{
  "abstract": "Joint-embedding predictive architectures (JEPA) learn representations by predicting in latent space, as in computer vision; retaining the action-conditioned predictor at inference turns them into latent world models, enabling planning in robotics (V-JEPA 2-AC). Bringing this design to EHR patient trajectories---a predictor that simulates a patient's trajectory in latent space---has not been explored. We use an LLM as the encoder, reading the hourly record as text, avoiding feature engineering and vocabulary harmonisation. But an LLM adapted by supervised fine-tuning does not organise its latent space around physiological dynamics, and freezing it to train the predictor, as in V-JEPA 2-AC, leaves the encoder unaware of the rollout signal: the predictor degrades under rollout. We instead co-train encoder and predictor under one latent-prediction objective, grounding the encoder in the dynamics its predictor must follow. Naïve co-training, however, is unstable: the untrained predictor drags the encoder toward collapse, and the predictor's rollout diverges as its target space moves. We present Clin-JEPA, a five-phase curriculum that stably co-trains an LLM encoder with a latent trajectory predictor on MIMIC-IV. Three evaluations support the design: (1) under 48-hour autoregressive rollout the co-trained predictor degrades least (predictor degradation $\\times$1.06, against $\\times$1.23--1.36 for two-stage designs and $\\times$6.3--66 for curriculum ablations) while the co-trained encoder resolves the progression of patient state most sharply (largest state displacement); (2) the co-trained encoder separates deteriorating from stable patients in its latent space with Cohen's $d{=}1.59$, against $\\leq$0.50 for two-stage encoders; (3) one set of embeddings serves 34 downstream tasks across three benchmarks, outperforming strong per-task tuned baselines and a pretrained EHR foundation model.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2605.10840",
  "author_total": 13,
  "authors": [
    {
      "name": "Yixuan Yang"
    },
    {
      "name": "Mehak Arora"
    },
    {
      "name": "Ryan Zhang"
    },
    {
      "name": "Baraa Abed"
    },
    {
      "name": "Junseob Kim"
    },
    {
      "name": "Tilendra Choudhary"
    },
    {
      "name": "Md Hassanuzzaman"
    },
    {
      "name": "Kevin Zhu"
    },
    {
      "name": "Ayman Ali"
    },
    {
      "name": "Chengkun Yang"
    },
    {
      "name": "Alasdair Edward Gent"
    },
    {
      "name": "Victor Moas"
    },
    {
      "name": "Rishikesan Kamaleswaran"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/9710wx9w99n33.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2605.10840",
  "fulltext_url": null,
  "id": "9710wx9w99n33",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2605.10840",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.AI"
    ],
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "q-bio.QM"
    ],
    [
      "openalex",
      "1702|Machine Learning in Healthcare"
    ]
  ],
  "openalex": "W7160957168",
  "pdf_url": "https://arxiv.org/pdf/2605.10840",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-16",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/9710wx9w99n33.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:22:24Z",
      "native_id": "2605.10840",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2605.10840"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "arxiv:q-bio.QM",
      "subfield": "bioinformatics"
    }
  ],
  "title": "Clin-JEPA: A Multi-Phase Co-Training Framework for Joint-Embedding Predictive Pretraining on EHR Patient Trajectories",
  "title_text": "Clin-JEPA: A Multi-Phase Co-Training Framework for Joint-Embedding Predictive Pretraining on EHR Patient Trajectories",
  "type": "preprint",
  "updated_at": "2026-10-30T12:22:24Z",
  "url": "https://research.draconitic.com/paper/9710wx9w99n33",
  "venue": "arXiv",
  "versions": []
}