{
  "id": 5043,
  "url": "https://arxiv.org/abs/2605.02973v3",
  "title": "Structured Diffusion Bridges: Inductive Bias for Denoising Diffusion Bridges",
  "summary": "Modality translation is inherently under-constrained, as multiple cross-modal mappings may yield the same marginals. Recent work has shown that diffusion bridges are effective for this task. However, most existing approaches rely on fully paired datasets, thereby imposing a single data-driven constraint. We propose a diffusion-bridge framework that characterizes the space of admissible solutions and restricts it via alignment constraints, treating paired supervision as an optional heuristic rath",
  "authors": "Eitan Kosman, Gabriele Serussi, Chaim Baskin",
  "category": "research",
  "topics": "bias-fairness,safety-alignment",
  "orgs": null,
  "regions": null,
  "published_at": "2026-05-03T16:17:38.000Z",
  "fetched_at": "2026-07-14T16:31:26.338Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/5043",
  "original_url": "https://arxiv.org/abs/2605.02973v3",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}