{
  "paper": {
    "arxiv_id": "2607.23718",
    "title": "Melo: A Production LLM-Powered Music Recommendation Agent",
    "url": "https://arxiv.org/abs/2607.23718",
    "organization": "NetEase Cloud Music / Zhejiang University of Technology"
  },
  "dataset": {
    "name": "MovieLens 100K",
    "users": 420,
    "items": 620
  },
  "setup": {
    "adapter": "melo",
    "same_split_and_candidates": true
  },
  "baseline": {
    "name": "catalog recommender without runtime repair",
    "hit_at_10": 0.05952380952380952,
    "ndcg_at_10": 0.026292301768631662,
    "fresh_hit_at_10": 0.020833333333333332,
    "head_share_at_10": 0.48333333333333334
  },
  "method": {
    "name": "five-node agent + entity grounding + reflective retry",
    "hit_at_10": 0.047619047619047616,
    "ndcg_at_10": 0.024147060037386987,
    "fresh_hit_at_10": 0.03125,
    "head_share_at_10": 0.15523809523809523
  },
  "relative": {
    "hit_at_10_percent": -20.0,
    "ndcg_at_10_percent": -8.159200933119065,
    "fresh_hit_at_10_percent": 50.00000000000001,
    "head_share_at_10_percent": -67.88177339901478
  },
  "stages": {
    "state_graph_nodes": [
      "intent",
      "entity grounding",
      "tool plan",
      "execution",
      "response"
    ],
    "search_index_verification": true,
    "reflective_retry": true,
    "selected_validation_blend": 0.30000000000000004,
    "validation_only_model_selection": {
      "hit_at_10": 0.08095238095238096,
      "ndcg_at_10": 0.04300575662567052,
      "fresh_hit_at_10": 0.06666666666666667,
      "head_share_at_10": 0.16666666666666666
    },
    "grounding_rejections": 0,
    "reflective_retries": 0
  },
  "paper_results": {
    "playlist_retention_pp_lower_bound": 2.0,
    "playlist_engagement_minutes_lower_bound": 1.0,
    "entity_misidentification_reduction_pp": 7.8,
    "retry_trigger_percent": 5.8,
    "retry_recovery_percent": 59.0
  },
  "scope": "实际执行 catalog entity gate、失败原因驱动的约束放宽和重试，不把简单热门回退冒充恢复；MovieLens item catalog 替代网易云音乐搜索索引和线上工具服务。",
  "schema_version": 2,
  "manifest_ref": "reproduction:melo",
  "evaluation_protocol": {
    "tier": "l2_public_dataset",
    "seeds": [
      42
    ],
    "formal_comparison": false,
    "claim_policy": "single/few-seed smoke result; do not claim a stable improvement"
  },
  "provenance": {
    "artifact_path": "docs/reproductions/2607.23718-melo/metrics/result-seed42.json",
    "historical_migration": "historical-metrics-v2-2026-08-09",
    "original_code_commit": "not recorded",
    "dataset_fingerprint": "not recorded in historical artifact"
  }
}