報(bào) 告 人：朱利平 教授

報(bào)告題目：A Goodness-of-fit Assessment for General Learning Procedure in High Dimensions

報(bào)告時(shí)間：2024年10月27日（周日）上午10：00

報(bào)告地點(diǎn)：靜遠(yuǎn)樓1506學(xué)術(shù)報(bào)告廳

主辦單位：數(shù)學(xué)研究院、數(shù)學(xué)與統(tǒng)計(jì)學(xué)院、科學(xué)技術(shù)研究院

報(bào)告人簡(jiǎn)介：

        朱利平，中國人民大學(xué)教授、博士生導(dǎo)師，學(xué)校和理工學(xué)部學(xué)術(shù)委員會(huì)委員，統(tǒng)計(jì)與大數(shù)據(jù)研究院院長，人民教育出版社普通高中教科書《數(shù)學(xué)》聯(lián)合主編，國家重大人才工程入選者，國家杰出青年科學(xué)基金獲得者，國家重點(diǎn)研發(fā)計(jì)劃首席科學(xué)家，兼任中國現(xiàn)場(chǎng)統(tǒng)計(jì)研究會(huì)生存分析分會(huì)理事長和高維數(shù)據(jù)統(tǒng)計(jì)分會(huì)副理事長等。先后受邀擔(dān)任國際統(tǒng)計(jì)學(xué)領(lǐng)域頂級(jí)學(xué)術(shù)期刊《統(tǒng)計(jì)年刊》、國際權(quán)威學(xué)術(shù)期刊《中華統(tǒng)計(jì)學(xué)》和《多元分析》等副主編，以及國內(nèi)統(tǒng)計(jì)學(xué)領(lǐng)域頂級(jí)學(xué)術(shù)期刊《中國科學(xué)·數(shù)學(xué)》（中、英文版）、《系統(tǒng)科學(xué)與數(shù)學(xué)》（中、英文版）和《應(yīng)用概率統(tǒng)計(jì)》等青年編委、編委和副主編等。

報(bào)告摘要：

        Black-box learners have demonstrated remarkable success across various fields due to their high predictive accuracy. However, the complexity of their learning procedures poses significant challenges in evaluating whether a given learner has achieved optimal performance on datasets with unknown data-generating mecha-nisms. We propose a general goodness-of-fit test for assessing different learning procedures involving high-dimensional predictors, encompassing methods from classical linear regression to advanced neural networks. Our goodness-of-fit test leverages data-splitting, utilizing the test set to evaluate the black-box learner trained on the training set. This evaluation is based on examining the cumulative covariance of the residuals. Extensive simulations and two real data analyses validate the effectiveness of our method.