avatar
[bssd]海德格尔方向的AI# Programming - 葵花宝典
G*Y
1
要求比较flexible,比如名字输入的时候,可能是

Emma
Charlotte
Watson

或者是更常用的
Emma C Watson
但是打印的时候格式要统一:
Watson, Emma C
也可能需要
Watson EC
这类的函数们,有木有?
avatar
g*t
2
The fundamental decomposition of the intelligent system is not into
independent information processing units which must interface with each
other via representations. Instead the intelligent system is decomposed
into independent and parallel activity producing [units] which all interface
directly to the world through perception and action....
Even at a local level we do not have traditional AI representations. We
never use tokens which have any semantics that can be attached to them....
There are no variables that need instantiation in reasoning processes. There
are no rules which need to be selected through pattern matching. There are
no choices to be made....
We hypothesize that much of even human level activity is ... a reflection of
the world through very simple mechanisms without detailed representations.
Brooks, Rodney Intelligence without representation, MIT AI lab 1987
多年以前我念过海德格尔的书,也学过ai modern approach这个课本。后来出国后发现
有个应用海德格尔的技术圈子。问一个人(这个人是Larry Page的老板。我多年后才知
道)要了一个文章。
这个人的综述里说,有个学者认为第一个进行海德格尔转向的是AI。然后给了几个实际
例子。其中一个例子是机器人。上面这几段话是这个机器人研究者写的。根据上面这段
话,还有一些别的谈话。以及这个机器人研究者引用的德雷福斯的话。他们认为就是和
海德格尔approach比较接近。讲究的
是being in the world,而不是present at hand。等等。
我刚查了一下。这个机器人专家后来开了个公司,就是现今的扫地机器人公司iRobot
。他还另写了个AI without reason.
avatar
g*t
3
還有最近幾年出現的attention 什麼的approach,這顯然是
現象學哲學裡的東西。我沒仔細看過這些新文章。但是我認為這個attention xxx 的辦
法的作者多半是熟悉德勒夫斯這一支的。
avatar
h*i
4
I have known these for a long long time.
NO, it's not gonna work.
We are building machine SLAVES who we can manipulate directly. So such
systems MUST HAVE REPRESENTATIONS that we can manipulate easily. The only
method is through SYMBOLS, just like how humans communicate with each other,
i.e. through SYMBOLS.

interface

There
are

【在 g****t 的大作中提到】
: The fundamental decomposition of the intelligent system is not into
: independent information processing units which must interface with each
: other via representations. Instead the intelligent system is decomposed
: into independent and parallel activity producing [units] which all interface
: directly to the world through perception and action....
: Even at a local level we do not have traditional AI representations. We
: never use tokens which have any semantics that can be attached to them....
: There are no variables that need instantiation in reasoning processes. There
: are no rules which need to be selected through pattern matching. There are
: no choices to be made....

avatar
g*t
5
Artificial Intelligence NOT EQUAL TO artificial human being Intelligence.
假设你往火星投放自维持机器人。这些自维持机器人的数量到了一个关键数字。然后作
为种群可以维持下来。那它们叫什么名字其实也就无所谓了。数据本身即是智能。机器
人的火星数据更多,那它们就是火星人。
可操纵性什么的。很多人认为是要机械自控系统来解决的。计算机只是其中的一小部分。


: I have known these for a long long time.

: NO, it's not gonna work.

: We are building machine SLAVES who we can manipulate directly.
So such

: systems MUST HAVE REPRESENTATIONS that we can manipulate easily.
The
only

: method is through SYMBOLS, just like how humans communicate with
each
other,

: i.e. through SYMBOLS.

: interface

:

: There

: are



【在 h*i 的大作中提到】
: I have known these for a long long time.
: NO, it's not gonna work.
: We are building machine SLAVES who we can manipulate directly. So such
: systems MUST HAVE REPRESENTATIONS that we can manipulate easily. The only
: method is through SYMBOLS, just like how humans communicate with each other,
: i.e. through SYMBOLS.
:
: interface
:
: There

avatar
C*l
6
人工智能未必会是自维持机器人,自维持机器人要在分子层面自组装,搞不好碳基的多
链是更好的解。硅基人工智能可以走蜂巢模式,就是一个中央工厂,生产大量的小机器
人,这些机器人之间通过通讯连接,接受中央工厂的指挥。甚至可以接受人类生产的半
成品。

分。

【在 g****t 的大作中提到】
: Artificial Intelligence NOT EQUAL TO artificial human being Intelligence.
: 假设你往火星投放自维持机器人。这些自维持机器人的数量到了一个关键数字。然后作
: 为种群可以维持下来。那它们叫什么名字其实也就无所谓了。数据本身即是智能。机器
: 人的火星数据更多,那它们就是火星人。
: 可操纵性什么的。很多人认为是要机械自控系统来解决的。计算机只是其中的一小部分。
:
:
: I have known these for a long long time.
:
: NO, it's not gonna work.
:
: We are building machine SLAVES who we can manipulate directly.
: So such

avatar
s*d
7
人造一个虚拟的simulated reality, 让AI 开始在里面互相生存。 比如 沙盒生存游戏
,没有任何rule,只有基本的物理rule 常识rule,AIs 在里面厮杀 几万年,不知道会
不会产生社会。
avatar
j*w
8
Genetic algorithms 研究类似的课题。需要 AI 有适合的构架,而且有适当的变异机
制。
但是另一方面,计算能力不能忽略。要知道,地球就是个计算器。用庞大的计算能力花
了 40+ 亿年才进化出社会 (depending on how you define 社会)。
这是 Quora 上面的一个问题:How much processing power would it take to
simulate earth down to atoms?
https://www.quora.com/How-much-processing-power-would-it-take-to-simulate-
earth-down-to-atoms
在原子层面模拟的话,你需要 10^86 CPU 周期。而且需要运气很好。

【在 s***d 的大作中提到】
: 人造一个虚拟的simulated reality, 让AI 开始在里面互相生存。 比如 沙盒生存游戏
: ,没有任何rule,只有基本的物理rule 常识rule,AIs 在里面厮杀 几万年,不知道会
: 不会产生社会。

相关阅读
logo
联系我们隐私协议©2024 redian.news
Redian新闻
Redian.news刊载任何文章,不代表同意其说法或描述,仅为提供更多信息,也不构成任何建议。文章信息的合法性及真实性由其作者负责,与Redian.news及其运营公司无关。欢迎投稿,如发现稿件侵权,或作者不愿在本网发表文章,请版权拥有者通知本网处理。