注册 登录  
   显示下一条  |  关闭
温馨提示!由于新浪微博认证机制调整,您的新浪微博帐号绑定已过期,请重新绑定!立即重新绑定新浪微博》  |  关闭







2017-02-06 05:25:34|  分类: 听的,读的,看的 |  标签: |举报 |字号 订阅

  下载LOFTER 我的照片书  |
今天读到的是Stephen Wolfram。
-Wolfram Mathematica

-A new kind of science.
Rule 30.
the principle of computational equivalence*

-wolframAlpha,knowledge engine。
(有没有search engine for design/art)
knowledge based computing


物质的本质就是计算。宇宙看作是necwork,需要evolve by新创造的rule。有些是hopeless universe,不必深入,你就能发现很多后悬空卷。

-Wolfram Mathematica
-I started off at a young age as a physicist using computers as tools.Then, I started drilling down, thinking about the computations I might want to do, trying to figure out what primitives they could be built up from and how they could be automated as much as possible. Eventually, I created a whole structure based on symbolic programming and so on that let me build Mathematica.
-I have to admit, actually, that I also had a very selfish reason for building Mathematica: I wanted to use it myself, a bit like Galileo got to use his telescope 400 years ago.But I wanted to look not at the astronomical universe, but at the computational universe.

drilling down深度研究,探讨
symbolic programming

-A new kind of science
- So we normally think of programs as being complicated things that we build for very specific purposes.But what about the space of all possible programs? 
-Here's a representation of a really simple program.So, if we run this program, this is what we get.let's try changing the rule for this program a little bit. Now we get another result, still very simple. But if we keep running this for a while, we find out that although the pattern we get is very intricate, it has a very regular structure. 

-Let's just run all possible programs of the particular type that we're looking at. They're called cellular automata. 
-You can see a lot of diversity in the behavior here. Most of them do very simple things, but if you look along all these different pictures, at rule number 30, you start to see something interesting going on. 

cellular automata 单元自动机(automata自动装置,机器人)

- So let's take a closer look at rule number 30 here. So here it is. We're just following this very simple rule at the bottom here, but we're getting all this amazing stuff. 

/Rule 30/

-To understand it,I eventually had to create a whole new kind of science.
-This science is different, more general, than the mathematics-based science that we've had for the past 300 or so years. You know, it's always seemed like a big mystery: how nature, seemingly so effortlessly,manages to produce so much that seems to us so complex. 

so much that

我觉得我们已经发现了其中的奥秘。这个知识我们能探索的计算空间的一个样本。它们都像30号规则。知道这个,我们开始能解释很多科学中长期以来的谜团。同时也带来新问题。例如computational irreducibility*计算的不可划规性。
Computations that cannot be sped up by means of any shortcut 
sped  up加速
by means of依靠,用
-we're used to having science let us predict things, but something like this is fundamentally irreducible. The only way to find its outcome is, effectively, just to watch it evolve.

It's connected to, what I call, the principle of computational equivalence*, which tells us that even incredibly simple systems can do computations as sophisticated as anything. 

*the principle of computational equivalence 计算等价性原理
Almost all processes that are not obviously simple can be viewed as computations of equivalent sophistication (Wolfram 2002, pp. 5 and 716-717).

“计算等价性原理(The Principle of Computational Equivalence)简单来说,就是认为任何看起来比较复杂的系统(流体、社会系统、蚁群,等等),他们的复杂度都是相同的,而且都达到了复杂性的极限——它们的复杂度,与宇宙中其他极为复杂的系统,例如大脑,是相同的。而进一步的,这个原理似乎从计算的视角,回答了「人能否理解宇宙」,这样的「终极问题」。” 来自知乎

-wolframAlpha:Knowledge based computing
-this has deep implications about the limits of science, about predictability and controllability of things like biological processes or economies, about intelligence in the universe, about questions like free will and about creating technology.
-ever since I was a kid, I'd been thinking about systematizing knowledge and somehow making it computable. 
 -I decided to just try to see how much of the systematic knowledge that's out there in the world we could make computable.

-But the goal is to democratize all of this knowledge, and to try and be an authoritative source in all areas. To be able to compute answers to specific questions that people have, not by searching what other people may have written down before, but by using built in knowledge to compute fresh new answers to specific questions.

-a crucial idea of Wolfram Alpha is that you can just ask it questions using ordinary human language,which means that we've got to be able to take all those strange utterances that people type into the input field and understand them. 

-Two big things happened: First, a bunch of new ideas about linguistics that came from studying the computational universe; and second, the realization that having actual computable knowledge completely changes how one can set about understanding language. 

-And, in fact, there's been an interesting coevolution that's been going on between Wolfram Alpha and its human users. Right now, if we look at web queries, more than 80 percent of them get handled successfully the first time. And if you look at things like the iPhone app, the fraction is considerably larger.搜索类似app之类的信息,成功的部分就更大了
get handled

I've realized that Wolfram Alpha actually gives one a whole new kind of computing that one can call knowledge-based computing, in which one's starting not just from raw computation, but from a vast amount of built-in knowledge. And when one does that, one really changes the economics of delivering computational things, whether it's on the web or elsewhere.
这个平台带给我们新的计算,称之knowledge-based computing

-I think the most exciting thing about this is that it really gives one the chance to democratize programming. I mean, anyone will be able to say what they want in plain language. Then, the idea is that Wolfram Alpha will be able to figure out what precise pieces of code can do what they're asking for and then show them examples that will let them pick what they need to build up bigger and bigger, precise programs. 
democratize programming 全民化编程

Wolfram Alpha and Mathematica are actually now full of algorithms that we discovered by searching the computational universe. And, for example, this — if we go back here — this has become surprisingly popular among composers finding musical forms by searching the computational universe.

-that leads to kind of an ultimate question: Could it be that someplace out there in the computational universe we might find our physical universe?
Perhaps there's even some quite simple rule, some simple program for our universe. Well, the history of physics would have us believe that the rule for the universe must be pretty complicated. But in the computational universe, we've now seen how rules that are incredibly simple can produce incredibly rich and complex behavior. 
-If the rules for the universe are simple, it's kind of inevitable that they have to be very abstract and very low level; operating, for example, far below the level of space or time, which makes it hard to represent things.

这些最终导向一个终极问题 有没有可能使这个计算空间 与我们的物理世界相融合? 也许存在简单的规则 一些简单的程序,对于我们的物理世界来说。 物理的历史让我们相信 宇宙的内部规则一定是很复杂的 但是在计算空间中 我们已经看到那些规则惊人的简单 却能够产生非常丰富和复杂的结果 所以,这可能是我们的物理世界的本质吗? 如果这个宇宙的规则很简单 不可避免的,他们一定是 十分抽象以及初级 远远运行于 时间、空间之下 这种运行方法很难表现某种东西

- But in at least a large class of cases, one can think of the universe as being like some kind of network, which, when it gets big enough, behaves like continuous space in much the same way as having lots of molecules can behave like a continuous fluid. Well, then the universe has to evolve by applying little rules that progressively update this network. And each possible rule, in a sense, corresponds to a candidate universe.
之后,宇宙进化就要依靠 应用这个网络中不断更新的简单规则。 并且,每一个可能的规则,在某种程度上说, 对应一个候选空间

-here are a few of the candidate universes that I've looked at.Some of these are hopeless universes, completely sterile, with other kinds of pathologies like no notion of space, no notion of time, no matter, other problems like that. But the exciting thing that I've found in the last few years is that you actually don't have to go very far in the computational universe before you start finding candidate universes that aren't obviously not our universe.


这里有几个候选空间 我正在研究的 一些是没希望的空间完全不能演化, 包括很多缺点,例如没有空间的观念 没有时间的概念,没有物质 或者类似的其他问题。

但是,我近几年发现的最令人激动的事 是你其实不必深入,在计算空间中 你就能发现与我们的物理空间。

- Here's the problem: Any serious candidate for our universe is inevitably full of computational irreducibility. Which means that it is irreducibly difficult to find out how it will really behave, and whether it matches our physical universe. A few years ago, I was pretty excited to discover that there are candidate universes with incredibly simple rules that successfully reproduce special relativity, and even general relativity and gravitation, and at least give hints of quantum mechanics*.
quantum mechanics* 

明显不同的候选空间问题在这里: 任何有可能的候选空间 不可避免地充满了计算不可化归性, 这意味着简化它的具体表现,是极其困难的,并且不易判断它是否符合我们的物理世界。

几年前,我非常兴奋地发现有些候选空间具有极其简单的规则,却能成功再现狭义相对论 和广义相对论以及重力,而且至少还给出了量子力学的暗示。

-I'm committed to seeing this project done, to see if, within this decade, we can finally hold in our hands the rule for our universeand know where our universe lies in the space of all possible universes
我们是否最终可以掌握 我们宇宙的规则 并且知道我们宇宙在 所有可能的宇宙空间的位置.

2017年02月05日 - Yuan - Yuan
阅读(27)| 评论(1)
推荐 转载




<#--最新日志,群博日志--> <#--推荐日志--> <#--引用记录--> <#--博主推荐--> <#--随机阅读--> <#--首页推荐--> <#--历史上的今天--> <#--被推荐日志--> <#--上一篇,下一篇--> <#-- 热度 --> <#-- 网易新闻广告 --> <#--右边模块结构--> <#--评论模块结构--> <#--引用模块结构--> <#--博主发起的投票-->


网易公司版权所有 ©1997-2017