新加坡科技設(shè)計(jì)大學(xué)林媚霞博士學(xué)術(shù)報(bào)告

報(bào)告人：林媚霞博士，新加坡科技設(shè)計(jì)大學(xué)工程系統(tǒng)與設(shè)計(jì)

報(bào)告題目：Exploring intrinsic structured sparsity in convex composite programming

報(bào)告時(shí)間：2022年11月25日，星期五，下午：14:00-15:00

報(bào)告地點(diǎn)：騰訊會(huì)議號(hào)：249 843 168; 密碼：1125

報(bào)告摘要：

Convex optimization models have been widely used in many applications such as machine learning and data science. However, the huge computation for the involved potentially large-scale problems has prevented their deployments in resource-limited devices. In our work, we design efficient second-order algorithms for the structured convex composite programming problems, which fully exploit the structure of the data and the underlying Hessians to highly reduce the computational cost. Dimension reduction techniques are also designed to further accelerate the computation, especially for the high-dimensional cases.

報(bào)告人簡介：

       林媚霞，新加坡科技設(shè)計(jì)大學(xué)助理教授。2020年在新加坡國立大學(xué)數(shù)學(xué)系取得博士學(xué)位，2016年在南京大學(xué)取得信息計(jì)算與科學(xué)學(xué)士學(xué)位。主要研究興趣為開發(fā)與設(shè)計(jì)大數(shù)據(jù)科學(xué)中的模型與算法，特別是高效求解機(jī)器學(xué)習(xí)，統(tǒng)計(jì)估計(jì)和運(yùn)籌學(xué)中涉及的超大規(guī)模優(yōu)化問題。以第一或通訊作者在高水平期刊和會(huì)議上發(fā)表多篇文章，包括SIAM Journal on Optimization, Mathematical Programming Computation, IEEE Transactions on Signal Processing及人工智能權(quán)威會(huì)議NIPS。