講座名稱：基于機(jī)器學(xué)習(xí)的特定應(yīng)用NoC的有效設(shè)計(jì)空間探索

講座時(shí)間：2019-06-22 10:00:00

講座地點(diǎn)：西電北校區(qū)東大樓221學(xué)術(shù)報(bào)告廳

講座人：Ulf Schlichtmann

講座人介紹：

Ulf Schlichtmann，博士，教授，德國慕尼黑工業(yè)大學(xué)全職講席教授，慕尼黑工業(yè)大學(xué)EDA實(shí)驗(yàn)中心主任，曾/現(xiàn)擔(dān)任德國英飛凌科技部門主管以及最高管理層成員，德國國家科學(xué)與工程院成員，新加坡慕尼黑工業(yè)大學(xué)研究中心主任。2007年至2013年擔(dān)任慕尼黑工業(yè)大學(xué)電子與計(jì)算機(jī)工程學(xué)院副院長和院長，2016年擔(dān)任慕尼黑工業(yè)大學(xué)董事會(huì)成員和教授任命委員會(huì)委員。Schlichtmann教授發(fā)表了250多篇頂級(jí)期刊和會(huì)議論文，獲得了多項(xiàng)會(huì)議的最佳論文提名和最佳論文獎(jiǎng)，包括Design, Automation and Test in Europe (DATE) 2017最佳論文獎(jiǎng)。

講座內(nèi)容：

In this talk, I will give a brief overview of the ongoing research topics in the Chair of Electronic Design Automation (EDA) at Technical University of Munich (TUM), covering analysis and optimization of electronic circuits and systems as well as emerging technologies such as biochips and photonic interconnect. Thereafter, our research on Multi-Processor Systems-on-Chip (MPSoCs) with an increasing number of heterogeneous processing elements will be presented in detail. Since the traffic requirements on the Network on Chip (NoC) become more and more unbalanced, the NoC network architecture must be customized according to the needs of the application at hand. In this talk, I will introduce two approaches for the acceleration of design space exploration (DSE) through machine learning techniques.

主辦單位：微電子學(xué)院