{
  "id": 19189,
  "url": "https://arxiv.org/abs/2608.12845v1",
  "title": "FSGR: Mitigating Token Frequency Bias for Fair SID-Based Generative Recommendation",
  "summary": "Semantic ID (SID)-based generative recommendation has recently achieved remarkable success. However, existing methods suffer from a previously overlooked fairness issue, which we term \\textbf{Token Frequency Bias}, where high-frequency SID tokens are systematically over-predicted while low-frequency SID tokens are under-predicted. This bias originates from the combined effects of imbalanced semantic codebooks during SID construction, and popularity bias together with the maximum likelihood estim",
  "authors": "Yuchen Zheng, Sihan Xu, Jingwen Yang, Xiangrui Cai, Haiwei Zhang, Xiaojie Yuan",
  "category": "research",
  "topics": "bias-fairness",
  "orgs": null,
  "regions": null,
  "published_at": "2026-08-13T05:34:51.000Z",
  "fetched_at": "2026-08-14T05:10:49.168Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/19189",
  "original_url": "https://arxiv.org/abs/2608.12845v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}