2601.17063v1.pdf

{
  "file/base": {
    "record": {
      "extension": ".pdf",
      "hash": "c1891af0ed6d3f7ba9976804b0605c273291c745bfa987322240e7047ec38d33",
      "media_type": "application/pdf",
      "media_type_prefix": "application",
      "name": "2601.17063v1.pdf",
      "size": 2248326
    },
    "source": {
      "execution_id": "9d7aed85-cf41-4144-ad8c-57296b4ceff0",
      "id": "file/base",
      "type": "Model",
      "version": "1.0.0"
    }
  },
  "file/pdf": {
    "private": false,
    "record": {
      "creation_date": "2026-01-27T01:02:24Z",
      "keywords": [
        "Mixture-of-Experts",
        "Offloading",
        "LLM Inference",
        "Scheduling",
        "Cache Replacement Policies"
      ],
      "modified_date": "2026-01-27T01:02:24Z",
      "page_count": 7,
      "producer": "pikepdf 8.15.1",
      "title": "FlashMoE: Reducing SSD I/O Bottlenecks via ML-Based Cache Replacement for Mixture-of-Experts Inference on Edge Devices",
      "version": "1.7"
    },
    "source": {
      "execution_id": "75f4639e-905b-450e-8e4a-32db763e5398",
      "id": "dorsal/pdf",
      "type": "Model",
      "version": "1.1.0"
    }
  }
}