Interpretable Language Model for Closed-Loop Type 1 Diabetes Control
Abstract: Type 1 Diabetes (T1D) is a chronic, life-threatening autoimmune condition characterized by the complete destruction of insulin-producing pancreatic beta cells. While Artificial Pancreas Systems (APS) powered by Reinforcement Learning (RL) have shown promise in automating insulin delivery, their ``black-box'' nature makes it hard for patients and doctors to trust them fully. This paper presents LLM-T1D, a promising approach that combines the precision of RL with the clear, human-like reasoning of Large Language Models (LLMs) to create a more transparent and reliable insulin pump controller. By training an expert RL system and distilling its knowledge into fine-tuned LLaMA 3.1 8B and Qwen3 8B models, we developed a controller that not only surpasses the RL system's performance but also explains its decisions in plain, understandable language. Tested on the FDA-approved UVA/Padova T1D simulator, the LLM controllers deliver excellent blood sugar control (73.5% Time in Range) while maintaining strict formal safety verification against hallucinations.
Submission history
Access Paper:
Current browse context:
References & Citations
BibTeX formatted citation
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .
Verified source · arXiv.org
Reported by arXiv.org. Open the original for full media and formatting.
More in Research
All news
ResearchPebble founder Eric Migicovsky says his 30-day warranty is all about trust
Pebble founder Eric Migicovsky says buyers of its new e-paper smartwatches should know what they're signing up for and trust Pebble to make things right if they run into issues, despite the short warranty. "I think the most important thing is trust," Migicovsky told me in an int…
Read at The Verge
ResearchThe war on ‘woke science’ comes for space research
The Trump administration is waging a culture war on science, and the latest salvo is in the form of a dry, bureaucratic proposal from the Office of Management and Budget (OMB) that could threaten the future of US science as we know it. The proposal would give political appointee…
Read at The VergeHG-RAG: Hierarchy-Guided Retrieval-Augmented Generation for Structured Knowledge Graphs
Retrieval Augmented Generation (RAG) has proven to be a widely successful process at improving the quality of outputs from a Large Language Model (LLM) for wider context. However, RAG systems typically retrieve context from flat document stores, which struggles when queries require hierarchical or relational reasoning across structured knowledge. I present HG-RAG (Hierarchy-Guided RAG), a framework that performs graph-traversal over a hierarchical knowledge graph to deliver structured context to a language model. My retrieval pipeline resolves a named entity anchor from the query, then expand…
Read at arXiv cs.AIIntelligent Three Level Learning Architecture for Autonomous UAV Swarms in Search and Rescue
This paper presents a novel three level hierarchical learning architecture for autonomous UAV swarms performing search and rescue operations. Unlike conventional approaches that apply a single learning paradigm across all hierarchy levels, the proposed architecture integrates three qualitatively different learning mechanisms corresponding to the biological hierarchy of reflexes, skills, and reasoning such as Hebbian neuroplasticity for individual agent adaptation, multi agent reinforcement learning with graph neural networks and behavior trees for tactical coordination, and model agnostic met…
Read at arXiv cs.AI