eprintid: 10117791 rev_number: 17 eprint_status: archive userid: 608 dir: disk0/10/11/77/91 datestamp: 2020-12-18 14:26:35 lastmod: 2021-12-10 23:38:56 status_changed: 2020-12-18 14:26:35 type: proceedings_section metadata_visibility: show creators_name: Zhu, T creators_name: Li, K creators_name: Georgiou, P title: Personalized Dual-Hormone Control for Type 1 Diabetes Using Deep Reinforcement Learning ispublished: pub divisions: UCL divisions: B02 divisions: DD4 note: This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. abstract: We introduce a dual-hormone control algorithm for people with Type 1 Diabetes (T1D) which uses deep reinforcement learning (RL). Specifically, double dilated recurrent neural networks are used to learn the control strategy, trained by a variant of Q-learning. The inputs to the model include the real-time sensed glucose and meal carbohydrate content, and the outputs are the actions necessary to deliver dual-hormone (basal insulin and glucagon) control. Without prior knowledge of the glucose-insulin metabolism, we develop a data-driven model using the UVA/Padova Simulator. We first pre-train a generalized model using long-term exploration in an environment with average T1D subject parameters provided by the simulator, then adopt importance sampling to train personalized models for each individual. In-silico, the proposed algorithm largely reduces adverse glycemic events, and achieves time in range, i.e., the percentage of normoglycemia, 93% for the adults and 83% for the adolescents, which outperforms previous approaches significantly. These results indicate that deep RL has great potential to improve the treatment of chronic diseases such as diabetes. date: 2020 date_type: published publisher: Springer official_url: https://doi.org/10.1007/978-3-030-53352-6_5 oa_status: green full_text_type: other language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 1836425 doi: 10.1007/978-3-030-53352-6_5 isbn_13: 9783030533519 lyricists_name: Li, Kezhi lyricists_id: KLIXX57 actors_name: Li, Kezhi actors_id: KLIXX57 actors_role: owner full_text_status: public series: Studies in Computational Intelligence publication: Studies in Computational Intelligence volume: 914 place_of_pub: Cham, Switzerland pagerange: 45-53 book_title: Explainable AI in Healthcare and Medicine editors_name: Shaban-Nejad, A editors_name: Michalowski, M editors_name: Buckeridge, DL citation: Zhu, T; Li, K; Georgiou, P; (2020) Personalized Dual-Hormone Control for Type 1 Diabetes Using Deep Reinforcement Learning. In: Shaban-Nejad, A and Michalowski, M and Buckeridge, DL, (eds.) Explainable AI in Healthcare and Medicine. (pp. pp. 45-53). Springer: Cham, Switzerland. Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10117791/1/Personalized%20Dual-Hormone%20Control%20for%20Type%201%20Diabetes%20Using%20DeepReinforcement%20Learning.pdf