Lingthusiasm - A podcast that's enthusiastic about linguistics

98: Helping computers decode sentences - Interview with Emily M. Bender

Nov 22, 2024
In this engaging conversation, Dr. Emily M. Bender, a Linguistics Professor at the University of Washington and co-host of Mystery AI Hype Theater 3000, explores how computers decode language. She reveals the difference between human and computer language learning, discussing the humorous failures that can arise. Emily emphasizes the significance of community over mere data, the ethical implications of AI training, and the creation of the Bender Rule, which promotes clarity in language use. Expect intriguing insights into the quirks of AI behavior and the future of linguistic technology!
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
ANECDOTE

From Startup Job To Grammar Matrix

  • Emily Bender described building a Japanese grammar at a startup to automate customer service responses.
  • That work led her to create the Grammar Matrix starter kit to help build grammars for many languages.
INSIGHT

Language Is Many Separate Problems

  • Solving “language” is vague because language involves many distinct skills like sound, syntax, semantics, and pragmatics.
  • True human-like understanding requires far more than linguistics alone, including multimodal and pragmatic reasoning.
ADVICE

Work With Communities First

  • Engage with communities first and ask what computational tools they want before building systems for their languages.
  • Train local community members as computational linguists rather than extracting their data without consent.
Get the Snipd Podcast app to discover more snips from this episode
Get the app