GAMES Webinar 2019 – 91期(深度学习可解释性专题课程) | 林宙辰（北京大学），Shixia Liu（清华大学）

【GAMES Webinar 2019-91期】
报告嘉宾：林宙辰，北京大学

报告时间：2019年4月18日 晚8:00-8:45（北京时间）
主持人：雷娜，大连理工大学（个人主页：http://faculty.dlut.edu.cn/leina/zh_CN/index.htm）
报告题目：优化与深度神经网络

报告摘要：

优化是机器学习的不可或缺的组成部分。在这个报告里，我将展示优化不仅能用于训练深度神经网络，还可以用于网络设计。报告基于最近发表的AAAI 2019和ACML 2018两篇论文。在AAAI 2019论文中，我们提出了提升近邻算子机（LPOM），来把前向神经网络的训练问题近似为一个多凸（Multi-Convex）优化问题，从而可以方便地求解。LPOM具有多个优点：每个变量块容易更新、存储量和随机梯度法相同、可以使用任何Lipschitz连续的非降激活函数、容易调参、可以异步并行等等。LPOM可能会成为目前主流的随机梯度法的替代。在ACML 2018论文中，我们首先证明具有固定权重的深度神经网络等价于用梯度法来最小化某目标函数。接着，基于快的算法对应于好网络的假设，我们用比梯度法快的算法来最小化该目标函数，从而得到相应的深度网络，ResNet和DenseNet都是我们设计框架的特例。

讲者简介：

林宙辰，北京大学信息科学技术学院教授。研究领域包括计算机视觉、图像处理、机器学习、模式识别和数值优化。他是CVPR 2014/2016/2019、ICCV 2015、NIPS 2015/2018/2019和AAAI 2019的领域主席，AAAI 2016/2017/2018和IJCAI 2016/2018/2019的高级程序委员。他是IEEE Transactions on Pattern Analysis and Machine Intelligence 和International Journal of Computer Vision的编委，IAPR和IEEE会士。

讲者个人主页：http://www.cis.pku.edu.cn/faculty/vision/zlin/zlin.htm

报告嘉宾：刘世霞，清华大学

报告时间：2019年4月18日 晚8:45-9:30（北京时间）
主持人：雷娜，大连理工大学（个人主页：http://faculty.dlut.edu.cn/leina/zh_CN/index.htm）
报告题目： Visual Analytics for Explainable Machine Learning

报告摘要：

Machine learning has demonstrated to be highly successful at solving many real-world applications ranging from information retrieval, data mining, and speech recognition, to computer graphics, visualization, and human-computer interaction. However, most users often treat the machine learning model as a “black box” because of its incomprehensible functions and unclear working mechanism. Without a clear understanding of how and why the model works, the development of high-performance models typically relies on a time-consuming trial-and-error procedure. This talk presents the major challenges of explainable machine learning and exemplifies the solutions with several visual analytics techniques and examples, including model understanding and diagnosis.

讲者简介：

Shixia Liu is a tenured associate professor at Tsinghua University. Her research interests include explainable machine learning, visual text analytics, visual social analytics, and text mining. Shixia is an associate editor-in-chief of IEEE Transactions on Visualization and Computer Graphics, an associate editor of IEEE Transactions on Big data and ACM Transactions on Interactive Intelligent Systems. She is one of the Papers Co-Chairs of IEEE VAST 2016 and 2017. She is on the editorial board of Information Visualization and Journal of visualization. She was the program co-chair of PacifcVis 2014 and VINCI 2012. Shixia was in the Steering Committee of VINCI 2013. She is on the organizing committee of IEEE VIS 2019, 2018, 2015 and 2014. She is/was in the Program Committee for CHI 2019, 2018, InfoVis 2015, 2014, VAST 2019, 2018, 2015, 2014, KDD 2015, 2014, 2013, ACM Multimedia 2009, etc.

讲者个人主页：http://cgcad.thss.tsinghua.edu.cn/shixia/

GAMES主页的“使用教程”中有 “如何观看GAMES Webinar直播？”及“如何加入GAMES微信群？”的信息；
GAMES主页的“资源分享”有往届的直播讲座的视频及PPT等。
观看直播的链接：http://webinar.games-cn.org