全部

From Model Optimization to Interpretable and Collaborative Deep Learning

  • 演讲者:刘日升(大连理工大学)

  • 时间:2019-03-20 16:15-17:15

  • 地点:慧园3栋 415报告厅

报告摘要:Model optimization plays the key role in many learning and vision tasks. However, designing numerical schemes always need high mathematical skills and rich domain knowledge. Moreover, how to apply the generally designed iterations in specific real-world scenario is always a challenging problem. In this talk, we introduce a series of paradigms to design task-specific optimization schemes based on (inexact) learnable architectures. The theoretical properties of these deeply trained propagations are carefully investigated. We demonstrate that we actually provide a new way to establish interpretable and collaborative deep learning models for different real-world applications. Some insights (e.g., the comparison to adversarial mechanism in GAN) will also be covered.

讲者信息:刘日升,大连理工大学计算数学博士,香港理工大学计算科学博士后。目前任职于大连理工大学国际信息与软件学院副教授、博士生导师,数字媒体技术系主任、几何计算与智能媒体研究所副所长。近年来在TPAMI、TNNLS、TIP、TMM等期刊和CVPR、NIPS、IJCAI、AAAI、ACM MM、ECCV、CIKM、ICDM、ACCV等会议发表论文70余篇。相关工作被引用超过1800次,H-index为15,最高单篇引用超过560次。获得ICME 2014和2015年最佳学生论文奖, VALSE 2018最受关注论文奖, ICME 2017最佳论文Finalist(两篇,Top 3%),ICIP 2015最佳10%论文,ICIMCS 2017最佳论文提名,IEEE智能计算亮点论文(Publication Spotlight)等。获得教育部自然科学二等奖1项、辽宁省自然科学二等奖1项。入选辽宁省青年拔尖人才、“香江学者”、大连市“青年科技之星”、ACM新星奖(大连Chapter)、大连理工大学“星海优青”。担任The Visual Computer、IET Image Processing和Journal of Electronic Imaging编委,AAAI 2019 高级程序委员(SPC)等。