EchoVoices combines k-NN-enhanced Whisper ASR, two-stage VITS TTS, and an LLM/RAG persona agent to create digital personas for seniors and children, reporting CER and similarity gains on two Mandarin datasets.
Generalization through memorization: Nearest neighbor language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
EchoVoices: Preserving Generational Voices and Memories for Seniors and Children
EchoVoices combines k-NN-enhanced Whisper ASR, two-stage VITS TTS, and an LLM/RAG persona agent to create digital personas for seniors and children, reporting CER and similarity gains on two Mandarin datasets.