stress-prediction

This paper evaluates on-device language models for privacy-preserving stress prediction using multimodal data on mobile devices, finding that lightweight models achieve low latency and predictable resource usage while highlighting practical constraints for mobile mental health applications.

This paper investigates weakly-supervised ByT5 fine-tuning for stress-aware sentence-level Filipino grapheme-to-phoneme conversion, achieving significant improvements in phoneme and character error rates on a manually-corrected test set.