SeniorTalk: Mandarin Chinese Speech Data for Seniors Aged 75-85

SeniorTalk provides 10,000 to 100,000 Mandarin Chinese speech records from individuals aged 75 to 85, produced by BAAI in 2025. It includes audio and text modalities to facilitate research in automatic speech recognition and speaker verification for the super-aged population.

Use Cases

• Automatic Speech Recognition (ASR) using audio segments and text transcriptions
• Speaker Verification (SV) for elderly voice biometric research
• Speaker Diarization (SD) to identify individual speakers in Mandarin conversations

Strengths

• Targeted demographic of 75-85 year olds
• 10,000 to 100,000 records
• Dual audio and text modalities

Limitations

• Narrow age demographic (75-85) limits use for general-purpose models
• CC BY-NC-SA 4.0 license prohibits commercial use

Provenance

Source

BAAI (Beijing Academy of Artificial Intelligence)

Freshness

Last updated January 2026

Geography

China