# CGM-JEPA: Learning Consistent Continuous Glucose Monitor Representations via Predictive Self-Supervised Pretraining

arXiv: https://arxiv.org/abs/2605.00933 · HTML: https://mlsanity.com/paper/2605.00933/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 38 references that are in the map.
- First posted on arXiv 1 May 2026.

## Builds on (most cited first, 17 of 17)

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Momentum Contrast for Unsupervised Visual Representation Learning](https://mlsanity.com/paper/1911.05722/) (2019)
- [A Time Series is Worth 64 Words: Long-term Forecasting with Transformers](https://mlsanity.com/paper/2211.14730/) (2022)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [Chronos: Learning the Language of Time Series](https://mlsanity.com/paper/2403.07815/) (2024)
- [A decoder-only foundation model for time-series forecasting](https://mlsanity.com/paper/2310.10688/) (2023)
- [A Survey on Contrastive Self-supervised Learning](https://mlsanity.com/paper/2011.00362/) (2020)
- [MOMENT: A Family of Open Time-series Foundation Models](https://mlsanity.com/paper/2402.03885/) (2024)
- [A Tutorial on Principal Component Analysis](https://mlsanity.com/paper/1404.1100/) (2014)
- [VL-JEPA: Joint Embedding Predictive Architecture for Vision-language](https://mlsanity.com/paper/2512.10942/) (2025)
- [Brain-JEPA: Brain Dynamics Foundation Model with Gradient Positioning and Spatiotemporal Masking](https://mlsanity.com/paper/2409.19407/) (2024)
- [SensorLLM: Aligning Large Language Models with Motion Sensors for Human Activity Recognition](https://mlsanity.com/paper/2410.10624/) (2024)
- [Self-Supervised Pre-Training with Joint-Embedding Predictive Architecture Boosts ECG Classification Performance](https://mlsanity.com/paper/2410.13867/) (2024)
- [A Large Sensor Foundation Model Pretrained on Continuous Glucose Monitor Data for Diabetes Management](https://mlsanity.com/paper/2412.09727/) (2024)
- [Self-supervised Regularization for Text Classification](https://mlsanity.com/paper/2103.05231/) (2021)
- [Use of Continuous Glucose Monitoring with Machine Learning to Identify Metabolic Subphenotypes and Inform Precision Lifestyle Changes](https://mlsanity.com/paper/2511.03986/) (2025)
- [Contextual Motifs: Increasing the Utility of Motifs using Contextual Data](https://mlsanity.com/paper/1703.02144/) (2017)

## Built on by (most cited first, 6 of 6)

- [GlucoFM: A Dual-Stream Foundation Model for Continuous Glucose Monitoring](https://mlsanity.com/paper/2605.30865/) (2026)
- [TRACE-TS: Attribution-Grounded and Traceable Sensor-Language Reasoning for Human Activity Understanding](https://mlsanity.com/paper/2608.00200/) (2026)
- [Transferable Evidence Reconstruction for Longitudinal Glucose Representations](https://mlsanity.com/paper/2609.28199/) (2026)
- [BeliefGraph-JEPA: Structured Latent World Models for Action-Conditioned Time Series](https://mlsanity.com/paper/2610.05409/) (2026)
- [GlucoPRISM: Protocol-Supervised Pretraining for Cross-Cohort Glucose Transfer](https://mlsanity.com/paper/2610.12528/) (2026)
- [TEX-MEX: Similarity-Guided Temporal Contrastive Learning for Time Series](https://mlsanity.com/paper/2610.19557/) (2026)

## Abstract (arXiv)

Continuous Glucose Monitoring (CGM) can detect early metabolic subphenotypes (insulin resistance, IR; β-cell dysfunction), but population-scale deployment faces two coupled problems. First, the same physiological state appears through multiple views (CGM time series, venous OGTT, Glucodensity summaries), so single-view representations fail to transfer when deployment shifts the modality or setting. Second, baselines perform inconsistently across these shifts. Both problems point to one remedy: representations that abstract away from any single view to capture higher-level temporal and distributional structure. We propose CGM-JEPA, a self-supervised pretraining framework which predicts masked latent representations rather than raw values, yielding abstraction that transfers across modalities. X-CGM-JEPA adds a masked Glucodensity cross-view objective for complementary distributional information. We pretrain on ~389k unlabeled CGM readings from 228 subjects and evaluate on two clinical cohorts (N=27 and N=17 public-release subsets) across three regimes (cohort generalization, venous-to-CGM transfer, home CGM) under 20-iteration × 2-fold cross-validation. X-CGM-JEPA ranks first or second on AUROC for both endpoints across all three regimes while no baseline does, exceeding the strongest baseline by up to +6.5 pp in cohort generalization and +3.6 pp in venous-to-CGM transfer (paired Wilcoxon, p<0.001). Under modality shift, it matches mean AUROC while redistributing toward weaker subgroups (ethnicity AUROC gap shrinks 25-54%); on sparse in-domain venous data, the distributional view lifts label-aware clustering (ARI +39%, NMI +40%). Code and weights: https://github.com/cruiseresearchgroup/CGM-JEPA