Our work on transforming non-speech captions with anchored generative models is headed to ASSETS 2025 in Denver!