behave as if computers have (free) will, are you more likely to assign them other internal states as well? (Like language understanding or creation of a mental model.)
(Apologies about the very slightly misleading title: I'm interested in learning more about people's beliefs here, not debating the qualities of the systems themselves. I also think more people don't think computers are people, but there are pretty strict character limits for YouTube video titles.)
Link to code (very much not polished): https://github.com/rctatman/Computer-...
Citations:
Scholl, B. J., & Tremoulet, P. D. (2000). Perceptual causality and animacy. Trends in cognitive sciences, 4(8), 299-309. https://perception.yale.edu/papers/00...
Mikropoulos, T. A., Misailidi, P., & Bonoti, F. (2003). Attributing human properties to computer artifacts: Developmental changes in children's understanding of the animate-inanimate distinction. https://www.researchgate.net/publicat...
Tong, Yu, Fuxing Wang, and Weijun Wang. "Fairies in the Box: Children’s Perception and Interaction towards Voice Assistants." Human Behavior and Emerging Technologies 2022 (2022). https://downloads.hindawi.com/journal...
----
Welcome! My name is Dr. Rachael Tatman and I'm a language technology educator. I have a doctorate in linguistics (my research was in computational sociolinguistics) and I've been a researcher, data analyst and developer advocate. My channel is for anyone who cares about language technology and other people.
If you'd like to support my work, the best way currently is through my ko-fi: https://ko-fi.com/rctatman/tiers
You can keep up with me here:
My Twitter: / rctatman
My Mastodon: https://mastodon.rctatman.com/@rctatman
My GitHub: https://github.com/rctatman
My website: http://www.rctatman.com/
#languagetechnology #naturallanguageprocessing #speechtechnology
Watch video NLP Deep Dive, Livecoding: If LLMs understand language, are computers people? online without registration, duration hours minute second in high quality. This video was added by user Rachael Tatman 01 January 1970, don't forget to share it with your friends and acquaintances, it has been viewed on our site 29 once and liked it 1 people.