DigiNews

Tech Watch by Johan Denoyer

← Back to articles

Show HN: I have trained 15 blood glucose forecasting models on my own data

Quality: 8/10 Relevance: 9/10

Summary

A GitHub repository showcases 15 blood glucose forecasting models trained on personal data, featuring a small transformer (T1DMAI D16) with LoRA adapters, evaluated on both personal and MetaboNet data against classical models like XGBoost and LightGBM. The work demonstrates AI-driven time-series forecasting in a biomedical context and provides open-source code and results, though it remains a research artifact rather than a medical device.

🚀 Service construit par Johan Denoyer