{
  "abstract": "Speculative decoding in Mixture-of-Experts (MoE) models faces the problem of unstable verification cost caused by input-dependent expert loading. To study the physics of this process, we formulate speculation-budget selection as an offline Stochastic Shortest Path (SSP) problem over reference sequences and build a diagnostic Oracle that uses counterfactual simulation to account for MoE verification cost. A detailed analysis of the Oracle's decisions on the Qwen3-Coder and EAGLE-3 pairing, in the space of marginal deltas (Delta Space), shows that rejected candidates form a strict linear boundary. This result demonstrates that a complex global optimization is locally governed by a necessary condition balancing marginal cost against expected progress ($\\frac{Δ\\mathbb{E}[Cost]}{Δ\\mathbb{E}[a]}$), providing a rigorous mathematical reference point for designing future adaptive online heuristics.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.22156",
  "author_total": 2,
  "authors": [
    {
      "name": "Aidar Amankulov"
    },
    {
      "name": "Denis Mamatin"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/3ywcc6adr7a30.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.22156",
  "fulltext_url": null,
  "id": "3ywcc6adr7a30",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.22156",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1702|Natural Language Processing Techniques"
    ]
  ],
  "openalex": "W7214002098",
  "pdf_url": "https://arxiv.org/pdf/2609.22156",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-01",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/3ywcc6adr7a30.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:23:33Z",
      "native_id": "2609.22156",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.22156"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "The Limits of Speculation: Bounding Speculative Decoding in Mixture-of-Experts",
  "title_text": "The Limits of Speculation: Bounding Speculative Decoding in Mixture-of-Experts",
  "type": "preprint",
  "updated_at": "2026-10-30T12:23:33Z",
  "url": "https://research.draconitic.com/paper/3ywcc6adr7a30",
  "venue": "arXiv",
  "versions": []
}