{
  "annotation_id": "57c3d79e-a47b-4d55-a0c3-af0edacdf3a2",
  "date_created": "2026-02-17T05:53:11.345000Z",
  "date_modified": "2026-02-17T05:53:11.345000Z",
  "file_hash": "0218e47115867ac986a491d73f03c1aaf7a3312a95b706eb09fc0d3c341a5a27",
  "private": false,
  "record": {
    "abstract": "Despite having hundreds of millions of speakers, Chinese dialects lag behind Mandarin in speech and language technologies. Most varieties are primarily spoken, making dialect-to-Mandarin speech-LLMs (large language models) more practical than dialect LLMs. Building dialect-to-Mandarin speech-LLMs requires speech representations with cross-dialect semantic alignment between Chinese dialects and Mandarin. In this paper, we achieve such a cross-dialect semantic alignment by training a speech encoder with ASR (automatic speech recognition)-only data, as demonstrated by speech-to-speech retrieval on a new benchmark of spoken Chinese varieties that we contribute. Our speech encoder further demonstrates state-of-the-art ASR performance on Chinese dialects. Together, our Chinese dialect benchmark, semantically aligned speech representations, and speech-to-speech retrieval evaluation lay the groundwork for future Chinese dialect speech-LLMs. We release the benchmark at https://github.com/kalvinchang/yubao.",
    "arxiv_id": "2601.07274",
    "authors": [
      "Kalvin Chang",
      "Yiwen Shao",
      "Jiahong Li",
      "Dong Yu"
    ],
    "categories": [
      "cs.CL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "title": "Towards Comprehensive Semantic Speech Embeddings for Chinese Dialects",
    "url": "https://arxiv.org/abs/2601.07274",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "c50aef3c-2ba4-44a0-b48b-5f50edd700f5",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}