?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=Personalized+Dual-Hormone+Control+for+Type+1+Diabetes+Using+Deep+Reinforcement+Learning&rft.creator=Zhu%2C+T&rft.creator=Li%2C+K&rft.creator=Georgiou%2C+P&rft.description=We+introduce+a+dual-hormone+control+algorithm+for+people+with+Type+1+Diabetes+(T1D)+which+uses+deep+reinforcement+learning+(RL).+Specifically%2C+double+dilated+recurrent+neural+networks+are+used+to+learn+the+control+strategy%2C+trained+by+a+variant+of+Q-learning.+The+inputs+to+the+model+include+the+real-time+sensed+glucose+and+meal+carbohydrate+content%2C+and+the+outputs+are+the+actions+necessary+to+deliver+dual-hormone+(basal+insulin+and+glucagon)+control.+Without+prior+knowledge+of+the+glucose-insulin+metabolism%2C+we+develop+a+data-driven+model+using+the+UVA%2FPadova+Simulator.+We+first+pre-train+a+generalized+model+using+long-term+exploration+in+an+environment+with+average+T1D+subject+parameters+provided+by+the+simulator%2C+then+adopt+importance+sampling+to+train+personalized+models+for+each+individual.+In-silico%2C+the+proposed+algorithm+largely+reduces+adverse+glycemic+events%2C+and+achieves+time+in+range%2C+i.e.%2C+the+percentage+of+normoglycemia%2C+++93%25++for+the+adults+and+++83%25++for+the+adolescents%2C+which+outperforms+previous+approaches+significantly.+These+results+indicate+that+deep+RL+has+great+potential+to+improve+the+treatment+of+chronic+diseases+such+as+diabetes.&rft.publisher=Springer&rft.contributor=Shaban-Nejad%2C+A&rft.contributor=Michalowski%2C+M&rft.contributor=Buckeridge%2C+DL&rft.date=2020&rft.type=Proceedings+paper&rft.language=eng&rft.source=+++++In%3A+Shaban-Nejad%2C+A+and+Michalowski%2C+M+and+Buckeridge%2C+DL%2C+(eds.)+Explainable+AI+in+Healthcare+and+Medicine.++(pp.+pp.+45-53).++Springer%3A+Cham%2C+Switzerland.+(2020)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10117791%2F1%2FPersonalized%2520Dual-Hormone%2520Control%2520for%2520Type%25201%2520Diabetes%2520Using%2520DeepReinforcement%2520Learning.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10117791%2F&rft.rights=open