Account Options

  1. Sign in
The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Google Groups Home
« Groups Home
Facebook OpenGraph 终于露峥嵘
There are currently too many topics in this group that display first. To make this topic appear first, remove this option from another topic.
There was an error processing your request. Please try again.
flag
  Messages 1 - 25 of 32 - Collapse all  -  Translate all to Translated (View all originals)   Newer >
The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
 
From:
To:
Cc:
Followup To:
Add Cc | Add Followup-to | Edit Subject
Subject:
Validation:
For verification purposes please type the characters you see in the picture below or the numbers you hear by clicking the accessibility icon. Listen and type the numbers you hear
 
Gary Wang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 5:20 am
From: Gary Wang <gary.wang1...@gmail.com>
Date: Sun, 26 Sep 2010 17:20:34 +0800
Local: Sun, Sep 26 2010 5:20 am
Subject: Facebook OpenGraph 终于露峥嵘
Facebook推出“实时个性化”功能 http://www.techcrunchchina.com/4517:Rotten
Tomatoes网站上已经有多个模块被Facebook化,如上图中圆圈的部分。最左上的圆圈是‘朋友们最喜欢的电影’,下方是‘根据你的评分推荐给你的电影 ’,右侧是‘最近的活动’。这就是Facebook在今年F8大会提出的“实时个性化”(Instant
Personalization)功能,而Rotten Tomatoes是首家实验合作伙伴。

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
谷文栋  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 7:11 am
From: 谷文栋 <wendell...@gmail.com>
Date: Sun, 26 Sep 2010 19:11:13 +0800
Local: Sun, Sep 26 2010 7:11 am
Subject: Re: [resys] Facebook OpenGraph 终于露峥嵘

搜索已死,推荐登基。

2010/9/26 Gary Wang <gary.wang1...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
邓同桐  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 7:12 am
From: 邓同桐 <infocean....@gmail.com>
Date: Sun, 26 Sep 2010 19:12:50 +0800
Local: Sun, Sep 26 2010 7:12 am
Subject: Re: [resys] Facebook OpenGraph 终于露峥嵘

赞!

在 2010年9月26日 下午7:11,谷文栋 <wendell...@gmail.com>写道:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Mahout  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 7:53 am
From: Mahout <mahout.c...@gmail.com>
Date: Sun, 26 Sep 2010 04:53:18 -0700 (PDT)
Local: Sun, Sep 26 2010 7:53 am
Subject: Re: Facebook OpenGraph 终于露峥嵘
为什么要把搜索和推荐对立起来呢

On 9月26日, 下午7时11分, 谷文栋 <wendell...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Boolean Tao  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 8:23 am
From: Boolean Tao <beas...@gmail.com>
Date: Sun, 26 Sep 2010 20:23:31 +0800
Local: Sun, Sep 26 2010 8:23 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

嘿嘿

中国儿童受多年教育世界上只能有一种声音、一个方向是正确的,推荐要登基了,那搜索不死行吗?

------------------------------------------------------------
Boolean Tao

2010/9/26 Mahout <mahout.c...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Zi-Ke ZHANG  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 8:31 am
From: Zi-Ke ZHANG <zhangz...@gmail.com>
Date: Sun, 26 Sep 2010 14:31:29 +0200
Local: Sun, Sep 26 2010 8:31 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

我觉得只要人的主观意愿存在,搜索就不会死 ^_^

2010/9/26 Boolean Tao <beas...@gmail.com>

zike zhang
--
Université de Fribourg
Département de Physique
Chemin du Musée 3
1700 Fribourg
Switzerland

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Boolean Tao  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 8:47 am
From: Boolean Tao <beas...@gmail.com>
Date: Sun, 26 Sep 2010 20:47:43 +0800
Local: Sun, Sep 26 2010 8:47 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

小谷是喊个口号提提气而已,没有说搜索要死的意思~~  有点夸张而已,关注主要观点啊,哈哈

------------------------------------------------------------
Boolean Tao

2010/9/26 Zi-Ke ZHANG <zhangz...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Gary Wang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 8:50 am
From: Gary Wang <gary.wang1...@gmail.com>
Date: Sun, 26 Sep 2010 20:50:43 +0800
Local: Sun, Sep 26 2010 8:50 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
搜索就是一种推荐,推荐就是搜索,这是从技术层面来看。从产品上,还是有不同的适用场景。死或者不死,还是要看大部分用户的使用习惯。
在 2010年9月26日 下午8:31,Zi-Ke ZHANG <zhangz...@gmail.com> 写道:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Xu Yizhi  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 11:38 am
From: Xu Yizhi <xuyi...@gmail.com>
Date: Sun, 26 Sep 2010 23:38:09 +0800
Local: Sun, Sep 26 2010 11:38 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

我觉得搜索跟推荐是一个东西,都是用的一个统计模型,只不过是表现形式不一样罢了。搜索是你输入信息后推荐给你需要的信息。推荐是你没有输入信息,根据你的以前 记录推荐给你信息。这么来看,推荐是搜索的一个特例而已。

2010/9/26 Gary Wang <gary.wang1...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Tinyfool  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 11:43 am
From: Tinyfool <tinyf...@gmail.com>
Date: Sun, 26 Sep 2010 23:43:56 +0800
Local: Sun, Sep 26 2010 11:43 am
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
大面上这么说也可以,不过实质上流行技术上来看,走向根本不同模型也没有什么太紧密的联系。

我认为有实质结合的地方再扯一生万物那一套比较好,不然的话,其实也就是纯净水。

2010/9/26 Xu Yizhi <xuyi...@gmail.com>:

--
Tinyfool的开发日记 http://www.tinydust.net/dev/
代码中国网 http://www.codechina.org
myTwitter: http://twitter.com/tinyfool

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Xu Yizhi  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 1:19 pm
From: Xu Yizhi <xuyi...@gmail.com>
Date: Mon, 27 Sep 2010 01:19:39 +0800
Local: Sun, Sep 26 2010 1:19 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

刚才灌的确实有点像纯净水,那我来点黄河水吧;0

我折腾过一段时间个性化搜索引擎,由于个性化,所以需要推荐系统。

做下去发现搜索和推荐用到的都是统计模型那一套,都是通过计算概率预测用户需求,都是用马尔科夫模型简化,都是通过各种数学模型进行rank,基本上大部分模块 推荐系统和搜索系统都可以共用。不同之处只是处理数据不一样。搜索是对所有web数据进行统计,推荐系统是根据用户日志建立一个最大熵模型,然后进行统计。

后来发现大部分人工智能都是在做统计模型,只是实现细节不一样罢了。比如开复同学的语音识别,证卷自动交易系统等。

当然,这只是个人瞎折腾的感受。

邮件组里各位做推荐系统的,系统核心算法难道不是用的统计模型么?

2010/9/26 Tinyfool <tinyf...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
xlvector  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:15 pm
From: xlvector <xlvec...@gmail.com>
Date: Sun, 26 Sep 2010 18:15:12 -0700 (PDT)
Local: Sun, Sep 26 2010 9:15 pm
Subject: Re: Facebook OpenGraph 终于露峥嵘
是的,就是统计模型,在统计模型之前是基于规则的,比如 产生式系统,数理逻辑

应该说这20-30年,模式识别的最大发展就是统计模型。

On Sep 27, 1:19 am, Xu Yizhi <xuyi...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Boolean Tao  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:36 pm
From: Boolean Tao <beas...@gmail.com>
Date: Mon, 27 Sep 2010 09:36:27 +0800
Local: Sun, Sep 26 2010 9:36 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

咱这技术人群就是矫情~~~

人家小谷喊个口号而已,口号嘛……别较真了啊~~~人家本意也没要把搜索和推荐对立起来……

------------------------------------------------------------
Boolean Tao

2010/9/27 xlvector <xlvec...@gmail.com>


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
raullew  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:44 pm
From: raullew <raul...@hotmail.com>
Date: Sun, 26 Sep 2010 18:44:29 -0700 (PDT)
Local: Sun, Sep 26 2010 9:44 pm
Subject: Re: Facebook OpenGraph 终于露峥嵘
我看下来用统计做模型和用统计做学习
思路还是很有差异的

On 9月26日, 上午10时19分, Xu Yizhi <xuyi...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Tinyfool  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:52 pm
From: Tinyfool <tinyf...@gmail.com>
Date: Mon, 27 Sep 2010 09:52:00 +0800
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
就像你自己说的一样,“大部分人工智能都是在做统计模型”,所以没有任何场景的强调搜索和推荐一致,没什么信息量,再者,你也说了其实也只有个性化搜索是推荐, 基本上倒排索引那套跟推荐系统还是不一样的。

真的明白的混在一起也就算了,其实也明白里面的区别。但是实际上这种调调很烦人,经常被一些半懂不懂的人拿出来用,就像,“哦,推荐系统啊,我知道,他们都说了 ,跟搜索也没什么区别嘛”

2010/9/27 raullew <raul...@hotmail.com>:

--
Tinyfool的开发日记 http://www.tinydust.net/dev/
代码中国网 http://www.codechina.org
myTwitter: http://twitter.com/tinyfool

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
George Zhang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:53 pm
From: George Zhang <birdzhangxi...@gmail.com>
Date: Mon, 27 Sep 2010 09:53:28 +0800
Local: Sun, Sep 26 2010 9:53 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

所谓统计模型也就是在无法完全定义黑箱内的物理规则与逻辑的时候,通过观测黑箱的表现特征,进行经验推测,确实都是属于概率模型
本身人类的调减反射什么的神经活动也是数据经验模型,或者说概率模型,我觉得这个是物理学发展遇到了复杂系统瓶颈时候的一个偷懒的方法,但目前为止还是很有效的

在 2010年9月27日 UTC+8上午9时15分12秒,xlvector <xlvec...@gmail.com>写道:

> 是的,就是统计模型,在统计模型之前是基于规则的,比如 产生式系统,数理逻辑

> 应该说这20-30年,模式识别的最大发展就是统计模型。

> --

Best Regard
George Zhang

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Gary Wang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:57 pm
From: Gary Wang <gary.wang1...@gmail.com>
Date: Mon, 27 Sep 2010 09:57:32 +0800
Local: Sun, Sep 26 2010 9:57 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
又跑题了,我想文栋说的意思是,终于有了一个推荐引擎应用的巨大场景,而不是仅仅捆绑在一个网站,一套数据上。

在 2010年9月27日 上午9:53,George Zhang <birdzhangxi...@gmail.com> 写道:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Tinyfool  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 9:57 pm
From: Tinyfool <tinyf...@gmail.com>
Date: Mon, 27 Sep 2010 09:57:50 +0800
Local: Sun, Sep 26 2010 9:57 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
问题是有些问题确实只有概率或者说统计意义

2010/9/27 George Zhang <birdzhangxi...@gmail.com>:

> 所谓统计模型也就是在无法完全定义黑箱内的物理规则与逻辑的时候,通过观测黑箱的表现特征,进行经验推测,确实都是属于概率模型
> 本身人类的调减反射什么的神经活动也是数据经验模型,或者说概率模型,我觉得这个是物理学发展遇到了复杂系统瓶颈时候的一个偷懒的方法,但目前为止还是很有效的

> 在 2010年9月27日 UTC+8上午9时15分12秒,xlvector <xlvec...@gmail.com>写道:

>> 是的,就是统计模型,在统计模型之前是基于规则的,比如 产生式系统,数理逻辑

>> 应该说这20-30年,模式识别的最大发展就是统计模型。

> --
> Best Regard
> George Zhang

--
Tinyfool的开发日记 http://www.tinydust.net/dev/
代码中国网 http://www.codechina.org
myTwitter: http://twitter.com/tinyfool

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
George Zhang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:00 pm
From: George Zhang <birdzhangxi...@gmail.com>
Date: Mon, 27 Sep 2010 10:00:36 +0800
Local: Sun, Sep 26 2010 10:00 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

推荐和搜索是两个发展时段的产物
在信息时代,为了方便查找信息产生了搜索这种行为
在电商时代,其实是因为电子商务平台的特殊性,使得更个性化的推荐系统有了基础

说白了,如果不是facebook,电商和google的一些竞争关系,google获得了人们的连续网络行为以后,做推荐系统又有何难?
所以技术与模型是一些工具,工具本身的通用性决定了可以用这些工具来解决各种类似的问题,而搜索和推荐系统使用的模型与工具应该是很接近的,他们的差异主要还是 在应用场景上

在 2010年9月27日 上午9:52,Tinyfool <tinyf...@gmail.com>写道:

> 就像你自己说的一样,“大部分人工智能都是在做统计模型”,所以没有任何场景的强调搜索和推荐一致,没什么信息量,再者,你也说了其实也只有个性化搜索是推荐, 基本上倒排索引那套跟推荐系统还是不一样的。

> 真的明白的混在一起也就算了,其实也明白里面的区别。但是实际上这种调调很烦人,经常被一些半懂不懂的人拿出来用,就像,“哦,推荐系统啊,我知道,他们都说了 ,跟搜索也没什么区别嘛”

--
Best Regard
George Zhang

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
George Zhang  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:05 pm
From: George Zhang <birdzhangxi...@gmail.com>
Date: Mon, 27 Sep 2010 10:05:58 +0800
Local: Sun, Sep 26 2010 10:05 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

呵呵,是呀,所以如果从商业角度,文栋的那句话是真实的,是其他大老们撬动google的手段

在 2010年9月27日 上午9:57,Gary Wang <gary.wang1...@gmail.com>写道:

> 又跑题了,我想文栋说的意思是,终于有了一个推荐引擎应用的巨大场景,而不是仅仅捆绑在一个网站,一套数据上。

> --

Best Regard
George Zhang

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
王立才  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:06 pm
From: "王立才" <wiiz...@gmail.com>
Date: Mon, 27 Sep 2010 10:06:25 +0800
Local: Sun, Sep 26 2010 10:06 pm
Subject: Re: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

对用户而言,搜索是一种主动行为,推荐是一种被动行为。

对系统而言,搜索和推荐都需要预测用户偏好或者Rank。

王立才
2010-09-27

发件人: George Zhang
发送时间: 2010-09-27 10:00:48
收件人: resys
抄送:
主题: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘

推荐和搜索是两个发展时段的产物
在信息时代,为了方便查找信息产生了搜索这种行为
在电商时代,其实是因为电子商务平台的特殊性,使得更个性化的推荐系统有了基础

说白了,如果不是facebook,电商和google的一些竞争关系,google获得了人们的连续网络行为以后,做推荐系统又有何难?
所以技术与模型是一些工具,工具本身的通用性决定了可以用这些工具来解决各种类似的问题,而搜索和推荐系统使用的模型与工具应该是很接近的,他们的差异主要还是 在应用场景上

在 2010年9月27日 上午9:52,Tinyfool <tinyf...@gmail.com>写道:

就像你自己说的一样,“大部分人工智能都是在做统计模型”,所以没有任何场景的强调搜索和推荐一致,没什么信息量,再者,你也说了其实也只有个性化搜索是推荐, 基本上倒排索引那套跟推荐系统还是不一样的。

真的明白的混在一起也就算了,其实也明白里面的区别。但是实际上这种调调很烦人,经常被一些半懂不懂的人拿出来用,就像,“哦,推荐系统啊,我知道,他们都说了 ,跟搜索也没什么区别嘛”

--
Best Regard
George Zhang


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
raullew  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:24 pm
From: raullew <raul...@hotmail.com>
Date: Sun, 26 Sep 2010 19:24:40 -0700 (PDT)
Local: Sun, Sep 26 2010 10:24 pm
Subject: Re: Facebook OpenGraph 终于露峥嵘
这些不是我说的啊。。。。。。。。。。

我说的是,虽然都是用统计做,不同的方法之间,至少我觉得,思路差异还挺大的,有些方法我读起来心里是比较抗拒的

On 9月26日, 下午6时52分, Tinyfool <tinyf...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Tinyfool  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:26 pm
From: Tinyfool <tinyf...@gmail.com>
Date: Mon, 27 Sep 2010 10:26:08 +0800
Local: Sun, Sep 26 2010 10:26 pm
Subject: Re: [resys] Re: Facebook OpenGraph 终于露峥嵘
跨线程回复而已,不用紧张

2010/9/27 raullew <raul...@hotmail.com>:

--
Tinyfool的开发日记 http://www.tinydust.net/dev/
代码中国网 http://www.codechina.org
myTwitter: http://twitter.com/tinyfool

 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Mahout  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:44 pm
From: Mahout <mahout.c...@gmail.com>
Date: Sun, 26 Sep 2010 19:44:33 -0700 (PDT)
Local: Sun, Sep 26 2010 10:44 pm
Subject: Re: Facebook OpenGraph 终于露峥嵘
这个说法恐怕不太准确

有几种主要的学习方法,bayesian (以及最大似然最大后验), PAC, 和VC维;bayesian当然是统计模型了,但别的这么说就不准
确,虽然也有人用统计模型来解释他们。

On 9月27日, 上午1时19分, Xu Yizhi <xuyi...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Mahout  
View profile   Translate to Translated (View Original)
 More options Sep 26 2010, 10:49 pm
From: Mahout <mahout.c...@gmail.com>
Date: Sun, 26 Sep 2010 19:49:02 -0700 (PDT)
Local: Sun, Sep 26 2010 10:49 pm
Subject: Re: Facebook OpenGraph 终于露峥嵘
呵呵,主要老看到有人把它们对立起来,比如经常有人会讨论推荐和搜索哪个更复杂

On 9月27日, 上午9时36分, Boolean Tao <beas...@gmail.com> wrote:


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Messages 1 - 25 of 32   Newer >
« Back to Discussions « Newer topic     Older topic »