লেখক : Ittai Dayan Holger R. রোথ Aoxiao Zhong Ahmed Harouni Amilcare Gentili Anas Z. Abidin Andrew Liu Anthony Beardsworth Costa Bradford J. Wood Chien-Sung Tsai Chih-Hung Wang Chun-Nan Hsu C. K. Lee Peiying Ruan Daguang Xu Dufan Wu Eddie Huang Felipe Campos Kitamura Griffin Lacey Gustavo César de Antônio Corradi Gustavo Nino Hao-Hsin Shin Hirofumi Obinata Hui Ren Jason C. Crane Jesse Tetreault Jiahui Guan John W. Garrett Joshua D. Kaggie Jung Gil Park Keith Dreyer Krishna Juluru Kristopher Kersten Marcio Aloisio Bezerra Cavalcanti Rockenbach Marius George Linguraru Masoom A. Haider Meena AbdelMaseeh Nicola Rieke Pablo F. Damasceno Pedro Mario Cruz e Silva Pochuan Wang Sheng Xu Shuichi Kawano Sira Sriswasdi Soo Young Park Thomas M. Grist Varun Buch Watsamon Jantarabenjakul Weichung Wang Won Young Tak Xiang Li Xihong Lin Young Joon Kwon Abood Quraini Andrew Feng Andrew N. Priest Baris Turkbey Benjamin Glicksberg Bernardo Bizzo Byung Seok Kim Carlos Tor-Díez Chia-Cheng Lee Chia-Jung Hsu Chin Lin Chiu-Ling Lai Christopher P. Hess Colin Compas Deepeksha Bhatia Eric K. Oermann Evan Leibovitz Hisashi Sasaki Hitoshi Mori Isaac Yang Jae Ho Sohn Krishna Nand Keshava Murthy Li-Chen Fu Matheus Ribeiro Furtado de Mendonça Mike Fralick Min Kyu Kang Mohammad Adil Natalie Gangai Peerapon Vateekul Pierre Elnajjar Sarah Hickman Sharmila Majumdar Shelley L. McLeod Sheridan Reed Stefan Gräf Stephanie Harmon Tatsuya Kodama Thanyawee Puthanakit Tony Mazzulli Vitor Lima de Lavor Yothin Rakvongthai Yu Rim Lee Yuhong Wen Fiona J. Gilbert Mona G. Flores Quanzheng Li লেখক : ইতালি ডায়ান Holger R. রোথ অক্সিজো জং আহমেদ হারুনি সহানুভূতিশীল অ্যানাস জি আবিদিন অ্যান্ড্রু লিও অ্যান্টনি Beardsworth কোস্টা ব্র্যাডফোর্ড জে উড গান্ধী সিং চি হং ওয়াং চুন-নান হুয়া C. K. লি রুয়াং রুয়াং ডায়াবেটিস ডুফান ওয়া এডি হুয়াং ফেলিপা ক্যাম্পোস কিটামুরা গ্রিফিন লেসি গুস্তাভো সিজার ডি অ্যান্টোনিও কররাদি গাস্টো নিনো হাও-হসিন শিন হিরোফুমি Obinata হুই রেন জেসন C. ক্রেইন জেসি টেট্রো জিয়াউই গুয়ান জন W. গ্যারেট জোসুয়া ডি ক্যাগি জং গিল পার্ক কিথ ড্রাইয়ার খ্রিস্টান খ্রিস্টান ক্রিস্টোফার ক্রিস্টো Marcio Aloisio Bezerra Cavalcanti রকেনবাক মারিয়াস জর্জ লিংগুরুরু মেসোম এ. হাইডার আব্দুল্লাহ নিকোলা রিক প্যাবলি এফ দামাসেনো মারিও ক্রুজ ও সিলভা পাওনা ওয়াং শং শু শুইচি কায়ানো স্যার সুদীর্ঘ তরুণ পার্ক টমাস এম গ্রিস ভার্জিন বই রাতের খাবার খাওয়া ওয়ান্ডা তরুণী জেনারেল শিয়াং লিও জিয়াং লিন তরুণ জোন কুয়ান কুরআন কুরআন অ্যান্ড্রু ফেন অ্যান্ড্রু ন. প্রাইস্ট টার্কবাই বেঞ্জামিন গ্লিকসবার্গ বার্নার্ডো বিজো খুঁজে বের করুন কিম কার্লোস টর-ডিজ চিয়া-চং লি চিয়া-জং হসু চিন লিন চু-লিং লাই ক্রিস্টোফার পি হেস কোলিন কম্পাস ডেপেকশা বাথিয়া এরিক K. ওয়ারম্যান ইভান লেইবোভিটজ হিজাজি সাসাকি হিটোশি মোরি ইজারা ইয়ান জে হো ছেলে ক্রিশ্না নান্দ কিশাভা মুরথি লু-চেন ফু ম্যাথিও রিবিরো ফুর্টাদো ডি মেনডোনসা মাইক ফ্রালিক মিন কিউ কং মোহাম্মদ আনুগত্য ন্যাটালি গান্ধী পিরিয়ড ভ্যাট পিটার এলনেজার স্যার হিকম্যান মাশরাফি মেধাবী শেলি এল ম্যাকলিওড শেরিডান রিড স্টিফেন গ্রাফ স্টিফেনি হারমন কুমিল্লা কুমিল্লা হেঁটে গেলেন। টনি মাজুলি ভিটর লিমা ওয়ার্ক ইয়াহুদী রাসূলুল্লাহ ইয় রিম লি ইয়াং ওয়াং ফিওনা জি গিলবার্ট মোনা জি ফ্লোরেস লিওনার্দো অবরোধ ফেডারেট সেন্সিং (এফএল) ডেটা একাধিক উত্স থেকে ডেটা ব্যবহার করে কারিগরি মডেল প্রশিক্ষণের জন্য ব্যবহৃত একটি পদ্ধতি যা ডেটা অ্যাননোমাইজেশন বজায় রাখে, এইভাবে ডেটা ভাগ করার জন্য অনেক বাধাগুলি সরিয়ে দেয়। এখানে আমরা বিশ্বব্যাপী 20 ইনস্টিটিউটের ডেটা ব্যবহার করে একটি এএফএল মডেল, যা ইলেকট্রনিক মেডিকেল রেকর্ড (ইএমআর) ব্রেস্ট জেনারেল এআই মডেল নামে বলা হয়, যা লক্ষণীয় রোগীদের অক্সিজেনের ভবিষ্যৎ প্রয়োজনীয়তা পূর্বাভাস করে, COVID-19 এর সাথে অংশগ্রহণকারী সমস্ত সাইটগুলির মধ্যে পরিমাপ করা AUC প্রধান বিজ্ঞান, একাডেমিক, মেডিকেল এবং ডাটা বিজ্ঞান সম্প্রদায়গুলি COVID-19 প্যানডেমিক সংকটের মুখে একত্রিত হয়েছে যাতে ধারাবাহিক সহযোগিতার স্বাভাবিক গোপনীয়তা এবং ডেটা মালিকানা বাধা ছাড়া ডেটা ভাগ এবং মডেল প্রশিক্ষণ এবং পরীক্ষাগুলির জন্য দ্রুত এবং নিরাপদ নতুন ধারণাগুলি মূল্যায়ন করতে পারে। , স্বাস্থ্যসেবা প্রদানকারী, গবেষক এবং শিল্প ক্রিকেটের দ্বারা সৃষ্ট অসম্পূর্ণ এবং গুরুত্বপূর্ণ ক্লিনিক্যাল চাহিদাগুলি মোকাবেলা করার জন্য তাদের মনোযোগ ঘুরিয়েছে, অবিশ্বাস্য ফলাফল সঙ্গে , , , , , , জাতীয় নিয়ন্ত্রক সংস্থা এবং একটি আন্তর্জাতিক সহযোগিতামূলক আত্মা দ্বারা ক্লিনিকাল গবেষণার নিয়োগ দ্রুত এবং সহায়তা করা হয়েছে , , ডেটা অ্যানালিটিক্স এবং এআই বৈজ্ঞাগুলি সবসময় খোলা এবং সহযোগিতামূলক পদ্ধতিগুলি উন্নীত করেছে, উদাহরণস্বরূপ ওপেন সোর্স সফটওয়্যার, পুনরাবৃত্তিযোগ্য গবেষণা, ডেটা রিপোজিটর এবং জনসাধারণের জন্য অ্যানিমাইজড ডেটা সেটগুলি উপলব্ধ করে। , প্যান্ডেমিটি দ্রুত ডেটা সহযোগিতা চালানোর প্রয়োজনকে উত্সাহিত করেছে যা দ্রুত বিবর্তিত এবং বিস্তৃত বিশ্বব্যাপী চ্যালেঞ্জগুলির প্রতিক্রিয়া করার সময় ক্লিনিকাল এবং বৈজ্ঞানিক সম্প্রদায়কে ক্ষমতা দেয়। , , . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 এই ধরনের সহযোগিতার একটি নির্দিষ্ট উদাহরণ হল আমাদের পূর্ববর্তী কাজ একটি এআই-ভিত্তিক SARS-COV-2 ক্লিনিকাল সিদ্ধান্ত সমর্থন (সিডিএস) মডেল. এই সিডিএস মডেলটি মেস জেনারেল ব্রিগাম (এমজিবি) এ বিকাশ করা হয়েছিল এবং একাধিক স্বাস্থ্য সিস্টেমের ডেটাতে সনাক্ত করা হয়েছিল। , , , CXR ইমেজিং ইনপুট হিসাবে নির্বাচিত হয়েছিল কারণ এটি ব্যাপকভাবে উপলব্ধ এবং সাধারণত ACR দ্বারা সরবরাহ করা নির্দেশাবলী দ্বারা নির্দেশ করা হয় Fleischner সমাজ WHO এর জাতীয় thoracic সমাজ , জাতীয় স্বাস্থ্য মন্ত্রণালয় COVID ম্যানুয়াল এবং বিশ্বব্যাপী রেডিওলজি সোসাইটি CDS মডেলের আউটপুট একটি স্কোর ছিল, যা CORISK নামে পরিচিত। , যা অক্সিজেন সমর্থন প্রয়োজনীয়তাগুলির সাথে সামঞ্জস্যপূর্ণ এবং শীর্ষস্থানীয় ক্লিনিকালদের দ্বারা রোগীদের পরীক্ষায় সহায়তা করতে পারে , , স্বাস্থ্যসেবা প্রদানকারীরা তাদের নিজস্ব ডেটা দ্বারা যাচাই করা মডেলগুলি পছন্দ করে। বর্তমানে অধিকাংশ এআই মডেল, উপরে উল্লেখিত সিডিএস মডেল সহ, প্রশিক্ষণ এবং "সীম" ডেটা উপর যাচাই করা হয়েছে যা প্রায়শই বৈচিত্র্যের অভাব। , , সম্ভাব্যভাবে overfitting এবং কম generalizability ফলাফল. এই ডেটা কেন্দ্রযোগ্যতা ছাড়া একাধিক সাইট থেকে বিভিন্ন ডেটা সঙ্গে প্রশিক্ষণ দ্বারা হ্রাস করা যেতে পারে যেমন ট্রান্সফার শেখার পদ্ধতি , FL একটি পদ্ধতি যা আলাদা ডেটা উৎসগুলিতে এআই মডেলগুলি প্রশিক্ষণ করার জন্য ব্যবহৃত হয়, যাতে ডেটাগুলি তাদের মূল অবস্থানের বাইরে পরিবহন করা বা প্রদর্শিত না হয়. FL অনেক শিল্পের জন্য প্রযোজ্য হলেও, FL সম্প্রতি ইনস্টিটিউশনাল স্বাস্থ্যসেবা গবেষণা জন্য প্রস্তাবিত হয়েছে . 18 19 20 21 22 23 24 25 26 27 28 29 30 27 31 32 33 34 35 36 ফেডারেট লার্নিং ডেটা অনুসরণযোগ্যতা উন্নত এবং অ্যালগরিদম পরিবর্তন এবং প্রভাব মূল্যায়ন সঙ্গে কেন্দ্রীয়ভাবে সংগঠিত পরীক্ষার দ্রুত চালু সমর্থন করে FL এর একটি পদ্ধতি, ক্লায়েন্ট-সার্ভার নামে পরিচিত, অন্য সার্ভারগুলিতে ("নোড") একটি "নোড" মডেল পাঠায় যা আংশিক প্রশিক্ষণ কাজ পরিচালনা করে এবং ফলাফলগুলি কেন্দ্রীয় ("ফেডারেটেড") সার্ভারে একত্রিত করার জন্য ফিরে পাঠায়। . 37 36 FL এর জন্য ডেটা গভর্নরতা স্থানীয়ভাবে বজায় রাখা হয়, গোপনীয়তা উদ্বেগগুলি হ্রাস করে, শুধুমাত্র মডেল ওজন বা গ্রিডিয়েন্ট সাইটগুলি এবং ফেডারেট সার্ভারগুলির মধ্যে যোগাযোগ করা হয় , FL ইতিমধ্যে সাম্প্রতিক মেডিকেল ইমেজিং অ্যাপ্লিকেশনে প্রতিশ্রুতিবদ্ধতা প্রদর্শন করেছে , , , COVID-19 এর বিশ্লেষণ , , একটি উল্লেখযোগ্য উদাহরণ SARS-COV-2-এর সাথে আক্রান্ত রোগীদের মৃত্যুর পূর্বাভাস মডেল যা ক্লিনিকাল বৈশিষ্ট্যগুলি ব্যবহার করে, যদিও মডেলগুলির সংখ্যা এবং স্কেলে সীমিত। . 38 39 40 41 42 43 8 44 45 46 আমাদের লক্ষ্য ছিল এমন একটি শক্তিশালী, সাধারণকরণযোগ্য মডেল তৈরি করা যা রোগীদের ট্রায়ারিংয়ে সহায়তা করতে পারে। আমরা তত্ত্বাবধান করেছি যে সিডিসি মডেলটি সফলভাবে ফেডারেশন করা যেতে পারে, যেহেতু এটি ক্লিনিকাল অনুশীলনে তুলনামূলকভাবে সাধারণ ডেটা ইনপুট ব্যবহার করে এবং যা অপারেটর-ভিত্তিক রোগীর অবস্থার মূল্যায়ন (যেমন ক্লিনিকাল প্রদর্শন বা রিপোর্ট করা সংক্রমণ) উপর ব্যাপকভাবে নির্ভর করে না। বরং, ল্যাবরেটরির ফলাফল, জীবনী লক্ষণ, একটি ইমেজিং গবেষণা এবং একটি সাধারণভাবে সংগৃহীত গণমাধ্যম (যেমন, বয়স) ব্যবহার করা আমাদের ধারণা ছিল যে এক্সএমটি স্থানীয় মডেলগুলির চেয়ে ভাল কাজ করবে এবং স্বাস্থ্যসেবা সিস্টেমগুলির মধ্যে ভাল জেনারেল করবে। ফলাফল পরীক্ষার মডেল আর্কিটেকচার পরীক্ষার মডেলটি উপরে উল্লেখিত CDS মডেলের উপর ভিত্তি করে। মোট ২০টি বৈশিষ্ট্য (ইএমআর থেকে ১৯টি এবং সিএক্সআর থেকে একটি) মডেলের ইনপুট হিসাবে ব্যবহৃত হয়। ফলাফল (যেমন, ‘মৃত্যু সত্য’) লেবেলগুলি রোগীর অক্সিজেন থেরাপি অনুযায়ী 24 এবং 72 ঘন্টা পরে জরুরি বিভাগের (ইডি) প্রাথমিক অন্তর্ভুক্ত হওয়ার পরে নির্ধারিত করা হয়েছে। . 27 1 রোগীদের ফলাফল ট্যাবলেটগুলি 0, 0.25, 0.50 এবং 0.75 হিসাবে স্থাপন করা হয়েছিল, নির্ভর করে যে রোগীটি পূর্বাভাসের উইন্ডোতে সর্বাধিক তীব্র অক্সিজেন থেরাপি পেয়েছিল। অক্সিজেন থেরাপি ক্যাটাগরিগুলি প্রতিটি রুমে বায়ু (আরএ), কম প্রবাহের অক্সিজেন (এলএফও), উচ্চ প্রবাহের অক্সিজেন (এইচএফও) / অনিনভাসিভ ভ্যালেন্টেশন (এনআইভি) বা যান্ত্রিক ভ্যালেন্টেশন (এমভি) ছিল। ইএমআর বৈশিষ্ট্যগুলির জন্য, কেবলমাত্র ইডিতে সংগৃহীত প্রথম মানগুলি ব্যবহার করা হয়েছিল এবং ডেটা প্রাক-প্রসেসিং অন্তর্ভুক্ত করেছিল ডেটা অন্তর্ভুক্ত করা, অনুপস্থিত মানের অন্তর্ভুক্ত করা এবং শূন্য-মাত্রা এবং ইউনিট ভ্যারান্সে স্বাভাবিক করা। সুতরাং মডেলটি ইএমআর এবং সিএক্সআর বৈশিষ্ট্যগুলি উভয় থেকে তথ্য একত্রিত করে, একটি 34-বৃদ্ধি convolutional নিউরাল নেটওয়ার্ক (ResNet34) ব্যবহার করে একটি সিএক্সআর এবং একটি গভীর এবং ক্রস নেটওয়ার্ক থেকে বৈশিষ্ট্যগুলি একত্রিত করার জন্য ইএমআর বৈশিষ্ট্যগুলি একত্রিত করে। ) মডেলের আউটপুট একটি ঝুঁকিপূর্ণ পয়েন্ট, যা EXAM পয়েন্ট বলা হয়, যা উপরে বর্ণিত লেবেলগুলির সাথে সংশ্লিষ্ট প্রতিটি 24 এবং 72 ঘন্টা পূর্বাভাসের জন্য 0-1 পরিসরে একটি অব্যাহত মান। পদ্ধতি মডেল ফেডারেশন এক্সএএম মডেলটি 16,148 টি ক্ষেত্রে একটি সংগঠন ব্যবহার করে প্রশিক্ষণ দেওয়া হয়েছিল, যা এটি কেবলমাত্র COVID-19 এর জন্য প্রথম FL মডেলগুলির মধ্যে নয়, তবে ক্লিনিকালভাবে প্রাসঙ্গিক আইটিতে একটি খুব বড় এবং মাল্টি-কনটেন্টিনাল উন্নয়ন প্রকল্প (ফিগার। ) সাইটগুলির মধ্যে ডেটা অপসারণের আগে সমন্বিত করা হয়নি এবং বাস্তব জীবনের ক্লিনিকাল কম্পিউটারিক্স পরিস্থিতিগুলির দিকে তাকিয়ে, ডেটা ইনপুটের একটি মৌলিক সমন্বয় লেখকদের দ্বারা পরিচালিত হয়নি। ) ১. বি ১. ডি , বিশ্ব মানচিত্র যা EXAM গবেষণায় অবদান রাখা 20 বিভিন্ন ক্লায়েন্ট সাইট দেখায়। , প্রতিটি প্রতিষ্ঠান বা সাইট দ্বারা অবদান রাখা মামলাগুলির সংখ্যা (ক্লায়েন্ট 1 সর্বাধিক মামলাগুলি অবদান রাখা সাইটের প্রতিনিধিত্ব করে)। , প্রতিটি ক্লায়েন্ট সাইটে ক্যান্সার ইন্টিগ্রেশন বিতরণ। , প্রতিটি ক্লায়েন্ট সাইটে রোগীর বয়স, সর্বনিম্ন এবং সর্বোচ্চ বয়স (আস্টারিস্ক), গড় বয়স (তিরঙ্গন) এবং স্ট্যান্ডার্ড বিপরীত (হাইড্রোজেন্টাল বার) দেখায়। . a b c d 1 আমরা প্রতিটি ক্লায়েন্টের টেস্ট ডেটাতে স্থানীয়ভাবে প্রশিক্ষিত মডেলগুলি বিশ্বব্যাপী FL মডেলের সাথে তুলনা করেছি। « 1 × 10–3, উইলকক্সন স্বাক্ষরিত র ্যাঙ্ক টেস্ট) 16% (যেমন মডেলটি প্রাসঙ্গিক স্থানীয় পরীক্ষা সেটগুলিতে চালানোর সময় গড় AUC দ্বারা সংজ্ঞায়িত: 0.795 থেকে 0.920, অথবা 12.5 শতাংশ পয়েন্ট) (চিত্র। ) এছাড়াও এটি 38% সাধারণকরণযোগ্যতা উন্নতি (যেমন সমস্ত পরীক্ষা সেটগুলিতে মডেলটি চালানোর সময় গড় AUC দ্বারা সংজ্ঞায়িত: 0.667 থেকে 0.920, অথবা 25.3 শতাংশ পয়েন্ট) 24 ঘন্টা অক্সিজেন চিকিত্সা পূর্বাভাসের জন্য সেরা বিশ্বব্যাপী মডেলের মডেলের তুলনায় মডেলগুলি শুধুমাত্র একটি সাইটের নিজস্ব ডেটা উপর প্রশিক্ষণ (ফিগ. ৭২ ঘন্টা অক্সিজেন চিকিত্সার পূর্বাভাসের ফলাফলগুলির জন্য, সেরা গ্লোবাল মডেল প্রশিক্ষণটি স্থানীয়ভাবে প্রশিক্ষিত মডেলগুলির তুলনায় 18% এর গড় কর্মক্ষমতা উন্নত করে, যখন গ্লোবাল মডেলের সাধারণকরণযোগ্যতা গড়ভাবে 34% উন্নত হয়। ) আমাদের ফলাফলের স্থিতিশীলতা বিভিন্ন র্যান্ডম ডেটা ভাগে স্থানীয় এবং FL প্রশিক্ষণ তিনটি রান পুনরাবৃত্তি করে নিশ্চিত করা হয়েছিল। P 2A ২ বি 1 , প্রতিটি ক্লায়েন্টের টেস্টের কর্মক্ষমতা সার্ভারে উপলব্ধ সেরা বিশ্বব্যাপী মডেলের (এফএল) তুলনায় স্থানীয় ডেটা শুধুমাত্র উপর প্রশিক্ষিত মডেলগুলির জন্য 24 ঘন্টা অক্সিজেন প্রক্রিয়াকরণের পূর্বাভাস। , Generalizability (অন্য সাইটের টেস্ট ডেটাতে গড় পারফরম্যান্স, যেমন AUC দ্বারা প্রতিনিধিত্ব করা হয়) একটি ক্লায়েন্টের ডেটা সেট আকারের ফাংশন হিসাবে (কোন ক্ষেত্রে)। ) এবং ক্লায়েন্ট 14 কেবলমাত্র রিআর চিকিত্সা সঙ্গে ক্ষেত্রে ছিল, যার ফলে মূল্যায়ন মিট্রিক (AUC) এই ক্ষেত্রে কোনও ক্ষেত্রে প্রযোজ্য ছিল না ( ক্লায়েন্ট 14 এর জন্য ডেটাও স্থানীয় মডেলগুলিতে গড় জেনারেলযোগ্যতার গণনা থেকে বাদ দেওয়া হয়েছিল। a b 1 পদ্ধতি স্থানীয় মডেলগুলি যা অস্থিতিশীল কোহল্টগুলি ব্যবহার করে প্রশিক্ষণ করা হয়েছিল (উদাহরণস্বরূপ, প্রধানত হালকা রোগের ক্ষেত্রে COVID-19) FL পদ্ধতি থেকে উল্লেখযোগ্যভাবে সুবিধা পেয়েছিল, শুধুমাত্র কয়েকটি ক্ষেত্রে ক্লায়েন্ট সাইটে 16 (অস্থিতিশীল ডেটা সেট), বেশিরভাগ রোগীদের হালকা রোগের তীব্রতা এবং শুধুমাত্র কয়েকটি গুরুতর ক্ষেত্রে। বিস্তৃত ডাটা ফিগার আরও গুরুত্বপূর্ণভাবে, FL মডেলের সাধারণকরণযোগ্যতা স্থানীয়ভাবে প্রশিক্ষিত মডেলের তুলনায় উল্লেখযোগ্যভাবে বৃদ্ধি পেয়েছে। ৩A 2 , ক্লায়েন্ট সাইট 16 এ ROC, অসামঞ্জস্যপূর্ণ ডেটা এবং বেশিরভাগ হালকা ক্ষেত্রে। ক্লায়েন্ট সাইটে স্থানীয় মডেলের ROC 12 (ছোট ডেটা সেট), বাম, মাঝামাঝি, ডান স্কোরের জন্য 72- ঘন্টা অক্সিজেন চিকিত্সা পূর্বাভাসের জন্য সেরা বিশ্বব্যাপী মডেলের ROC, বেসরকারি পাঁচটি গ্রাহক সাইটে বেসরকারি ডেটা সেটগুলির উপর প্রশিক্ষিত মডেলের গড় ROC। ) of the EXAM risk score are shown. Pos and neg denote the number of positive and negative cases, respectively, as defined by this range of EXAM score. a b t In the case of client sites with relatively small datasets, the best FL model markedly outperformed not only the local model but also those trained on larger datasets from five client sites in the Boston area of the USA (Fig. ). 3b গ্লোবাল মডেলটি 24/72 ঘন্টায় COVID ইতিবাচক এবং নেতিবাচক রোগীদের অক্সিজেন চাহিদা পূর্বাভাসে ভালভাবে কাজ করেছিল (বর্ধিত ডেটা চিত্র। ). 3 Validation at independent sites Following initial training, EXAM was subsequently tested at three independent validation sites: Cooley Dickinson Hospital (CDH), Martha’s Vineyard Hospital (MVH) and Nantucket Cottage Hospital (NCH), all in Massachusetts, USA. The model was not retrained at these sites and it was used only for validation purposes. The cohort size and model inference results are summarized in Table , and the ROC curves and confusion matrices for the largest dataset (from CDH) are shown in Fig. . The operating point was set to discriminate between nonmechanical ventilation and mechanical ventilation (MV) treatment (or death). The FL global trained model, EXAM, achieved an average AUC of 0.944 and 0.924 for 24- and 72-h prediction tasks, respectively (Table , যা প্রশিক্ষণ পরীক্ষায় ব্যবহৃত সাইটগুলির মধ্যে গড় কর্মক্ষমতা অতিক্রম করে। 24 ঘন্টায় এমভি চিকিত্সা (অথবা মৃত্যুর) পূর্বাভাসের জন্য, এক্সএনএমএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্সএক্স 2 4 2 , , Performance (ROC) (top) and confusion matrices (bottom) of the EXAM FL model on the CDH dataset for prediction of oxygen requirement at 24 h ( ) and 72 h ( ). ROCs for three different cutoff values ( ) of the EXAM risk score are shown. a b a b t For MV at CDH at 72 h, EXAM had a low false-negative rate of 7.1%. Representative failure cases are presented in Extended Data Fig. , showing two false-negative cases from CDH where one case had many missing EMR data features and the other had a CXR with a motion artifact and some missing EMR features. 4 Use of differential privacy A primary motivation for healthcare institutes to use FL is to preserve the security and privacy of their data, as well as adherence to data compliance measures. For FL, there remains the potential risk of model ‘inversion’ or even the reconstruction of training images from the model gradients themselves . To counter these risks, security-enhancing measures were used to mitigate risk in the event of data ‘interception’ during site-server communication . We experimented with techniques to avoid interception of FL data, and added a security feature that we believe could encourage more institutions to use FL. We thus validated previous findings showing that partial weight sharing, and other differential privacy techniques, can successfully be applied in FL . Through investigation of a partial weight-sharing scheme , , , we showed that models can reach a comparable performance even when only 25% of weight updates are shared (Extended Data Fig. ) 47 48 49 50 50 51 52 5 Discussion This study features a large, real-world healthcare FL study in terms of number of sites and number of data points used. We believe that it provides a powerful proof-of-concept of the feasibility of using FL for fast and collaborative development of needed AI models in healthcare. Our study involved multiple sites across four continents and under the oversight of different regulatory bodies, and thus holds the promise of being provided to different regulated markets in an expedited way. The global FL model, EXAM, proved to be more robust and achieved better results at individual sites than any model trained on only local data. We believe that consistent improvement was achieved owing to a larger, but also a more diverse, dataset, the use of data inputs that can be standardized and avoidance of clinical impressions/reported symptoms. These factors played an important part in increasing the benefits from this FL approach and its impact on performance, generalizability and, ultimately, the model’s usability. For a client site with a relatively small dataset, two typical approaches could be used for fitting a useful model: one is to train locally with its own data, the other is to apply a model trained on a larger dataset. For sites with small datasets, it would have been virtually impossible to build a performant deep learning model using only their local data. The finding, that these two approaches were outperformed on all three prediction tasks by the global FL model, indicates that the benefit for client sites with small datasets arising from participation in FL collaborations is substantial. This is probaby a reflection of FL’s ability to capture more diversity than local training, and to mitigate the bias present in models trained on a homogenous population. An under-represented population or age group in one hospital/region might be highly represented in another region—such as children who might be differentially affected by COVID-19, including disease manifestations in lung imaging . 46 সনাক্তকরণের ফলাফল নিশ্চিত করেছে যে গ্লোবাল মডেলটি শক্তিশালী, যা আমাদের অনুমানকে সমর্থন করে যে এএফএল প্রশিক্ষিত মডেলগুলি স্বাস্থ্যসেবা সিস্টেমগুলিতে জেনারেলযোগ্য। তারা COVID-19 রোগীর যত্নে পূর্বাভাসের অ্যালগরিদমগুলির ব্যবহার এবং মডেল তৈরিতে এবং পরীক্ষায় FL এর ব্যবহারের জন্য একটি বিশ্বাসযোগ্য ক্ষেত্রে সরবরাহ করে। এই গবেষণায় অংশগ্রহণ করে ক্লায়েন্ট সাইটগুলি পরীক্ষার অ্যাক্সেস পেয়েছে, যা কোনও নিয়ন্ত্রক অনুমোদন অথবা ভবিষ্যতে ক্লিনিক্যাল যত্নে প্রবেশের পূর্বে আরও সনাক্ত করা হবে। পাশাপাশি বিভিন্ন জায়গায় যা পরীক্ষার অংশ ছিল না। 53 COVID-19 রোগীদের সিদ্ধান্ত নেওয়ার জন্য 200 টিরও বেশি পূর্বাভাস মডেল প্রকাশিত হয়েছে . Unlike the majority of publications focused on diagnosis of COVID-19 or prediction of mortality, we predicted oxygen requirements that have implications for patient management. We also used cases with unknown SARS-COV-2 status, and so the model could provide input to the physician ahead of receiving a result for PCR with reverse transcription (RT–PCR), making it useful for a real-life clinical setting. The model’s imaging input is used in common practice, in contrast with models that use chest computed tomography, a nonconsensual diagnostic modality. The model’s design was constrained to objective predictors, unlike many published studies that leveraged subjective clinical impressions. The data collected reflect varied incidence rates, and thus the ‘population momentum’ we encountered is more diverse. This implies that the algorithm can be useful in populations with different incidence rates. 19 রোগীর কোহল্ট সনাক্তকরণ এবং ডেটা সমন্বয় গবেষণা এবং ডেটা বিজ্ঞান মধ্যে নতুন বিষয় নয় , but are further complicated, when using FL, given the lack of visibility on other sites’ datasets. Improvements to clinical information systems are needed to streamline data preparation, leading to better leverage of a network of sites participating in FL. This, in conjunction with hyperparameter engineering, can allow algorithms to ‘learn’ more effectively from larger data batches and adapt model parameters to a particular site for further personalization—for example, through further fine-tuning on that site . A system that would allow seamless, close-to real-time model inference and results processing would also be of benefit and would ‘close the loop’ from training to model deployment. 54 39 Because data were not centralized they are not readily accessible. Given that, any future analysis of the results, beyond what was derived and collected, is limited. Similar to other machine learning models, EXAM is limited by the quality of the training data. Institutions interested in deploying this algorithm for clinical care need to understand potential biases in the training. For example, the labels used as ground truth in the training of the EXAM model were derived from 24- and 72-h oxygen consumption in the patient; it is assumed that oxygen delivered to the patient equates the oxygen need. However, in the early phase of the COVID-19 pandemic, many patients were provided high-flow oxygen prophylactically regardless of their oxygen need. Such clinical practice could skew the predictions made by this model. Since our data access was limited, we did not have sufficient available information for the generation of detailed statistics regarding failure causes, post hoc, at most sites. However, we did study failure cases from the largest independent test site, CDH, and were able to generate hypotheses that we can test in the future. For high-performing sites, it seems that most failure cases fall into one of two categories: (1) low quality of input data—for example, missing data or motion artifact in CXR; or (2) out-of-distribution data—for example a very young patient. ভবিষ্যতে, আমরা রোগের অগ্রগতির বিভিন্ন পর্যায়ে ‘সংখ্যক ড্রাইভ’ এর সম্ভাবনাও তদন্ত করতে যাচ্ছি. আমরা বিশ্বাস করি যে, ২০টি সাইটের মধ্যে বৈচিত্র্যের কারণে এই ঝুঁকি হ্রাস করা যেতে পারে। A feature that would enhance these kinds of large-scale collaboration is the ability to predict the contribution of each client site towards improving the global FL model. This will help in client site selection, and in prioritization of data acquisition and annotation efforts. The latter is especially important given the high costs and difficult logistics of these large-consortia endeavors, and it will enable these endeavors to capture diversity rather than the sheer quantity of data samples. Future approaches may incorporate automated hyperparameter searching নিউরাল আর্কিটেকচার অনুসন্ধান and other automated machine learning approaches to find the optimal training parameters for each client site more efficiently. 55 56 57 Known issues of batch normalization (BN) in FL motivated us to fix our base model for image feature extraction to reduce the divergence between unbalanced client sites. Future work might explore different types of normalization techniques to allow the training of AI models in FL more effectively when client data are nonindependent and identically distributed. 58 49 FL সেটিংয়ের মধ্যে গোপনীয়তা আক্রমণ সম্পর্কে সাম্প্রতিক কাজগুলি মডেল প্রশিক্ষণের সময় ডেটা ছড়ানোর বিষয়ে উদ্বেগ সৃষ্টি করেছে এদিকে, সুরক্ষা অ্যালগরিদমগুলি অজানা এবং একাধিক কারণ দ্বারা সীমাবদ্ধ থাকে, যদিও পার্থক্যযুক্ত গোপনীয়তা অ্যালগরিদমগুলি , , ভাল সুরক্ষা প্রদর্শন করে, তারা মডেলের কর্মক্ষমতা দুর্বল করতে পারে। , maintain performance but may substantially increase message size and training time. A quantifiable way to measure privacy would allow better choices for deciding the minimal privacy parameters necessary while maintaining clinically acceptable performance , , . 59 36 48 49 60 36 48 49 পরবর্তী যাচাইয়ের পরে, আমরা এডি সেটিংয়ের মধ্যে এক্সএএম মডেলের বিতরণকে রোগীর ও জনসংখ্যার স্তরে ঝুঁকি মূল্যায়ন করার উপায় হিসাবে বিবেচনা করি, এবং রোগীদের ট্রায়ারিংয়ের প্রায়ই কঠিন কাজটি করার সময় ক্লিনিক্সদের একটি অতিরিক্ত রেফারেন্স পয়েন্ট সরবরাহ করি। পদ্ধতি নৈতিক অনুমোদন All procedures were conducted in accordance with the principles for human experimentation as defined in the Declaration of Helsinki and International Conference on Harmonization Good Clinical Practice guidelines, and were approved by the relevant institutional review boards at the following validation sites: CDH, MVH, NCH and at the following training sites: MGB, Mass General Hospital (MGH), Brigham and Women’s Hospital, Newton-Wellesley Hospital, North Shore Medical Center and Faulkner Hospital (all eight of these hospitals were covered under MGB’s ethics board reference, no. 2020P002673, and informed consent was waived by the instititional review board (IRB). Similarly, participation of the remaining sites was approved by their respective relevant institutional review processes: Children’s National Hospital in Washington, DC (no. 00014310, IRB certified exempt); NIHR Cambridge Biomedical Research Centre (no. 20/SW/0140, informed consent waived); The Self-Defense Forces Central Hospital in Tokyo (no. 02-014, informed consent waived); National Taiwan University MeDA Lab and MAHC and Taiwan National Health Insurance Administration (no. 202108026 W, informed consent waived); Tri-Service General Hospital in Taiwan (no. B202105136, informed consent waived); Kyungpook National University Hospital in South Korea (no. KNUH 2020-05-022, informed consent waived); Faculty of Medicine, Chulalongkorn University in Thailand (nos. 490/63, 291/63, informed consent waived); Diagnosticos da America SA in Brazil (no. 26118819.3.0000.5505, informed consent waived); University of California, San Francisco (no. 20-30447, informed consent waived); VA San Diego (no. H200086, IRB certified exempt); University of Toronto (no. 20-0162-C, informed consent waived); National Institutes of Health in Bethesda, Maryland (no. 12-CC-0075, informed consent waived); University of Wisconsin-Madison School of Medicine and Public Health (no. 2016-0418, informed consent waived); Memorial Sloan Kettering Cancer Center in New York (no. 20-194, informed consent waived); and Mount Sinai Health System in New York (no. IRB-20-03271, informed consent waived). MI-CLAIM guidelines for reporting of clinical AI models were followed (Supplementary Note ) 2 সেটিং গবেষণা The study included data from 20 institutions (Fig. ): MGB, MGH, Brigham and Women’s Hospital, Newton-Wellesley Hospital, North Shore Medical Center and Faulkner Hospital; Children’s National Hospital in Washington, DC; NIHR Cambridge Biomedical Research Centre; The Self-Defense Forces Central Hospital in Tokyo; National Taiwan University MeDA Lab and MAHC and Taiwan National Health Insurance Administration; Tri-Service General Hospital in Taiwan; Kyungpook National University Hospital in South Korea; Faculty of Medicine, Chulalongkorn University in Thailand; Diagnosticos da America SA in Brazil; University of California, San Francisco; VA San Diego; University of Toronto; National Institutes of Health in Bethesda, Maryland; University of Wisconsin-Madison School of Medicine and Public Health; Memorial Sloan Kettering Cancer Center in New York; and Mount Sinai Health System in New York. Institutions were recruited between March and May 2020. Dataset curation started in June 2020 and the final data cohort was added in September 2020. Between August and October 2020, 140 independent FL runs were conducted to develop the EXAM model and, by the end of October 2020, EXAM was made public on NVIDIA NGC , , তিনটি স্বাধীন সাইটের ডেটা স্বাধীন যাচাইয়ের জন্য ব্যবহার করা হয়েছিল: CDH, MVH এবং NCH, মার্কিন যুক্তরাষ্ট্রের ম্যাসাউজেটস, এই তিনটি হাসপাতালে প্রশিক্ষণ সাইটের তুলনায় রোগীর জনসংখ্যার বৈশিষ্ট্যগুলি আলাদা ছিল। ১A 61 62 63 ডাটা সংগ্রহ The 20 client sites prepared a total of 16,148 cases (both positive and negative) for the purposes of training, validation and testing of the model (Fig. ) চিকিত্সা ডেটা অ্যাক্সেস করা হয়েছিল যারা গবেষণার অন্তর্ভুক্তির শর্তাবলী পূরণ করে। ক্লায়েন্ট সাইটগুলি ২০১৯ সালের ডিসেম্বরে পানডেমির শুরু থেকে এবং যতক্ষণ না তারা এক্সএএম গবেষণার জন্য স্থানীয় প্রশিক্ষণ শুরু করে ততক্ষণ পর্যন্ত সমস্ত COVID-প্রতিক্রিয়াশীল রোগীদের অন্তর্ভুক্ত করার চেষ্টা করেছিল। সমস্ত স্থানীয় প্রশিক্ষণ ৩০ সেপ্টেম্বর 2020 পর্যন্ত শুরু হয়েছিল। ১ বি একটি ‘কেস’ একটি CXR এবং রোগীর মেডিকেল রেকর্ড থেকে নেওয়া প্রয়োজনীয় ডেটা ইনপুট অন্তর্ভুক্ত করে। . The distribution and patterns of CXR image intensity (pixel values) varied greatly among sites owing to a multitude of patient- and site-specific factors, such as different device manufacturers and imaging protocols, as shown in Fig. . Patient age and EMR feature distribution varied greatly among sites, as expected owing to the differing demographics between globally distributed hospitals (Extended Data Fig. ). ১ বি ১. ডি 6 Patient inclusion criteria Patient inclusion criteria were: (1) patient presented to the hospital’s ED or equivalent; (2) patient had a RT–PCR test performed at any time between presentation to the ED and discharge from the hospital; (3) patient had a CXR in the ED; and (4) patient’s record had at least five of the EMR values detailed in Table লক্ষ্য, CXR, ল্যাবরেটরি ফলাফল এবং Vitales ব্যবহৃত প্রথম ED পরিদর্শন সময় ক্যাপচার জন্য উপলব্ধ ছিল. মডেল কোন CXR, ল্যাবরেটরি ফলাফল বা Vitales ED ছেড়ে যাওয়ার পর অন্তর্ভুক্ত ছিল না. 1 ইনপুট মডেল মোট 21 EMR বৈশিষ্ট্যগুলি মডেলের ইনপুট হিসাবে ব্যবহৃত হয়েছিল. ফলাফল (যেমন, জমির সত্য) লেবেলগুলি রোগীর প্রয়োজনীয়তাগুলির উপর ভিত্তি করে 24 এবং 72 ঘন্টা পরে ইডিতে প্রাথমিক অন্তর্ভুক্ত হওয়ার পরে নির্ধারণ করা হয়েছিল. অনুরোধ করা EMR বৈশিষ্ট্য এবং ফলাফলগুলির একটি বিস্তারিত তালিকা টেবিলটিতে দেখতে পাওয়া যায় . 1 The distribution of oxygen treatment using different devices at different client sites is shown in Extended Data Fig. ডাটা সেট বিতরণের মধ্যে পার্থক্য বৃহত্তম এবং সবচেয়ে ছোট ক্লায়েন্ট সাইটগুলির মধ্যে দেখা যায় Extended Data Fig. . 7 8 COVID-19 এর ইতিবাচক রোগের সংখ্যা, যেমনটি একটি একক RT-PCR পরীক্ষা দ্বারা নিশ্চিত করা হয় যা ED প্রদানের এবং হাসপাতাল থেকে বেরিয়ে আসার মধ্যে যে কোনও সময়ে পাওয়া যায়, এটি অতিরিক্ত টেবিলে তালিকাভুক্ত করা হয়। . Each client site was asked to randomly split its dataset into three parts: 70% for training, 10% for validation and 20% for testing. For both 24- and 72-h outcome prediction models, random splits for each of the three repeated local and FL training and evaluation experiments were independently generated. 1 EXAM model development There is wide variation in the clinical course of patients who present to hospital with symptoms of COVID-19, with some experiencing rapid deterioration in respiratory function requiring different interventions to prevent or mitigate hypoxemia , . A critical decision made during the evaluation of a patient at the initial point of care, or in the ED, is whether the patient is likely to require more invasive or resource-limited countermeasures or interventions (such as MV or monoclonal antibodies), and should therefore receive a scarce but effective therapy, a therapy with a narrow risk–benefit ratio due to side effects or a higher level of care, such as admittance to the intensive care unit বিপরীতভাবে, একটি রোগী যিনি আক্রমণাত্মক অক্সিজেন থেরাপি প্রয়োজন কম ঝুঁকির মধ্যে আছে একটি নিয়মিত পার্ক হিসাবে একটি কম তীব্র যত্ন সেটিংস স্থাপন করা যেতে পারে, বা এমনকি বাড়িতে চলমান স্ব-নিরীক্ষা জন্য এডি থেকে মুক্ত করা যেতে পারে . EXAM was developed to help triage such patients. 62 63 64 65 Of note, the model is not approved by any regulatory agency at this time and it should be used only for research purposes. পরীক্ষার ফলাফল EXAM FL ব্যবহার করে প্রশিক্ষণ দেওয়া হয়েছিল; এটি একটি ঝুঁকিপূর্ণ স্কোর (কৃত্রিম EXAM স্কোর) উত্পাদন করে যা CORISK এর মত। (Extended Data Fig. ) এবং রোগীদের সার্টিফিকেশন করার জন্য একইভাবে ব্যবহার করা যেতে পারে. এটি একটি রোগীর অক্সিজেন সমর্থন প্রয়োজনীয়তা দুই উইন্ডো - 24 এবং 72 ঘন্টা - এর পরে প্রাথমিক উপস্থাপনা ED. বিস্তৃত ডেটা ফিগার. কিভাবে CORISK এবং EXAM স্কোর রোগীর সারিতে ব্যবহার করা যেতে পারে তা ব্যাখ্যা করে। 27 9a ৯বি ব্রেস্ট জেনারেটর ইমেজগুলি সামনের অবস্থান ইমেজটি নির্বাচন করার জন্য প্রাক প্রক্রিয়াকরণ করা হয়েছিল এবং পার্শ্ববর্তী দৃশ্য ইমেজগুলি বাদ দেয়া হয়েছিল, এবং তারপর 224 × 224 এর একটি রেজোলিউশন পর্যন্ত স্ক্যালার করা হয়েছিল। , মডেলটি ইএমআর এবং সিএক্সআর বৈশিষ্ট্য উভয় থেকে তথ্য সংমিশ্রণ করে (সম্পাদিত ResNet34 এর উপর ভিত্তি করে স্থানীয় মনোযোগ CheXpert ডেটা সেট (CheXpert Dataset) and the Deep & Cross network এই বিভিন্ন ডেটা টাইপগুলি সংমিশ্রণ করার জন্য, প্রতিটি CXR ইমেজ থেকে একটি 512-দৈর্ঘ্য বৈশিষ্ট্য ভেক্টর প্রাক ট্রেনিং ResNet34 ব্যবহার করে, স্থানীয় মনোযোগ দিয়ে, তারপর গভীর এবং ক্রস নেটওয়ার্কের জন্য ইনপুট হিসাবে EMR বৈশিষ্ট্যগুলির সাথে সংযুক্ত করা হয়েছিল। . We used cross-entropy as the loss function and ‘Adam’ as the optimizer. The model was implemented in Tensorflow using the NVIDIA Clara Train SDK . শ্রেণীকরণ কাজগুলির জন্য গড় AUC (≥LFO, ≥HFO/NIV বা ≥MV) হিসাবে গণনা করা হয়েছিল এবং চূড়ান্ত মূল্যায়ন মিট্রিক হিসাবে ব্যবহার করা হয়েছিল, যাতে নূন্য-মাত্রা এবং ইউনিট ভ্যারানেন্সে স্বাভাবিক করা হয়েছিল। ) 9a 66 67 68 9b 69 70 27 Feature imputation and normalization A MissForest algorithm স্থানীয় প্রশিক্ষণ ডাটা সেটের উপর ভিত্তি করে EMR বৈশিষ্ট্যগুলি অনুমান করার জন্য ব্যবহৃত হয়েছিল. যদি একটি ক্লায়েন্ট সাইট ডাটা সেট থেকে একটি EMR বৈশিষ্ট্য সম্পূর্ণরূপে অনুপস্থিত হতো, তাহলে MGB ক্লায়েন্ট সাইটগুলির ডেটা থেকে কেবলমাত্র হিসাব করা সেই বৈশিষ্ট্যের গড় মান ব্যবহার করা হয়েছিল. তারপর, EMR বৈশিষ্ট্যগুলি MGB ক্লায়েন্ট সাইটগুলির ডেটা থেকে হিসাব করা পরিসংখ্যানের উপর ভিত্তি করে শূন্য গড় এবং ইউনিট ভারেন্সে পুনর্নির্মাণ করা হয়েছিল। 71 Details of EMR–CXR data fusion using the Deep & Cross network To model the interactions of features from EMR and CXR data at the case level, a deep-feature scheme was used based on a Deep & Cross network architecture . ইএমআর ইনপুটগুলির জন্য বাইনারি এবং বিভাগীয় বৈশিষ্ট্যগুলি, পাশাপাশি সিএক্সআরতে 512-দৈর্ঘ্য ইমেজ বৈশিষ্ট্যগুলি, অন্তর্ভুক্ত এবং স্ট্যাকিং স্তরগুলির মাধ্যমে বাস্তব মূল্যগুলির সংমিশ্রিত ঘন ভেক্টরগুলিতে রূপান্তরিত করা হয়েছিল। রূপান্তরিত ঘন ভেক্টরগুলি ফিউশন ফ্রেমওয়ার্কে ইনপুট হিসাবে কাজ করেছিল, যা নির্দিষ্টভাবে বিভিন্ন উত্স থেকে ইনপুটগুলির মধ্যে ফিউশন প্রয়োগ করার জন্য একটি ক্রসিং নেটওয়ার্ক ব্যবহার করেছিল। ক্রসিং নেটওয়ার্কটি তার স্তরগুলির মধ্যে স্পষ্ট বৈশিষ্ট্য ক্রসিং করেছিল, মূল ইনপ 68 বিস্তারিত Arguably the most established form of FL is implemention of the federated averaging algorithm as proposed by McMahan et al. এই অ্যালগরিদমটি একটি ক্লায়েন্ট-সার্ভার সেটিং ব্যবহার করে বাস্তবায়ন করা যেতে পারে যেখানে প্রতিটি অংশগ্রহণকারী সাইট একটি ক্লায়েন্ট হিসাবে কাজ করে। এএফএলকে একটি পদ্ধতি হিসাবে চিন্তা করা যেতে পারে যা একটি গ্লোবাল হার্ট ফাংশনকে কমাতে লক্ষ্য করে যা প্রতিটি সাইটে অনুমান করা হয়। প্রতিটি ক্লায়েন্ট সাইটে স্থানীয় হার্টকে সর্বনিম্ন করে দেয়া এবং একটি কেন্দ্রীয় সংমিশ্রণ সার্ভারে শিক্ষিত ক্লায়েন্ট সাইটে ওজনগুলি সমন্বয় করে, প্রতিটি ক্লায়েন্ট সাইটে স্থানীয়ভাবে একটি সেট ডাটা সেট অ্যাক্সেসের প্রয়োজন ছাড়াই গ্লোবাল হার্ট কমাতে পারে। প্রতিটি ). 72 9c এফএলএল এর একটি পেসডো অ্যালগরিদম অতিরিক্ত নোটে প্রদর্শিত হয় আমাদের পরীক্ষায়, আমরা ফেডারেট রুটগুলির সংখ্যাটি সেট করেছি = 200, প্রতি রাউন্ডে একটি স্থানীয় প্রশিক্ষণ সময় প্রতিটি গ্রাহকের সংখ্যা, , ক্লায়েন্টের নেটওয়ার্ক সংযোগের উপর নির্ভর করে বা নির্দিষ্ট টার্গেট ফলাফল সময়ের জন্য উপলব্ধ ডেটা (24 বা 72 ঘন্টা) পর্যন্ত ছিল। , প্রতিটি ক্লায়েন্টের ডাটা সেটের আকারের উপর নির্ভর করে এবং প্রতিটি ক্লায়েন্টের অবদানকে ফেডারেট মাত্রায় মডেল ওজন সংমিশ্রণ করার সময় ব্যবহার করা হয়। FL প্রশিক্ষণ কাজের সময়, প্রতিটি ক্লায়েন্ট সাইট তার স্থানীয় যাচাইকরণ সেটের উপর মডেলের কর্মক্ষমতা ট্র্যাকিং দ্বারা তার সেরা স্থানীয় মডেল নির্বাচন করে। 1 T t K nk k অ্যাডাম অপ্টিমাইজারটি স্থানীয় প্রশিক্ষণ এবং এলএলএল উভয় ক্ষেত্রে ব্যবহৃত হয়েছিল 5 × 10-5 এর প্রাথমিক শিখার হার এবং প্রতি 40 যুগের পরে 0.5 ফ্যাক্টর দিয়ে ধাপে ধাপে শিখার হার হ্রাসের সাথে, যা ফেডারেট গড় সংমিশ্রণের জন্য গুরুত্বপূর্ণ। র ্যান্ডম আফিন রূপান্তর, রুট, অনুবাদ, কাটা, স্ক্যালিং এবং র ্যান্ডম তীব্রতা শব্দ এবং বিনিময় সহ, প্রশিক্ষণের সময় ডেটা বৃদ্ধির জন্য ইমেজগুলিতে প্রয়োগ করা হয়েছিল। 73 BN layers এর সংবেদনশীলতার কারণে when dealing with different clients in a nonindependent and identically distributed setting, we found the best model performance occurred when keeping the pretrained ResNet34 with spatial attention ফ্লি প্রশিক্ষণের সময় স্থিতিশীল প্যারামিটার (অর্থাৎ, এই স্তরগুলির জন্য একটি শিখার হার ব্যবহার করে)। গভীর ও ক্রস নেটওয়ার্ক যা ইমেজ বৈশিষ্ট্যগুলি ইএমআর বৈশিষ্ট্যগুলির সাথে সংযুক্ত করে, BN স্তরগুলি অন্তর্ভুক্ত করে না এবং অতএব BN অস্থিতিশীলতা সমস্যাগুলির দ্বারা প্রভাবিত হয়নি। 58 47 In this study we investigated a privacy-preserving scheme that shares only partial model updates between server and client sites. The weight updates were ranked during each iteration by magnitude of contribution, and only a certain percentage of the largest weight updates was shared with the server. To be exact, weight updates (also known as gradients) were shared only if their absolute value was above a certain percentile threshold, (গ) বিস্তৃত ডাটা ফিগার। ) , যা সমস্ত অ-নুলো গ্রিডিয়েন্ট থেকে গণনা করা হয়েছিল, Δ প্রতিটি গ্রাহকের জন্য এটি আলাদা হতে পারে। in each FL round এই সিস্টেমের বিকল্পগুলি বড় gradients বা diferential privacy schemes এর অতিরিক্ত ক্লিপিং অন্তর্ভুক্ত করতে পারে। that add random noise to the gradients, or even to the raw data, before feeding into the network . k 5 জিপিএ(টি) k t 49 51 Statistical analysis আমরা স্থানীয়ভাবে প্রশিক্ষিত মডেল এবং 24 এবং 72 ঘন্টা সময় পয়েন্টের জন্য FL মডেলের মধ্যে কর্মক্ষমতা উন্নতির পর্যবেক্ষণ করা উন্নতির গুরুত্ব নিশ্চিত করার জন্য একটি Wilcoxon স্বাক্ষরিত র ্যাংক পরীক্ষা পরিচালনা করেছি। বিস্তৃত ডাটা ফিগার (১) ন্যূনতম হাইপোথেসি একদমই প্রত্যাখ্যান করা হয়েছে। « 1 × 10–3 in both cases. 2 1 P ফরেক্স ডেভেলপমেন্ট সিস্টেম (একটি ডেভেলপমেন্ট ডেভেলপমেন্ট সিস্টেম) ব্যবহার করে ডেভেলপমেন্ট ডেভেলপমেন্ট সিস্টেম (একটি ডেভেলপমেন্ট ডেভেলপমেন্ট সিস্টেম) ব্যবহার করে ডেভেলপমেন্ট ডেভেলপমেন্ট সিস্টেম ( = 0.43, = 0.035, স্বাধীনতা ডিগ্রি (df) = 17 24 ঘন্টা মডেল এবং = 0.62, = 0.003, df = 16 72-h মডেলের জন্য)। r P r P To compare ROC curves from the global FL model and local models trained at different sites (Extended Data Fig. ), we bootstrapped 1,000 samples from the data and computed the resulting AUCs. We then calculated the difference between the two series and standardized using the formula = (AUC1 – AUC2) কোথায় নমনীয় পার্থক্য হল, এটি bootstrap পার্থক্যগুলির স্ট্যান্ডার্ড বিপরীততা এবং AUC1 এবং AUC2 হল সংশ্লিষ্ট bootstrapped AUC সিরিজ। with normal distribution, we obtained the অতিরিক্ত টেবিলে চিত্রিত মূল্য ফলাফলগুলি দেখায় যে নূন্য হাইপোথেশনটি খুবই নিম্ন পর্যায়ে প্রত্যাখ্যান করা হয়েছিল। পরিসংখ্যান অনুযায়ী এফএল ফলাফলের অগ্রাধিকারের পরিসংখ্যানগত গুরুত্ব প্রদর্শন করে। মূল্যগুলি pROC লাইব্রেরিতে R এ পরিচালিত হয়েছিল . 3 D s D s D P 2 P P 74 যেহেতু মডেলটি একটি আলাদা ফলাফল পূর্বাভাস করে, 0 থেকে 1 এর মধ্যে একটি ক্রমাগত স্কোর, একটি qqplot মত সরাসরি ক্যালিব্রেশন মূল্যায়ন সম্ভব নয়। ) আমরা চারটি ভূগর্ভস্থ সত্য শ্রেণীর (আরএ, এলএফও, এইচএফও, এমভি) মধ্যে স্থানীয় এবং এএফএল মডেল স্কোরগুলি তুলনা করার জন্য একটি দিকনির্দেশিত বিশ্লেষণ বিশ্লেষণ (ANOVA) পরীক্ষা পরিচালনা করেছি। -statistic, calculated as the variation between the sample means divided by variation within the samples and representing the degree of dispersion among different groups, was used to quantify the models. Our results show that the -পাঁচটি ভিন্ন স্থানীয় সাইটের মান 245.7, 253.4, 342.3, 389.8 এবং 634.8, যখন এফএল মডেলের মান 843.5। -values মানে যে গ্রুপগুলি আরও বিভক্ত করা যায়, আমাদের FL মডেল থেকে পয়েন্ট স্পষ্টভাবে চারটি ভিত্তিক সত্য বিভাগগুলির মধ্যে বৃহত্তর বিচ্ছিন্নতা দেখায়। value of the ANOVA test on the FL model is <2 × 10–16, indicating that the FL prediction scores are statistically significantly different among the different prediction classes. 10 F F F P রিপোর্ট সংক্ষিপ্ত Further information on research design is available in the linked to this article. Nature Research Reporting Summary Data availability The dataset from the 20 institutes that participated in this study remains under their custody. These data were used for training at each of the local sites and were not shared with any of the other participating institutions or with the federated server, and they are not publicly available. Data from the independent validation sites are maintained by CAMCA, and access can be requested by contacting Q.L. Based on determination by CAMCA, a data-sharing review and amendment of IRB for research purposes can be conducted by MGB research administration and in accordance with MGB IRB and policy. Code availability এই গবেষণায় ব্যবহৃত সমস্ত কোড এবং সফটওয়্যার NGC এ পাবলিকভাবে উপলব্ধ। অ্যাক্সেস করতে, অতিথি হিসাবে লগইন করতে বা একটি প্রোফাইল তৈরি করতে, নিম্নলিখিত ইউআরএলগুলির মধ্যে একটি প্রবেশ করুন. প্রশিক্ষিত মডেল, ডেটা প্রস্তুতি নির্দেশিকা, প্রশিক্ষণ কোড, মডেলের যাচাইকরণ পরীক্ষা, readme ফাইল, ইনস্টলেশন নির্দেশিকা এবং লাইসেন্স ফাইল NVIDIA NGC এ পাবলিকভাবে উপলব্ধ : ফেডারেট শেখার সফ্টওয়্যারটি ক্লারার ট্রেন এসডিক এর অংশ হিসাবে উপলব্ধ: বিকল্পভাবে, এই কমান্ডটি ব্যবহার করে "wget --content-disposition" মডেলটি ডাউনলোড করুন -একটি সূচক আছে, যার মধ্যে একটি হল: "O clara_train_covid19_exam_ehr_xray_1.zip". 61 https://ngc.nvidia.com/catalog/models/nvidia:med:clara_train_covid19_exam_ehr_xray https://ngc.nvidia.com/catalog/containers/nvidia:clara-train-sdk https://api.ngc.nvidia.com/v2/models/nvidia/med/clara_train_covid19_exam_ehr_xray/versions/1/zip রেফারেন্স Budd, J. et al. Digital technologies in the public-health response to COVID-19. , 1183–1192 (2020). Nat. Med. 26 Moorthy, V., Henao Restrepo, A. M., Preziosi, M.-P. & Swaminathan, S. Data sharing for novel coronavirus (COVID-19). , 150 (2020). Bull. World Health Organ. 98 Chen, Q., Allot, A. & Lu, Z. Keep up with the latest coronavirus research. , 193 (2020). Nature 579 Fabbri, F., Bhatia, A., Mayer, A., Schlotter, B. & Kaiser, J. BCG IT spend pulse: how COVID-19 is shifting tech priorities. (2020). https://www.bcg.com/publications/2020/how-covid-19-is-shifting-big-it-spend Candelon, F., Reichert, T., Duranton, S., di Carlo, R. C. & De Bondt, M. The rise of the AI-powered company in the postcrisis world. (2020). https://www.bcg.com/en-gb/publications/2020/business-applications-artificial-intelligence-post-covid Chao, H. et al. Integrative analysis for COVID-19 patient outcome prediction. , 101844 (2021). Med. Image Anal. 67 Zhu, X. et al. Joint prediction and time estimation of COVID-19 developing severe symptoms using chest CT scan. , 101824 (2021). Med. Image Anal. 67 Yang, D. et al. Federated semi-supervised learning for Covid region segmentation in chest ct using multi-national data from China, Italy, Japan. , 101992 (2021). Med. Image Anal. 70 Minaee, S., Kafieh, R., Sonka, M., Yazdani, S. & Jamalipour Soufi, G. Deep-COVID: predicting COVID-19 from chest X-ray images using deep transfer learning. , 101794 (2020). Med. Image Anal. 65 COVID-19 Studies from the World Health Organization Database. (2020). https://clinicaltrials.gov/ct2/who_table ACTIV. (2020). https://www.nih.gov/research-training/medical-research-initiatives/activ Coronavirus Treatment Acceleration Program (CTAP). US Food and Drug Administration (2020). https://www.fda.gov/drugs/coronavirus-covid-19-drugs/coronavirus-treatment-acceleration-program-ctap Gleeson, P., Davison, A. P., Silver, R. A. & Ascoli, G. A. A commitment to open source in neuroscience. , 964–965 (2017). Neuron 96 Piwowar, H. et al. The state of OA: a large-scale analysis of the prevalence and impact of open access articles. , e4375 (2018). PeerJ. 6 European Society of Radiology (ESR). What the radiologist should know about artificial intelligence – an ESR white paper. , 44 (2019). Insights Imaging 10 Pesapane, F., Codari, M. & Sardanelli, F. Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine. , 35 (2018). Eur. Radiol. Exp. 2 Price, W. N. 2nd & Cohen, I. G. Privacy in the age of medical big data. , 37–43 (2019). Nat. Med. 25 Liang, W. et al. Development and validation of a clinical risk score to predict the occurrence of critical illness in hospitalized patients with COVID-19. , 1081–1089 (2020). JAMA Intern. Med. 180 Wynants, L. et al. Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisal. , m1328 (2020). Brit. Med. J. 369 Zhang, L. et al. D-dimer levels on admission to predict in-hospital mortality in patients with Covid-19. , 1324–1329 (2020). J. Thromb. Haemost. 18 Sands, K. E. et al. Patient characteristics and admitting vital signs associated with coronavirus disease 2019 (COVID-19)-related mortality among patients admitted with noncritical illness. (2020). https://doi.org/10.1017/ice.2020.461 American College of Radiology. CR recommendations for the use of chest radiography and computed tomography (CT) for suspected COVID-19 infection. (2020). https://www.acr.org/Advocacy-and-Economics/ACR-Position-Statements/Recommendations-for-Chest-Radiography-and-CT-for-Suspected-COVID19-Infection Rubin, G. D. et al. The role of chest imaging in patient management during the COVID-19 pandemic: a multinational consensus statement from the Fleischner Society. , 172–180 (2020). Radiology 296 World Health Organization. Use of chest imaging in COVID-19. (2020). https://www.who.int/publications/i/item/use-of-chest-imaging-in-covid-19 Jamil, S. et al. Diagnosis and management of COVID-19 disease. , 10 (2020). Am. J. Respir. Crit. Care Med. 201 Redmond, C. E., Nicolaou, S., Berger, F. H., Sheikh, A. M. & Patlas, M. N. Emergency radiology during the COVID-19 pandemic: The Canadian Association of Radiologists Recommendations for Practice. , 425–430 (2020). Can. Assoc. Radiologists J. 71 Buch, V. et al. Development and validation of a deep learning model for prediction of severe outcomes in suspected COVID-19 Infection. Preprint at (2021). https://arxiv.org/abs/2103.11269 Lyons, C. & Callaghan, M. The use of high-flow nasal oxygen in COVID-19. , 843–847 (2020). Anaesthesia 75 Whittle, J. S., Pavlov, I., Sacchetti, A. D., Atwood, C. & Rosenberg, M. S. Respiratory support for adult patients with COVID-19. , 95–101 (2020). J. Am. Coll. Emerg. Physicians Open 1 Ai, J., Li, Y., Zhou, X. & Zhang, W. COVID-19: treating and managing severe cases. , 370–371 (2020). Cell Res. 30 Esteva, A. et al. A guide to deep learning in healthcare. , 24–29 (2019). Nat. Med. 25 Cahan, E. M., Hernandez-Boussard, T., Thadaney-Israni, S. & Rubin, D. L. Putting the data before the algorithm in big data addressing personalized healthcare. , 78 (2019). NPJ Digit. Med. 2 Thrall, J. H. et al. Artificial intelligence and machine learning in radiology: opportunities, challenges, pitfalls, and criteria for success. , 504–508 (2018). J. Am. Coll. Radiol. 15 Shilo, S., Rossman, H. & Segal, E. Axes of a revolution: challenges and promises of big data in healthcare. , 29–38 (2020). Nat. Med. 26 Gao, Y. & Cui, Y. Deep transfer learning for reducing health care disparities arising from biomedical data inequality. , 5131 (2020). Nat. Commun. 11 Rieke, N. et al. The future of digital health with federated learning. , 119 (2020). NPJ Dig. Med. 3 Yang, Q., Liu, Y., Chen, T. & Tong, Y. Federated machine learning: concept and applications. , 12 (2019). ACM Trans. Intell. Syst. Technol. 10 Ma, C. et al. On safeguarding privacy and security in the framework of federated learning. , 242–248 (2020). IEEE Netw. 34 Brisimi, T. S. et al. Federated learning of predictive models from federated Electronic Health Records. , 59–67 (2018). Int. J. Med. Inform. 112 Roth, H. R. et al. Federated learning for breast density classification: a real-world implementation. In , (eds. Albarqouni, S. et al.) Vol. 12,444, 181–191 (Springer International Publishing, 2020). Proc. Second MICCAI Workshop, DART 2020 and First MICCAI Workshop, DCL 2020 Domain Adaptation and Representation Transfer, and Distributed and Collaborative Learning Sheller, M. J. et al. Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data. , 12598 (2020). Sci. Rep. 10 Remedios, S. W., Butman, J. A., Landman, B. A. & Pham, D. L. in (eds Remedios, S. W. et al.) (Springer, 2020). Federated Gradient Averaging for Multi-Site Training with Momentum-Based Optimizers Xu, Y. et al. A collaborative online AI engine for CT-based COVID-19 diagnosis. Preprint at (2020). https://www.medrxiv.org/content/10.1101/2020.05.10.20096073v2 Raisaro, J. L. et al. SCOR: A secure international informatics infrastructure to investigate COVID-19. , 1721–1726 (2020). J. Am. Med. Inform. Assoc. 27 Vaid, A. et al. Federated learning of electronic health records to improve mortality prediction in hospitalized patients with COVID-19: machine learning approach. , e24207 (2021). JMIR Med. Inform. 9 Nino, G. et al. Pediatric lung imaging features of COVID-19: a systematic review and meta-analysis. , 252–263 (2021). Pediatr. Pulmonol. 56 Fredrikson, M., Jha, S. & Ristenpart, T. Model inversion attacks that exploit confidence information and basic countermeasures. In 1322–1333, (2015). Proc. 22nd ACM SIGSAC Conference on Computer and Communications Security https://doi.org/10.1145/2810103.2813677 Zhu, L., Liu, Z. & Han, S. in (eds Wallach, H. et al.) 14774–14784 (Curran Associates, Inc., 2019). Advances in Neural Information Processing Systems 32 Kaissis, G. A., Makowski, M. R., Rückert, D. & Braren, R. F. Secure, privacy-preserving and federated machine learning in medical imaging. , 305–311 (2020). Nat. Mach. Intell. 2 Li, W. et al. in 133–141 (Springer, 2019). Privacy-Preserving Federated Brain Tumour Segmentation Shokri, R. & Shmatikov, V. Privacy-preserving deep learning. In (2015). Proc. 53rd Annual Allerton Conference on Communication, Control, and Computing (Allerton) https://doi.org/10.1109/allerton.2015.7447103 Li, X. et al. Multi-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results. , 101765 (2020). Med. Image Anal. 65 Estiri, H. et al. Predicting COVID-19 mortality with electronic medical records. , 15 (2021). NPJ Dig. Med. 4 Jiang, G. et al. Harmonization of detailed clinical models with clinical study data standards. , 65–74 (2015). Methods Inf. Med. 54 Yang, D. et al. in . (2019). Searching Learning Strategy with Reinforcement Learning for 3D Medical Image Segmentation https://doi.org/10.1007/978-3-030-32245-8_1 Elsken, T., Metzen, J. H. & Hutter, F. Neural architecture search: a survey. , 1–21 (2019). J. Mach. Learning Res. 20 Yao, Q. et al. Taking human out of learning applications: a survey on automated machine learning. Preprint at (2019). https://arxiv.org/abs/1810.13306 Ioffe, S. & Szegedy, C. Batch normalization: accelerating deep network training by reducing internal covariate shift. In , PMLR , 448–456 (2015). Proc. 32nd International Conf. Machine Learning 37 Kaufman, S., Rosset, S. & Perlich, C. Leakage in data mining: formulation, detection, and avoidance. In , 556–563 (2011). Proc. 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining Zhang, C. et al. BatchCrypt: efficient homomorphic encryption for cross-silo federated learning. In , 493–506 (2020). Proc. 2020 USENIX Annual Technical Conference, ATC 2020 . (2020). Nvidia NGC Catalog: COVID-19 Related Models https://ngc.nvidia.com/catalog/models?orderBy=scoreDESC&pageNumber=0&query=covid&quickFilter=models&filters Marini, J. J. & Gattinoni, L. Management of COVID-19 respiratory distress. , 2329–2330 (2020). JAMA 323 Cook, T. M. et al. Consensus guidelines for managing the airway in patients with COVID-19: Guidelines from the Difficult Airway Society, the Association of Anaesthetists the Intensive Care Society, the Faculty of Intensive Care Medicine and the Royal College of Anaesthetist. , 785–799 (2020). Anaesthesia 75 Galloway, J. B. et al. A clinical risk score to identify patients with COVID-19 at high risk of critical care admission or death: an observational cohort study. , 282–288 (2020). J. Infect. 81 Kilaru, A. S. et al. Return hospital admissions among 1419 COVID-19 patients discharged from five U.S. emergency departments. , 1039–1042 (2020). Acad. Emerg. Med. 27 He, K., Zhang, X., Ren, S. & Sun, J. Deep residual learning for image recognition. In (2016). Proc. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) https://doi.org/10.1109/cvpr.2016.90 Irvin, J. et al. CheXpert: a large chest radiograph dataset with uncertainty labels and expert comparison. , 590–597 (2019). Proc. AAAI Conf. Artif. Intell. 33 Wang, R., Fu, B., Fu, G. & Wang, M. Deep & Cross network for Ad Click predictions. In Article no. 12 (2017). Proc. ADKDD’17 Abadi, M. et al. TensorFlow: asystem for large-scale machine learning. In , USENIX Association 265–283 (2016). 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16) . (2020). NVIDIA Clara Imaging https://developer.nvidia.com/clara-medical-imaging Stekhoven, D. J. & Bühlmann, P. MissForest–non-parametric missing value imputation for mixed-type data. , 112–118 (2012). Bioinformatics 28 McMahan, H., Moore, E., Ramage, D., Hampson, S. & y Arcas, B. A. Communication-efficient learning of deep networks from decentralized data. (2017). http://proceedings.mlr.press/v54/mcmahan17a.html Hsieh, K., Phanishayee, A., Mutlu, O. & Gibbons, P. B. The non-IID data quagmire of decentralized machine learning. In PMLR 119 (2020). Proc. 37th International Conf. Machine Learning Robin, X. et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. , 77 (2011). BMC Bioinformatics 12 স্বীকৃতি এই গবেষণায় প্রকাশ করা গ্রাহক ডেটা বিভাগের মাধ্যমে প্রকাশ করা গ্রাহকদের মতামতগুলি এই গবেষণায় অ্যাক্সেস করা হয় এবং অবশ্যই NHS, NIHR, স্বাস্থ্য ও সামাজিক যত্নের বিভাগ বা গ্রাহকদের সাথে সংযুক্ত কোনও সংস্থার মতামতগুলি নয়। MGB তাদের সমর্থনের জন্য নিম্নলিখিত ব্যক্তিদের ধন্যবাদ জানান: জে. ব্রিনক, রেডিওলজি বিভাগ, ম্যাসাজুসেট জেনারেল হাসপাতাল, হার্ভার্ড মেডিকেল স্কুল, বস্টন, এমএ; এম. গুও, রেডিওলজি বিভাগ, ম্যাসাজুসেট জেনারেল হাসপাতাল, হার্ভার্ড জেনারেল মেডিকেল স্কুল, হার্ভার্ড মেডিক মেডিকেল ফাংশন, চুলালংকর্ন বিশ্ববিদ্যালয়ের মাধ্যমে মেডিকেল ফাংশন, রাতচ্যাপাক্সেক Sompoch Endowment Fund RA (PO) (নম্বর 001/63) এর জন্য COVID-19-র সাথে সম্পর্কিত ক্লিনিকাল ডেটা এবং জৈবিক নমুনা সংগ্রহ এবং ব্যবস্থাপনা করার জন্য ধন্যবাদ। NIHR ক্যামব্রিজ বায়োমেডিক্যাল গবেষণা কেন্দ্রকে ধন্যবাদ, যা NIHR (ক্যামব্রিজ বিশ্ববিদ্যালয়ের হাসপাতালগুলির NHS ফাউন্ডেশন ফাউন্ডেশনের ক্যামব্রিজ বায়োমেডিক্যাল গবেষণা কেন্দ্র) দ্বারা সমর্থিত। জাতীয় তাইওয়ান বিশ্ববিদ্যালয় MeDA ল্যাব এবং MAHC এবং তাইওয়ান জাতীয় স্বাস্থ্য নিরাপত্তা প্রশাসন https://data.ucsf.edu/covid19 এই কাগজটি CC by 4.0 Deed (Attribution 4.0 International) লাইসেন্সের অধীনে প্রাকৃতিকভাবে পাওয়া যায়। এই কাগজটি CC by 4.0 Deed (Attribution 4.0 International) লাইসেন্সের অধীনে প্রাকৃতিকভাবে পাওয়া যায়।