{
  "abstract": "Joint Energy-Based Models (JEM) unify classification and generation within a single network and support out-of-distribution (OOD) detection. Canonical JEM training relies on stochastic gradient Langevin dynamics (SGLD); a theoretically motivated alternative, the Predictor-Corrector (PC) sampler, has not previously undergone a systematic replication test on the canonical model. We reproduce canonical JEM on WideResNet-28-10 without normalisation layers on two independent runs and test a fixed-noise PC adaptation - with the degenerate annealed-noise predictor replaced by a deterministic gradient step - across three protocols: the adapted sampler replacing SGLD throughout the full training trajectories (115-132 epochs); cold-start generation (FID); and refinement-style multi-OOD detection (AUROC). The reconstruction reaches 92.88% test accuracy and buffer-FID 44.46 (canonical: 92.90% and 38.40). We document two failure modes: catastrophic late-training divergence with the signature of the canonical outlier-buffer mechanism (all four runs), and run-dependent SVHN OOD-discrimination dynamics. No consistent method-level advantage of the adaptation over SGLD is observed on any protocol: refinement AUROC differences stay below 0.007 across ten checkpoint-OOD pairs; seeded cold-start generation favours SGLD by about five FID points; on the training protocol a hierarchical seed-by-image bootstrap gives a 95% confidence interval on the macro-averaged AUROC difference that contains zero, while a seed-level equivalence test with two runs per method cannot establish formal equivalence. The training-protocol data are consistent both with equivalence and with a small directional effect. This outcome is consistent with theory: the guarantees of the annealed-noise PC framework do not transfer to the constant-noise regime of canonical JEM.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": "Journal reference: Discover Computing 29, 755 (2026)",
  "arxiv": "2608.05025",
  "author_total": 1,
  "authors": [
    {
      "name": "Dmytro Knopov"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/2yenac7d0jexx.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2608.05025",
  "fulltext_url": null,
  "id": "2yenac7d0jexx",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2608.05025",
  "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": null,
  "pdf_url": "https://arxiv.org/pdf/2608.05025",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-14",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/2yenac7d0jexx.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:29:42Z",
      "native_id": "2608.05025",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2608.05025"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Comparing SGLD and a fixed-noise predictor-corrector adaptation in canonical joint energy-based models on CIFAR-10",
  "title_text": "Comparing SGLD and a fixed-noise predictor-corrector adaptation in canonical joint energy-based models on CIFAR-10",
  "type": "preprint",
  "updated_at": "2026-10-30T12:29:42Z",
  "url": "https://research.draconitic.com/paper/2yenac7d0jexx",
  "venue": "arXiv",
  "versions": [
    {
      "doi": "10.1007/s10791-026-10679-9",
      "pid": null,
      "relation": "published-as",
      "title": null,
      "venue": null
    }
  ]
}