{
  "abstract": "Time series foundation models (TSFMs) commonly adapt to new data by attaching a single trainable head to a frozen backbone, a one-size-fits-all setup that underfits heterogeneous regimes. Replacing the head with a mixture of experts is the standard upgrade, but on instance-normalized backbones (the dominant TSFM design class) it fails: routing entropy collapses to zero and one expert absorbs every input, a failure we call normalization-induced routing collapse. Standard MoE rescue mechanisms do not repair it, because the cause is in the router's input, not its optimization. Pre-encoder normalization strips the statistics a router would need to tell regimes apart. A mutual-information decomposition makes this precise and yields a signal-ratio that, computed before training, predicts dataset vulnerability (Spearman $ρ= -0.88$). Eight causal controls, including a vision-modality replication, isolate instance normalization as the cause. The prescription is a minimal causal intervention: Raw-Routed Mixture of Adapters (RR-MoA), which routes on the raw, pre-normalization input. Under a strictly frozen backbone, RR-MoA wins 54/54 comparisons against the strongest fixed adapter and significantly outperforms LoRA, TRACE, AdaMix, and full fine-tuning. The effect generalizes across six backbones and an imputation task. Frozen RR-MoA also beats full fine-tuning by 12-79% (the Frozen Paradox); two architecturally distinct variants confirm the principle generalizes beyond this specific router.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.39445",
  "author_total": 4,
  "authors": [
    {
      "name": "Hung Phan"
    },
    {
      "name": "Thuy T. Nguyen"
    },
    {
      "name": "Minh Ngoc Dinh"
    },
    {
      "name": "Nhat-Quang Tran"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/548f1a9edvqwm.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.39445",
  "fulltext_url": null,
  "id": "548f1a9edvqwm",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.39445",
  "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|Domain Adaptation and Few-Shot Learning"
    ]
  ],
  "openalex": "W7215029075",
  "pdf_url": "https://arxiv.org/pdf/2609.39445",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-30",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/548f1a9edvqwm.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:25:58Z",
      "native_id": "2609.39445",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.39445"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Raw-Routed Mixture of Adapters: A Causal Intervention for Routing Collapse in Time Series Foundation Models",
  "title_text": "Raw-Routed Mixture of Adapters: A Causal Intervention for Routing Collapse in Time Series Foundation Models",
  "type": "preprint",
  "updated_at": "2026-10-30T12:25:58Z",
  "url": "https://research.draconitic.com/paper/548f1a9edvqwm",
  "venue": "arXiv",
  "versions": []
}