報告題目:Solving an inverse source problem by deep neural network method with convergence and error analysis
報 告 人:劉繼軍 教授
所在單位:東南大學
報告時間:2022年11月3日 星期四 下午14:00-15:00
報告地點:騰訊會議 ID:939-312-800 密碼:3359
會議鍊接:https://meeting.tencent.com/dm/fIrxzuOBE2Rw
校内聯系人:張德悅 dyzhang@jlu.edu.cn
報告摘要:For the inverse source problem of an elliptic system using noisy internal measurement as inversion input, we approximate its solution by the neural network function, which is yielded by optimizing an empirical loss function with some regularizing terms. We analyze the convergence of the general loss with noisy input in Deep Galerkin Method by the regularizing empirical loss function. Based on the upper bound of the expected loss function by its regularizing empirical form, we establish the upper bound of the expected loss function at the minimizer of regularizing empirical noisy loss function in terms of the number of sampling points well as the noise level quantitatively, for suitably chosen regularizing parameters and regularizing terms. Then, by specifying the number of sampling points in terms of noise level of inversion input data, we establish the error orders representing the difference between the neural network solution and the exact one, under some {\it a-priori} restrictions on the source. Finally, we give numerical implementations for several examples to verify our theoretical results. This is a joint work with Dr. Hui Zhang.
報告人簡介:劉繼軍,男,1965年出生,博士。東南大學二級教授,博士生導師,享受國務院政府特殊津貼專家。現任南京應用數學中心常務副主任,全國大學生數學建模競賽組委會委員,中國工業與應用數學學會數學建模競賽專業委員會委員,江蘇省計算數學學會副理事長。國家精品資源共享課《數學建模與數學實驗》主持人。曆任中國工業與應用數學學會常務理事,中國計算數學學會常務理事,江蘇省工業與應用數學學會第五屆、第六屆理事會理事長。 長期從事數學物理反問題、大規模科學計算和介質成像的數學理論和方法的研究。主持完成國家自然科學基金重大研究計劃培育項目等多項基金項目。已在SIAM J., Inverse Problems等發表學術論文130餘篇,在科學出版社出版學術專著2本。曾受中國NSFC、德國DAAD、韓國21Brain Project等資助赴國外開展合作研究。2012-2017年任Inverse Problems in Sciences and Engineering編委,2018年起任J. Inverse and Ill-posed Problems編委。入選江蘇省青藍工程青年骨幹教師,青藍工程中青年學術帶頭人,江蘇省333工程第三層次培養人選。獲寶鋼教育基金會全國優秀教師一等獎,作為主持人獲江蘇省教學成果一等獎、江蘇省自然科學三等獎、教育部自然科學二等獎。