Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Future-aware blood glucose forecasting using knowledge distillation with transformer-based sequence-to-sequence

Xiaoyu Sun1, Hongru Li2, Xia Yu2

  • 1College of Information Science and Technology, Northeastern University, Shenyang, 110819, China. sunxiaoyu1@ise.neu.edu.cn.

Scientific Reports
|March 1, 2026
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

Improving Translational Accuracy02:07

Improving Translational Accuracy

15.3K
Base complementarity between the three base pairs of mRNA codon and the tRNA anticodon is not a failsafe mechanism. Inaccuracies can range from a single mismatch to no correct base pairing at all. The free energy difference between the correct and nearly correct base pairs can be as small as 3 kcal/ mol. With complementarity being the only proofreading step, the estimated error frequency would be one wrong amino acid in every 100 amino acids incorporated. However, error frequencies observed in...
15.3K
Improving Translational Accuracy02:07

Improving Translational Accuracy

3.7K
3.7K

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Endocytic motif on HIV-1 Env regulates cleavage status and antibody neutralization of cell-free and cell-to-cell infection.

Journal of virology·2026
Same author

Corrosion Inhibition of Imidazolidinyl Urea-Based Zinc Anode Stabilization for Achieving Superbly Reversible Aqueous Zinc Metal Batteries.

Langmuir : the ACS journal of surfaces and colloids·2026
Same author

Butyrate protects against Klebsiella pneumoniae-induced oxidative stress in alveolar macrophages via p62-Keap1-Nrf2 pathway.

Redox biology·2026
Same author

How should we interpret the use of vonoprazan and proton pump inhibitors during immunotherapy for hepatocellular carcinoma?

International journal of clinical oncology·2026
Same author

1064 nm InGaAsP multi-junction laser power converters via reverse-bias method.

Optics express·2026
Same author

Research on the influence of micro-pile layout of stability of landslide resistance.

Scientific reports·2026
Same journal

Turbulent flow in a vortex separator with a directed pipe inlet.

Scientific reports·2026
Same journal

Systematic characteristic evaluation of clay-based cementitious material derived from calcium carbide residue and waste tile powder.

Scientific reports·2026
Same journal

Retraction Note: Improvement of a rapid diagnostic application of monoclonal antibodies against avian influenza H7 subtype virus using Europium nanoparticles.

Scientific reports·2026
Same journal

Applying large language models to spam detection in the Kazakh low-resource language setting.

Scientific reports·2026
Same journal

An open-source 3D printing system enabling in-situ freeze-thaw processing of hydrogels.

Scientific reports·2026
Same journal

An enhanced EfficientNet framework for automated waste classification using cosine annealing and label smoothing.

Scientific reports·2026
See all related articles

This study introduces a future-aware learning framework for improved blood glucose level (BGL) forecasting in diabetes management. The method enhances prediction accuracy by using future data during training, enabling reliable real-time glucose monitoring.

Area of Science:

  • Biomedical Engineering
  • Artificial Intelligence in Healthcare
  • Diabetes Technology

Background:

  • Accurate blood glucose level (BGL) forecasting is vital for diabetes self-management and clinical decisions.
  • Current deep learning models using continuous glucose monitoring (CGM) data often exclude future factors like insulin or meals, limiting real-time forecasting.
  • This gap hinders the development of precise and reliable glucose prediction tools.

Purpose of the Study:

  • To develop a novel future-aware learning framework for multi-step BGL prediction.
  • To leverage privileged future information during training while ensuring deployability at inference.
  • To enhance the accuracy and clinical reliability of BGL forecasting models.

Main Methods:

  • A Transformer-based teacher model was trained offline using historical CGM data and future disturbance information (insulin, meals).

Related Experiment Videos

  • A student model, trained via knowledge distillation, learned to approximate the teacher's predictions using only historical data.
  • The framework was evaluated on the OhioT1DM and AZT1D datasets for prediction horizons of 30-120 minutes.

    • Main Results:

    • The proposed framework demonstrated consistent reductions in root mean squared error (RMSE) and mean absolute error (MAE) compared to established methods.
    • Over 90% of predictions achieved clinical reliability, as validated by Clarke error grid analysis.
    • The future-aware approach significantly improved glucose forecasting performance under realistic deployment constraints.

    Conclusions:

    • Future-aware training strategies can substantially enhance glucose forecasting accuracy and reliability.
    • The developed framework offers a practical solution for real-time BGL prediction without requiring future data at deployment.
    • This approach holds significant potential for improving diabetes self-management and clinical decision-making.