Live in the future, then build what's missing!

2014working weibo点滴记录6 agent future

tony /
分类 | 生活 
标签 | weibo 

2014-8-6

Whoever Owns The Agent Owns The Future

PC -> SEARCH -> PHONE -> JARVIS

2014-8-8

dnn放到移动终端来做,大有可为。关注移动终端异构计算的deep learning吧。 //@phunter_lau:想起来这个哥们还做了个拍几张照片就能实时认出来自己家猫在哪里的demo http://petewarden.com/2014/04/08/how-to-add-a-brain-to-your-smart-phone/ iPhone SDK和app都能下载,还是很萌的,20m内存就行 @djvu9 libccv的作者写了篇怎样把ImageNet上训练的CNN分类器在iPhone上跑起来 http://libccv.org/post/how-to-run-a-state-of-the-art-image-classifier-on-a-iphone/ 大致功能是拍个照片能认出来里面是什么。做法是减小网络规模,把权值聚类,然后用png压缩。模型的大小大概是原来的10%左右。这种模型大概有几千万的参数吧。让我想起了《钢铁苍穹》里的那台手机。。

这个很好,希望大家在Word and Paraphase Representation 这个领域继续深入,多提供些开源工具。 @刘知远THU 斯坦福Richard Socher在EMNLP2014发表新作:GloVe: Global Vectors for Word Representation 粗看是融合LSA等算法的想法,利用global word co-occurrence信息提升word vector学习效果,很有意思,在word analogy task上准确率比word2vec提升了11%。 http://www.socher.org/index.php/Main/GloveGlobalVectorsForWordRepresentation

http://petewarden.com/2014/04/08/how-to-add-a-brain-to-your-smart-phone/ petewarden.com How to add a brain to your smart phone Pete Warden做了不少有趣的事情呢!

Facebook Is Hiring To Make Linux Networking Better Than FreeBSD. Apple要在freebsd上多给予支持啊,智能硬件时代也少不了freebsd。 @zolker Facebook Is Hiring To Make Linux Networking Better Than FreeBSD.Our goal over the next few years is for the Linux kernel network stack to rival or exceed that of FreeBSD.Facebook招Linux内核开发,目标几年内在网络协议栈赶超FreeBSD[哈哈]

How to optimize Raspberry Pi code using its GPU Why engineers shouldn’t disdain frontend development What does the future hold for deep learning? Why is everyone so excited about deep learning? Deep learning on the Raspberry Pi! @tony的微博 O网页链接 petewarden.com How to add a brain to your smart phone Pete Warden做了不少有趣的事情呢!

Ever tried. Ever failed. No matter. Try Again. Fail again. Fail better. @tony的微博 O网页链接 petewarden.com How to add a brain to your smart phone Pete Warden做了不少有趣的事情呢!

https://github.com/jetpacapp/DeepBeliefSDK The SDK for Jetpac's iOS Deep Belief image recognition framework http://libccv.org/ A Modern Computer Vision Library how to run a state of the art image classifier on a iPhone @tony的微博 O网页链接 petewarden.com How to add a brain to your smart phone Pete Warden做了不少有趣的事情呢!

nbic的时代终于来临了,赶紧开始搞robotics。 //@阳志平1949:时间真快。。。2011年到现在,短短三年,如今突破如期而来。看来我的预测是对的,目前的确是nbic技术站在风口上了,以神经科技、人工智能为主导的技术,将引领下一个时代变革。 @安人心智 【IBM认知计算项目取得重大突破! 】仿脑芯片已拥有超级计算机能力,相对2011年只能模拟256个神经元来说,目前已能模拟1m个神经元、256m个突触!并且低功耗,以每秒30帧播放400*240像素的视频,仅消耗63毫瓦!最新Science论文请点击:http://www.sciencemag.org/content/345/6197/668 IBM官方介绍请点击:http://www.research.ibm.com/cognitive-computing/neurosynaptic-chips.shtml#fbid=9lLuBo8aGco

http://www.idsia.ch/ The Swiss AI Lab IDSIA We focus on machine learning (artificial neural networks, reinforcement learning), optimal universal artificial intelligence and optimal rational agents, operations research, complexity theory, and robotics.

Google acquires for ~$500m the company deepmind co-founded by our former PhD student Shane Legg (PDF) and Demis Hassabis and Mustafa Suleyman. Deepmind also employs Prof. J. Schmidhuber's former SNF-funded PhD students Dr. Daan Wierstra, Dr. Tom Schaul, Dr. Alex Graves. @tony的微博 O网页链接 The Swiss AI Lab IDSIA We focus on machine learning (artificial neural networks, reinforcement learning), optimal universal artificial intelligence and optimal rational agents, operations research, complexity theory, and robotics.

http://people.idsia.ch/~juergen/deeplearning.html Jürgen Schmidhuber 比赛的成绩太多了,还搞中文手写识别呢。 //@tony的微博:Google acquires for ~$500m the company deepmind co-founded by our former PhD student Shane Legg (PDF) and Demis Hassabis and Mustafa Suleyman. Deepmind also empl @tony的微博 O网页链接 The Swiss AI Lab IDSIA We focus on machine learning (artificial neural networks, reinforcement learning), optimal universal artificial intelligence and optimal rational agents, operations research, complexity theory, and robotics.

http://people.idsia.ch/~ciresan/index.htm idsia确实有不少像Dan Claudiu Cireșan这种牛人。 //@tony的微博:O网页链接 Jürgen Schmidhuber 比赛的成绩太多了,还搞中文手写识别呢。 //@tony的微博:Google acquires for ~$500m the company deepmind co-founded by our former PhD student Sha @tony的微博 O网页链接 The Swiss AI Lab IDSIA We focus on machine learning (artificial neural networks, reinforcement learning), optimal universal artificial intelligence and optimal rational agents, operations research, complexity theory, and robotics.

2014-8-10

赞!!!改变命运看心气、格局、智力。提高心气,文养气,诗洗心;提高格局,一本一本好书一寸一寸拓展开;提高智力,神经智力+经验智力+反省智力。 @阳志平1949 【穷小子怎么改变命运?】@非言语 说:情商可自保,但做事得有真本事。这句话极赞。所谓情商,国人常理解为狡猾、算计、操控。靠这些,最多自保,乱世就是炮灰。改变命运看心气、格局、智力。提高心气,文养气,诗洗心;提高格局,一本一本好书一寸一寸拓展开;提高智力,神经智力+经验智力+反省智力。

http://carlos.bueno.org/2012/08/writers-reading-list.html carlos.bueno.org A Computer Science Writer's Reading List The Structure of Scientific Revolutions

GoogLeNet的效果,支持Yann LeCun吧! @fs_Yangqing 我们在Google的队伍今年刷新了ImageNet challenge的新纪录:在classification上6.6%的错误率比去年降低44%,在detection上43.9%的检测率比去年提高49%。具体技术将在ECCV奉上:)

2014-8-20

MIT Technology Review评选出了2014年的TR35,Quoc Le也上榜了。 @王威廉 MIT Technology Review评选出了2014年的TR35得主: http://www.technologyreview.com/lists/innovators-under-35/2014/ CMU机器学习系新任助理教授,深度学习专家Quoc Le榜上有名;另外著名众包语言学习网站Duolingo的共同创始人,CMU计算机系博士Severin Hacker也榜上有名。

2014-8-22

Suchanek和Weikum要在VLDB做tutorial,讲turn web into knowledge base,很赞!!! @波多野丽猪 过几天Suchanek和Weikum在VLDB做tutorial,有机会的一定要去听,国内去的最好帮忙录一下。他俩之前在sigmod的tutorial就非常给力,思路清晰,把data->knowledge的脉络讲全了,slides请戳http://resources.mpi-inf.mpg.de/yago-naga/sigmod2013-tutorial/sigmod2013-slides.pdf

未来十年值得关注的两个人工智能领域:以实体为基础的知识图谱,以及智能的对话系统。rl-based agents and dl-based knowledge. //@王威廉:见谷歌大神Ni Lao和Kevin Murphy等人的KDD 2014神作Knowledge Vault: http://www.cs.cmu.edu/~nlao/publication/2014.kdd.pdf //@我爱机器学习:#以实体为基础的知识图谱#这个有推荐的参考资料吗? @王威廉 机器学习大神,谷歌前副总裁,CMU计算机学院院长Andrew Moore今天给我们系学生研讨会做报告,介绍了他认为未来十年值得关注的两个人工智能领域:以实体为基础的知识图谱,以及智能的对话系统。另外,Andrew觉得算法及系统方面的研究也会一直是重点。

2014-8-23

Systems Performance: Enterprise and the Cloud DTrace: Dynamic Tracing in Oracle Solaris, Mac OS X and FreeBSD Solaris Performance and Tools: DTrace and MDB Techniques for Solaris 10 and OpenSolaris //@老师木:一直想修炼这本书的技能,若成,则超越jeff dean 指日可待 //@fleeting001: / @网路冷眼 【Linux性能资源大全】http://www.brendangregg.com/linuxperf.html 收集了有关Linux观测、基准测试和调优工具,文档,演讲资料(Slides和视频)等资源。里面的工具图清晰地标明对不同对象和工具的映射关系,大赞!!!感兴趣请转发并拿走。@何_登成 @孢子响马 @Linux中国 @developerWorks @极客头条

2014-8-28

Cache Line、Cache Size、False Sharing,持续优化! @何_登成 如果说理解一个技术的最高境界,就是能够用最简单的方式将这个技术表达出来的话,那么Igor对于CPU Cache的理解绝对达到了此境界。他的博文:Gallery of Processor Cache Effects http://igoro.com/archive/gallery-of-processor-cache-effects/ 7个简单至极的代码示例,覆盖了Cache Line、Cache Size、False Sharing等重要知识点,不得不服[赞]

2014-8-29

赶紧升级ddr4吧! 单条64GB或128GB, 2133Mbps。 @快科技-原驱动之家 【单条64GB!全球首个3D DDR4内存量产】三星宣布,已经开始批量投产全球第一款采用3D TSV立体硅穿孔封装技术的DDR4内存条,单条容量达64GB,32颗颗粒,每颗封装四个1Gb芯片,2xnm工艺制造。不得不说,三星共和国的技术还是很牛X的,之前还量产了第二代3D NAND闪存,密度也是最高。O网页链接

SK海力士 DDR4 DRAM和NAND Flash。 @集邦DRAMeXchange 【SK海力士冲NAND Flash产能、重庆厂下月量产】SK海力士22日表示,重庆厂将封装测试韩国清州厂生产的NAND Flash芯片,预定下个月开始量产。SK海力士将加值产品如DDR4 DRAM和NAND Flash,大量转往中国无锡厂生产,外包给韩国供货商的数量骤减。

2014-8-30

Skip lists、Bloom filters、 Dancing links、Tries,很强大。 @网路冷眼 【StackOverflow的神贴:那些少人所知而又有用的数据结构】每个程序猿都熟悉链表、二叉树和哈希表等数据结构,可是Skip lists、Bloom filters、 Dancing links、Tries、Rops...又知道几个呢?涨姿势了!建议果断收藏放入工具箱,以备不时之需。http://stackoverflow.com/questions/500607/what-are-the-lesser-known-but-useful-data-structures @蒋涛CSDN @刘江总编 @西祠响马

unix和c的哲学和文化影响会一直持续下去。//Chris Dixon认为:每条原始的Unix命令,都会变成一项互联网服务。Grep->Google, rsync->Dropbox, man->stack overflow, cron->ifttt @王威廉 Chris Dixon认为:每条原始的Unix命令,都会变成一项互联网服务。Grep->Google, rsync->Dropbox, man->stack overflow, cron->ifttt...... 要创业的同学,选一条命令快上? [懒得理你]


前一页     后一页