To avoid serious diabetic complications, people with type 1 diabetes must keep their blood glucose levels (BGLs) as close to normal as possible. Insulin dosages and carbohydrate consumption are important considerations in managing BGLs. Since the 1960s, models have been developed to forecast blood glucose levels based on the history of BGLs, insulin dosages, carbohydrate intake, and other physiological and lifestyle factors. Such predictions can be used to alert people of impending unsafe BGLs or to control insulin flow in an artificial pancreas. In past work, we have introduced an LSTM-based approach to blood glucose level prediction aimed at "what if" scenarios, in which people could enter foods they might eat or insulin amounts they might take and then see the effect on future BGLs. In this work, we invert the "what-if" scenario and introduce a similar architecture based on chaining two LSTMs that can be trained to make either insulin or carbohydrate recommendations aimed at reaching a desired BG level in the future. Leveraging a recent state-of-the-art model for time series forecasting, we then derive a novel architecture for the same recommendation task, in which the two LSTM chain is used as a repeating block inside a deep residual architecture. Experimental evaluations using real patient data from the OhioT1DM dataset show that the new integrated architecture compares favorably with the previous LSTM-based approach, substantially outperforming the baselines. The promising results suggest that this novel approach could potentially be of practical use to people with type 1 diabetes for self-management of BGLs.
College of Health Solutions, Arizona State University（亚利桑那州立大学健康解决方案学院）；School of Computing and Augmented Intelligence, Arizona State University（亚利桑那州立大学计算与增强智能学院）；Department of Endocrinology, Mayo Clinic Arizona（梅奥诊所亚利桑那分部内分泌科）
Personalised Insulin Adjustment with Reinforcement Learning: An In-Silico Validation for People with Diabetes on Intensive Insulin Treatment
ARTORG Center, University of Bern, Bern, Switzerland；Graduate School for Cellular；Biomedical Sciences, University of Bern, Switzerland；Diabetes Center, National；Kapodistrian University of Athens, Athens, Greece；Department of Endocrinology；Nephrology, Nordsjællands Hospital, Hillerød, Denmark；CARIM School for Cardiovascular Diseases, Maastricht University, Maastricht, The Netherlands；P \& A Kyriakou’ Children’s Hospital, Athens, Greece；Department of Kidney；Department of Clinical Medicine, Faculty of Health；Medical Sciences, University of Copenhagen, Copenhagen, Denmark；Department of Internal Medicine, Maastricht University Medical Centre+, Maastricht, The Netherlands；Department of Internal Medicine, Radboud University Medical Centre, Nijmegen, The Netherlands
MetaboNet-Bench: A Multi-modal Benchmark for Glucose Forecasting in Type 1 Diabetes
Department of Genetics, Stanford University School of Medicine（斯坦福大学医学院遗传学系）；Replica Health；Boston Children’s Hospital, Harvard Medical School（哈佛医学院波士顿儿童医院）；Diabetes Research Institute, Mills-Peninsula Medical Center（米尔斯半岛医学中心糖尿病研究所）
GUIDE: Reinforcement Learning for Behavioral Action Support in Type 1 Diabetes
College of Health Solutions, Arizona State University（亚利桑那州立大学健康解决方案学院）；School of Computing and Augmented Intelligence, Arizona State University（亚利桑那州立大学计算与增强智能学院）；School for Engineering of Matter Transport and Energy, Arizona State University（亚利桑那州立大学物质传输与能源工程学院）；Institute for Foundation of Machine Learning, The University of Texas at Austin（德克萨斯大学奥斯汀分校机器学习基础研究所）；Sony AI（索尼人工智能）
GlyTwin: Digital Twin for Glucose Control in Type 1 Diabetes Through Optimal Behavioral Modifications Using Patient-Centric Counterfactuals
College of Health Solutions, Arizona State University（亚利桑那州立大学健康解决方案学院）；School of Computing and Augmented Intelligence, Arizona State University（亚利桑那州立大学计算与增强智能学院）；Department of Endocrinology, Mayo Clinic Arizona（梅奥诊所亚利桑那分部内分泌科）
