WildElder
收藏资源简介:
WildElder 是一个从网络视频中收集的中文老年人口语语料库,通过精细的人工标注丰富了数据集。该数据集结合了野外数据的多样性和现实性,以及专家手动标注和严格的质量检查的可靠性。每个语句都被标注了正字法转录、说话人年龄、性别和口音强度。数据集涵盖了广泛的通用主题,包括家庭和日常生活、自然和环境、科学和教育、历史和文化以及健康和疾病。WildElder 为自动语音识别和说话人分析的研究提供了强大的支持,并展示了老年人口语处理在现实世界中的挑战和潜力。
WildElder is a colloquial speech corpus of elderly Chinese speakers collected from online videos, enriched by meticulous manual annotations. This dataset integrates the diversity and realism of in-the-wild data with the reliability brought by expert manual annotations and strict quality checks. Each utterance in the corpus is annotated with orthographic transcription, the speaker’s age, gender, and accent intensity. It covers a wide range of general topics, including family and daily life, nature and the environment, science and education, history and culture, as well as health and diseases. WildElder provides robust support for research on automatic speech recognition and speaker analysis, and showcases the real-world challenges and potential of spoken language processing for elderly populations.
- 1SafeToolBench: Pioneering a Prospective Benchmark to Evaluating Tool Utilization Safety in LLMs北京理工大学,香港中文大学,北京航空航天大学,百度 · 2025年



