題目: Learning theory for distributed learning

報(bào)告人：林紹波，bwin必贏唯一官網(wǎng)教授

時(shí)間：2019年11月4日（周一)下午15：00-16：30

地點(diǎn)：管院313會(huì)議室

歡迎廣大師生前來參加！

報(bào)告內(nèi)容: We analyze the performance of distributed learning based on a divide-and-conquer strategy.  This scheme applies a specified learning algorithm to data subsets that are distributively stored on multiple servers to produce individual output functions, and then takes a weighted average of the individual output functions as a final estimator.  We study that under which condition distributed learning is feasible and conduct several strategies such as semi-supervised distributed learning and communications to improve its performance further.

報(bào)告人簡(jiǎn)介：林紹波，bwin必贏唯一官網(wǎng)教授，青年拔尖人才。博士畢業(yè)于bwin必贏唯一官網(wǎng)數(shù)學(xué)與統(tǒng)計(jì)學(xué)院。曾先后于香港城市大學(xué)數(shù)學(xué)系、香港理工大學(xué)應(yīng)用數(shù)學(xué)系、香港城市大學(xué)數(shù)據(jù)科學(xué)學(xué)院擔(dān)任博士后、副研究員、研究員。研究方向?yàn)榇髷?shù)據(jù)分析、分布式計(jì)算及深度學(xué)習(xí)。主持或以核心成員參與國(guó)家自然科學(xué)基金9項(xiàng)。 在Journal of Machine Learning Research, Applied and Computational Harmonics Analysis, IEEE Transactions on Signal Processing, IEEE Transactions on Neural Networks and Learning Systems等著名期刊發(fā)表論文50余篇。