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

2014reading weibo信息轨迹6 neural science

tony /
分类 | 生活 
标签 | weibo 

2014-5-16

deepmind http://arxiv.org/pdf/1312.5602v1.pdf

Demis Hassabis http://demishassabis.com/publications/

Shane Legg http://www.vetta.org/publications/

http://www.sciencemag.org/content/318/5858/1844.1.full sciencemag.org BREAKTHROUGH OF THE YEAR

2014-5-17

dileep george http://www.dileepgeorge.com/cgi-sys/suspendedpage.cgi

http://redwood.berkeley.edu/wiki/Publications The Redwood Center for Theoretical Neuroscience

http://vicarious.com/about.html vicarious team

http://www.kurzweilai.net/vicarious-ai-breaks-captcha-turing-test kurzweilai.net Vicarious AI breaks CAPTCHA ‘Turing test’

谷歌主页之所以如此简洁,主要是因为公司的两位联合创始人Sergey Brin和Larry Page根本不懂HTML编码技术。在很长的一段时间内,谷歌用户只能通过点击回车键来展开搜索,因为谷歌主页上根本就没有“提交”按钮。正常,不是学不会HTML,而是不屑于学HTML。哈哈!

谷歌通过reCAPTCHA技术让计算机逐渐学会了如何正确读取文本。由于谷歌每天使用到2000亿次CAPTCHA技术,因此谷歌计算机已经学会了该如何辨别图书上的文字,即便是在这些文字已经被扭曲的情况下 这个发明太有创意了。书籍是人类历史的最大财富!

http://idl.baidu.com/en/IDL-team.html baidu idl team

http://lcsl.mit.edu/ldr-workshop/Schedule.html Learning Data Representation: Hierarchies and Invariance The goals: to investigate advances and challenges in learning "good representations" from data, in particular representations that can reduce the complexity of later supervised learning stages.

http://openreview.net/venue/iclr2014 ICLR 2014 ICLR 2013

一切学习本质上都是建立神经网络和功能联结,不管是行为主义、建构主义、信息加工理论、社会认知理论、动机论和自我调节论,只不过神经科学严谨科学有余,而活学活用不够。 赞!

book-2012 Learning Theories An Educational Perspective 6th Edition

http://deeplearning.net/tutorial/rnnrbm.html#rnnrbm Modeling and generating sequences of polyphonic music with the RNN-RBM

http://www.researchgate.net/publication/257072406_Deep_generative_learning_of_location-invariant_visual_word_recognition researchgate.net Deep generative learning of location-invariant visual word recognition.

http://www.researchgate.net/publication/236654800_Deep_Unsupervised_Learning_on_a_Desktop_PC_A_Primer_for_Cognitive_Scientists researchgate.net Deep Unsupervised Learning on a Desktop PC: A Primer for Cognitive Scientists.

2014-5-18

计算认知科学、计算社会科学、计算神经科学的区别?哪个更适合申请?A. 三者理论上区别很大,计算神经科学同义词更等同于神经信息学,偏生理;计算认知科学狭义词可以等同于认知建模(尤其是贝叶斯),偏人工智能;计算社会科学狭义词等于多主体建模,偏社会学。- 阳志平

三个学科的数学入门难度。计算神经科学我了解很少,不予发表意见;计算认知科学的数学基础主要在于贝叶斯、神经网络两者,对它的数学训练基本等同于cs专业的训练,与机器学习类要求与训练一致;计算社会科学的数学基础主要在于图论与网络科学,相对前者,更简单,但该领域已经人才泛滥了。- 阳志平

Tomaso Poggio 拿到25 Million的项目!http://mcgovern.mit.edu/news/news/nsf-awards-25-million-to-establish-center-for-brains-minds-and-machines/ Tommy是机器学习领域的旗帜性人物之一,很少有学者像他那样把统计学习理论,认知学,大脑神经科学等一气呵成,融汇贯通。 You need to have language, you need to see things, you need to have language and vision work together.

http://vislab.bjmu.edu.cn/blog/hwangxin/ vislab.bjmu.edu.cn 让Kindle支持扫描版PDF重排 Koreader把K2pdfopt用的很好啊,看来最好自己也根据需要定制一下,tesseract的效果不太好,pdf转文字一直还有不少问题。

http://www.hi-pda.com/forum/viewthread.php?tid=1078988 hi-pda.com Koreader -- 更新日志,功能建议和Bug反馈 这个做的很用心啊!

http://www.hi-pda.com/forum/viewthread.php?tid=913606&extra=page%3D1%26amp%3Bfilter%3Dtype%26amp%3Btypeid%3D72 hi-pda.com 很好的资源分享论坛

http://tpgit.github.io/UnOfficialLeptDocs/leptonica/index.html leptonica doc

http://www.imagemagick.org/ imagemagick doc

http://ghostscript.com/doc/current/Readme.htm ghostscript doc

K2pdfopt has built-in PDF translation (via the MuPDF library) but will try to use Ghostscript if Ghostscript is available and the internal (MuPDF) translation fails. but you can force Ghostscript to be used with the -gs option.

http://www.mupdf.com/docs/ mupdf doc


前一页     后一页