2014-7-11
https://dialrc.org/ The Dialog Research Center DialRC conducts community-wide spoken dialog challenges (SDC). There are a variety of ways to participate in this effort, which is aimed at bringing the community together and providing common data and tasks.
http://mi.eng.cam.ac.uk/research/dialogue/ The Dialogue Systems group which is part of the CUED Speech Group is led by Prof Steve Young . The work of the group centres on the use of statistical approaches to Spoken Dialogue Systems. The aim is to design systems that can be traine @tony的微博 O网页链接 The Dialog Research Center DialRC conducts community-wide spoken dialog challenges (SDC). There are a variety of ways to participate in this effort, which is aimed at bringing the community together and providing common data and tasks.
The potential benefits include reduced deployment costs, more robust operation in adverse environments and the ability to adapt on-line. Recent work has primarily focused on the application of Partially-Observable Markov Decision Processes and machine learning techniques to dialo @tony的微博 O网页链接 The Dialog Research Center DialRC conducts community-wide spoken dialog challenges (SDC). There are a variety of ways to participate in this effort, which is aimed at bringing the community together and providing common data and tasks.
Currently the dialogue systems group is working on the EU funded Probabalistic Adaptive Learning And Natural Conversational Engine (PARLANCE) project. The goal of this project is to to design and build mobile applications that approach human performance in con- versational intera @tony的微博 O网页链接 The Dialog Research Center DialRC conducts community-wide spoken dialog challenges (SDC). There are a variety of ways to participate in this effort, which is aimed at bringing the community together and providing common data and tasks.
http://www.iwsds.org/ INTERNATIONAL WORKSHOP SERIES ON SPOKEN DIALOGUE SYSTEMS TECHNOLOGY The IWSDS Workshop series provides an international forum for the presentation of research and applications and for lively discussions among researchers as well as industrialists, with a spe @tony的微博 O网页链接 The Dialog Research Center DialRC conducts community-wide spoken dialog challenges (SDC). There are a variety of ways to participate in this effort, which is aimed at bringing the community together and providing common data and tasks.
医疗健康是个好方向，不过门槛也挺高的。真正需要知识积累！！！ @StephanieYR 移动医疗！诊断app！请各位记得这两事…或许程序猿们天天沉浸在代码的世界并不知道，外面的人从上市公司到投资机构还有医疗单位，都在绞尽脑汁的想接近做这方面的IT团队…一个个快急哭了都…醒醒吧…亲们～也就只有你们把人脸识别先应用到查询AV女优内种事情！运动感应机器人什么的拿去先做游戏了…
2014-7-12
我们的工具箱越来越丰富了，全民maker时代！！！ //@南大周志华: //@vinW: //@老师木: //@搞笑人士: //@王斌_ICTIR:M //@梁斌penny: 真有诚意啊 //@52nlp: //@阿邦dd: 马上去看 @赵家平USC Jeff Hinton组把deep CNN(CovNets)在ImageNet上train好的模型放到网上了，试了下classification, retrieval, image2text的在线demo, amazing! http://deeplearning.cs.toronto.edu/ 最重要的是他们的source code以及installation & documentation 也一并公布，超过Rob Fergus学生的Clarifai http://www.clarifai.com/
看来腾讯也要在deep learning这方面发力了，内部做的一些事情也开始放开。 (评论给 深度神经网络DNN的多GPU数据并行框架 及其在语音识别的应用 - 云聚网 http://www.yunjuu.com/info/122875.html )
2014-7-14
http://webscalesql.org/ “We're Gonna Need A Bigger Database” Who is behind WebScaleSQL? WebScaleSQL currently includes contributions from MySQL engineering teams at Facebook, Google, LinkedIn, and Twitter. Together, we are working to share a common base of code ... //@tony的微博: @郑昀 因为要摆脱Oracle对MySQL用户的潜在制约（如MySQL在5.5.30过渡到5.5.31版本期间逐渐闭源，不再公开补丁的测试数据和修订历史等），继苹果、维基百科等迁移数据库之后，Google高级系统工程师Jeremy Cole透露，谷歌的开源数据中心将由MySQL迁移至MariaDB，至此，Google的数据库已大部转向MariaDB10.0。
http://blog.mimvp.com/ blog.ithomer.net Python爬虫抓站的一些技巧 看来做了不少事情了，哈哈！！！
http://obmem.info/?p=476 obmem.info 原来在这里 http://obmem.info/?p=848 使用python/casperjs编写终极爬虫-客户端App的抓取 @tony的微博 O网页链接 blog.ithomer.net Python爬虫抓站的一些技巧 看来做了不少事情了，哈哈！！！
https://github.com/scrapy/scrapy scrapy现在用的人应该还是不少吧，github的关注度挺高的！！！ //@tony的微博:O网页链接 obmem.info 原来在这里 O网页链接 使用python/casperjs编写终极爬虫-客户端App的抓取 @tony的微博 O网页链接 blog.ithomer.net Python爬虫抓站的一些技巧 看来做了不少事情了，哈哈！！！
http://licstar.net/archives/262 维基百科简体中文语料的获取 wiki的数据确实不错，就是少了点。 如果有人能将百度百科，互动百科，百度问答，百度知道， 搜搜问问等等数据源统一起来，构造一个大的生活百科类语料库就好了。 微博也是，多好的知识源啊。能开放点就好了。
http://www.reddit.com/r/MachineLearning/comments/20my9s/from_word2vec_to_doc2vec_an_approach_driven_by/?utm_source=twitterfeed&utm_medium=twitter From word2vec to doc2vec: an approach driven by Chinese restaurant process From Frequency to Meaning: Vector Space Models of Semantics @算文解字
ACL2014# 出现更多深度学习，特别是MT领域，从BBN将Bengio的NNLM扩展到同时考虑目标和源语言而完爆基线的神作，到MSR的直接优化翻译模型得到向量短语表示和结合两种NN做MT的文章。另外用到tensor的文章明显变多，其中一篇的亮点是通过优化而不是分解得到低纬近似的表征和参数。
http://lewuathe.com/blog/2014/02/23/trying-word2vec-from-twitter-corpus/ lewuathe.com Trying Word2vec From Twitter Corpus I used MeCab for the morphological analysis. 日本人也在研究这些哈哈！果然在用mecab！ //@tony的微博:看来大家都在热心研究word2vec和gensim啊！！！ //@梁斌penny: @西瓜大丸子汤 刚才说到python优化，举个具体的例子 Gensim的作者把word2vec(深度学习)做了几个经典优化：循环，numpy/BLAS，cython，多线程（真的可以）结果效率提高了上千倍，比Google开源出来的原始C版本还快3倍。他最近还写了个word2vec教程。无论是学习word2vec还是python优化，都不可不看 O网页链接
2014-7-15
https://hacklab.to/ hacklab.to Toronto's hacker collective 他们在做不少有趣的事情。 We use the term "hacking" in the MIT sense. We make things, repurpose things, program things, invent things, and make lights blink!
HackLab.TO is a community space with a diverse membership, including artists, computer programmers, web designers, and hardware hackers. It is inspired by the philosophies of the global hackerspaces movement which encourages people to socialize, share knowledge, and work together @tony的微博 O网页链接 hacklab.to Toronto's hacker collective 他们在做不少有趣的事情。 We use the term "hacking" in the MIT sense. We make things, repurpose things, program things, invent things, and make lights blink!
http://snapforbeginners.com/ snapforbeginners.com 大家正在用haskell做一些很实用的东西。
We also show that with Mio, McNettle (an SDN controller written in Haskell) can scale effectively to 40+ cores, reach a thoroughput of over 20 million new requests per second on a single machine, and hence become the fastest of all existing SDN controllers. @tony的微博 O网页链接 snapforbeginners.com 大家正在用haskell做一些很实用的东西。
http://engineering.silk.co/post/90354057868/announcing-rest-a-haskell-rest-framework Announcing rest - A Haskell REST framework This API can then be run in different web frameworks like happstack, snap, or wai. //@tony的微博:We also show that with Mio, McNettle (an SDN controller @tony的微博 O网页链接 snapforbeginners.com 大家正在用haskell做一些很实用的东西。
http://www.cocoachina.com/industry/20140516/8449.html GHC 7.8.1最近已经发布，为Haskell生态系统带来了多项改善。Haskell现在可以针对iOS编译，并且增加了多种新特性，如闭合类型族、角色、重载列表、模式同义词等以及做了一些其他性能改进。 Haskell可以通过clang编译成对iOS可用 //@to @tony的微博 O网页链接 snapforbeginners.com 大家正在用haskell做一些很实用的东西。
2014-7-16
https://github.com/pkelsey/libuinet This is a user-space port of the FreeBSD TCP/IP stack, begun with the FreeBSD 9.1-RELEASE sources and many pieces of Kip Macy's user-space port of an earlier version of the FreeBSD stack, libplebnet. 这个确实很有意思！ //@tony的微博:系统级的优 @bnu_chenshuo C1000k新思路：今年BSDCan14上有人把FreeBSD 9的TCP/IP协议栈移植到了用户态，意味着并发TCP连接不占用系统文件数，只占内存。优化也更直接，不再是调黑盒参数组合，而是直接上profiling，再改代码。用户态的吞吐量比不上内核，不过对C1000k应该不成问题。搜 libuinet。
http://www.bsdcan.org/2014/schedule/events/447.en.html BSDCan 2014 The Technical BSD Conference Userspace Networking with libuinet A portable and performant TCP/IP stack-in-a-box
libuinet is a userspace library version of the FreeBSD TCP/IP stack that also includes extensions to the base stack functionality that make it particularly useful in network infrastructure equipment. @tony的微博 O网页链接 BSDCan 2014 The Technical BSD Conference Userspace Networking with libuinet A portable and performant TCP/IP stack-in-a-box
libuinet was originally conceived as a way to bring highly scalable transparent proxy functionality to the free, portable TCP proxy WANProxy (http://wanproxy.org/ 系统和网络玩的好，真实什么东西都可以做啊！ //@tony的微博:libuinet is a userspace library version of the FreeBSD TCP/ @tony的微博 O网页链接 BSDCan 2014 The Technical BSD Conference Userspace Networking with libuinet A portable and performant TCP/IP stack-in-a-box
http://www.wanproxy.org/index.shtml WANProxy is a free, portable TCP proxy which makes TCP connections send less data, which improves TCP performance and throughput over lossy links, slow links and long links. This is just what you need to improve performance over satellite, wire @tony的微博 O网页链接 BSDCan 2014 The Technical BSD Conference Userspace Networking with libuinet A portable and performant TCP/IP stack-in-a-box
WANProxy also supports optimizing SSH traffic specifically, in addition to other TCP protocols. WANProxy is the first WAN optimization software to support SSH WAN optimization. As part of the work towards making WANProxy usable as a transparent proxy, a library libuinet. //@tony @tony的微博 O网页链接 BSDCan 2014 The Technical BSD Conference Userspace Networking with libuinet A portable and performant TCP/IP stack-in-a-box
http://www.cs.toronto.edu/~ranzato/ Marc'Aurelio Ranzato research scientist at the Facebook AI Research lab in Menlo Park CA. I previously worked at Google in the Brain team, and before that, I was a post-doctoral fellow in Machine Learning, University of Toronto.
I worked in Geoffrey Hinton's lab for two wonderful years. I did my Ph.D. in Computer Science at New York University in Yann LeCun's group. I am originally from Padova in Italy, where I graduated in Electronics Engineering. @tony的微博 O网页链接 Marc'Aurelio Ranzato research scientist at the Facebook AI Research lab in Menlo Park CA. I previously worked at Google in the Brain team, and before that, I was a post-doctoral fellow in Machine Learning, University of Toronto.
facebook现在确实网络了不少牛人，在硬件和ai等多方面发展，和google以前走过的路很像！ //@tony的微博:I worked in Geoffrey Hinton's lab for two wonderful years. I did my Ph.D. in Computer Science at New York University in Yann LeCun's group. I am originally from Padova in Italy, where @tony的微博 O网页链接 Marc'Aurelio Ranzato research scientist at the Facebook AI Research lab in Menlo Park CA. I previously worked at Google in the Brain team, and before that, I was a post-doctoral fellow in Machine Learning, University of Toronto.
semantic的问题，经历了很多尝试，移动智能时代可以做出一些真正的产品出来。 @西瓜大丸子汤 Denny Vrandečić的新文章Wikidata: A Free Collaborative Knowledge Base http://semanticweb.memect.com/?p=270 Denny以前的两个项目: Semantic Mediawiki和Wikidata都是相当成功的。现在他在谷歌的硅谷分舵，当然无疑问的，在知识图谱组。这个组召集的牛人实在是太多了: Noy, Guha,Dan Brickley,David Huynh...
类似google glass这类的交互工具，开发出更好的智能交互应用，开启全民知识库构建时代。 //@王海勋haixun:Semantic Web和Semantic Network不是一回事。Semantic Web是一套规则，早先依赖于人工annotation，之后转向Linked Data，依赖结构数据内在的语义。两者都不成功。人工智能，特别是自然语言理解， @西瓜大丸子汤 Jim Hendler今天的视频和PPT: Semantic Web: The Inside Story 强烈推荐搞人工智能的同仁都看看 O网页链接 语义网作为符号主义走向应用的尝试，也曾获得与深度学习类似的投资与眼球。结合前两天关于AI winter的讨论，其在今天尤其有参考意义 O网页链接 @王海勋haixun @Gary南京