Chenfiona的个人博客分享 http://blog.sciencenet.cn/u/Chenfiona

博文

【重磅推荐】MIT Tomaso Poggio关于深度学习机理的研究

已有 483 次阅读 2017-3-16 16:31 |个人分类:好文推荐|系统分类:论文交流|关键词:深度学习,类脑,认知,综述,机器学习,机器人

上周,在IJAC优先在线发表的几篇论文中,麻省理工美国人文与科学院院士Tomaso Poggio的一篇有关深度学习的综述成为一大亮点。Poggio教授在这篇文章中阐述了有关神经网络深度学习的一个基本问题:为什么深层网络比浅层网络更好?



    文章内容延续了Poggio教授在2016年8月在中国人工智能大会(2016CCAI)上的演讲报告《The Science and the Engineering ofIntelligence》。

图1 来源于2016 CCAI 演讲PPT


图2 来源于2016 CCAI 演讲PPT


  深度学习架构和机器学习模式的搭建,来自于神经学方面的研究进展,换句话说,同样的架构是存在于大脑皮质当中的。

  关于深度学习,已经有成千上万的研究者在不同领域进行这方面的研究,比如无人驾驶、语音识别等等。可是我们还不清楚,为什么深度学习在这些工程应用中会起作用,深度学习的机理是什么?

   我们对这个问题很感兴趣另外一个原因是:探讨深度学习的机理也将有助于我们理解‘为什么大脑皮质会存在一些不同的层次?’

Poggio教授在这篇文章中,将为您解读深度学习的关键理论、最新成果和开放式研究问题。

同时这篇文章也是IJAC即将发表的Special Issue on Human Inspired Computing中的一篇文章。该专题其他热文将陆续优先在线发表,敬请期待。


一点点题外话:小编在去年的CCAI大会上有幸拜访了Poggio教授,教授博学、谦逊而富有亲和力的形象给小编也留下深刻印象。他曾提到:期望能帮助年轻人更好的了解神经科学、理解机器学习。如果要在智能方面走得远,不能只靠计算机,还需要与人类本身的研究相互结合,才能碰撞出更多的东西。


接下来,小编将为您奉上这篇文章的具体信息,以及IJAC近期在线发表的其他文章链接,欢迎下载阅读


TitleWhy and when can deep-but not shallow-networksavoid the curse of dimensionality: A review

AuthorTomaso Poggio, Hrushikesh Mhaskar, LorenzoRosasco, Brando Miranda, Qianli Liao

AbstractThe paper reviews and extends an emerging bodyof theoretical results on deep learning including the conditions under which itcan be exponentially better than shallow learning. A class of deepconvolutional networks represent an important special case of these conditions,though weight sharing is not the main reason for their exponential advantage.Implications of a few key theorems are discussed, together with new results,open problems and conjectures.

KeywordsMachine learning, neural networks, deep andshallow networks, convolutional neural networks, function approximation, deeplearning

Full Texthttps://link.springer.com/article/10.1007/s11633-017-1054-2

Publish datePublished online March 14, 2017


The other recentlypublished online papers include:


TitleImprovement of wired drill pipe data qualityvia data validation and reconciliatio

AuthorDan Sui, Olha Sukhoboka, Bernt Sigve Aadnøy

KeywordsData quality, wired drill pipe (WDP), datavalidation and reconciliation (DVR), drilling models

Full Texthttps://link.springer.com/article/10.1007/s11633-017-1068-9

Publish datePublished online March 4, 2017


TitleReaction torque control of redundantfree-floating space robot

AuthorMing-He Jin, Cheng Zhou, Ye-Chao Liu, Zi-QiLiu, Hong Liu

KeywordsRedundant space robot, reaction torque,reactionless control, base disturbance minimization, Linux/real timeapplication interface (RTAI)

Full Texthttps://link.springer.com/article/10.1007/s11633-017-1067-x

Publish datePublished online March 4, 2017


TitleA piecewise switched linear approach fortraffic flow modeling

AuthorAbdelhafid Zeroual,Nadhir Messai, SihemKechida, Fatiha Hamdi

KeywordsSwitched systems, modeling, macroscopic,traffic flow, data calibration

Full Texthttps://link.springer.com/article/10.1007/s11633-017-1060-4

Publish datePublished online March 4, 2017


TitleNavigation of non-holonomic mobile robot usingneuro-fuzzy logic with integrated safe boundary algorithm

AuthorA. Mallikarjuna Rao, K. Ramji, B. S. K.Sundara Siva Rao, V. Vasu, C. Puneeth

KeywordsRobotics, autonomous mobile robot (AMR),navigation, fuzzy logic, neural networks, adaptive neuro-fuzzy inference system(ANFIS), safe boundary algorithm

Full Texthttps://link.springer.com/article/10.1007/s11633-016-1042-y

Publish datePublished online March 4, 2017

TitleMethod for visual localization of oil and gaswellhead based on distance function of projected features

AuthorYing Xie, Xiang-Dong Yang, Zhi Liu, Shu-NanRen, Ken Chen

KeywordsRobot vision, visual localization, 3D objectlocalization, model based pose estimation, distance function of projectedfeatures, nonlinear least squares, random sample consensus (RANSAC)

Full Texthttps://link.springer.com/article/10.1007/s11633-017-1063-1

Publish datePublished online March 4, 2017


TitleVirtual plate based controlling strategy oftoy play for robots communication development in JA space

AuthorWei Wang, Xiao-Dan Huang

KeywordsHuman robot cooperation, joint attention (JA)space, reachable space, toy play ability, a virtual plate

Full Texthttps://link.springer.com/article/10.1007/s11633-016-1022-2

Publish datePublished online February 21, 2017


阅读更多IJAC优先在线出版论文:https://link.springer.com/journal/11633




http://blog.sciencenet.cn/blog-749317-1039818.html

上一篇:【2017年学术会议资讯(下)】-自动控制类&机器人
下一篇:【IJAC推文】疯狂存储数据的你,是否了解高存储介质背后的尖端技

0

该博文允许注册用户评论 请点击登录 评论 (0 个评论)

数据加载中...

Archiver|科学网 ( 京ICP备14006957 )

GMT+8, 2017-5-30 19:09

Powered by ScienceNet.cn

Copyright © 2007-2017 中国科学报社