g****e
发帖数: 222 | 1 anyone intersted in how the brain processes some infomation? ;P
here is one issue
imagine you are asked to spell a word, like 'internet'
you get the knowledge of the spelling of this from your
memory of mental representation,
load this into a temporary buffer (working memory, whatever
we call it), then spell out letter by letter.
does this plausible or make any sense?
evidence convergent that in this buffer not only hold
the elements(letters) and their orders, but also some
other features or str |
E*****y
发帖数: 33 | 2 这简直就是台计算机，拼单词的过程就象让扬声器或显示器
发某个声音或显示某个东西类似。。。
查找存在硬盘或其他媒体上的数据库
把所需东西载入内存
在内存中以一定的地址顺序存放相关信息
以不同地址来表示顺序，CPU(大脑)处理内存(神经元)的数据是
按地址顺序来的。。。呵呵
计算机的内存是要不断刷新的。。。
计算机的内存(RAM)刷新好象不是这样的，具体忘了。。。
计算机的内存刷新也会出错，不过出错几率很小。。。
呵呵，建议去看看计算机原理。。。真是很相似
【在 g****e 的大作中提到】
: anyone intersted in how the brain processes some infomation? ;P
: here is one issue
: imagine you are asked to spell a word, like 'internet'
: you get the knowledge of the spelling of this from your
: memory of mental representation,
: load this into a temporary buffer (working memory, whatever
: we call it), then spell out letter by letter.
: does this plausible or make any sense?
: evidence convergent that in this buffer not only hold
: the elements(letters) and their orders, but also some
|
g****e
发帖数: 222 | 3 ft.of coz ah! computer borrows idea from brain,
we borrow terminology from CS
of coz not necessarily the same
just a possible solution ah
【在 E*****y 的大作中提到】
: 这简直就是台计算机，拼单词的过程就象让扬声器或显示器
: 发某个声音或显示某个东西类似。。。
:
: 查找存在硬盘或其他媒体上的数据库
: 把所需东西载入内存
: 在内存中以一定的地址顺序存放相关信息
: 以不同地址来表示顺序，CPU(大脑)处理内存(神经元)的数据是
: 按地址顺序来的。。。呵呵
: 计算机的内存是要不断刷新的。。。
: 计算机的内存(RAM)刷新好象不是这样的，具体忘了。。。
|
E*****y
发帖数: 33 | 4 我感觉大脑不应该跟笨蛋计算机一样的运作，它应该有自己的运算方式，
并行 + 协同 + 模糊算法 可能是大脑的主要运转方式。。。 计算机下
棋是很难有全局观念的，只有在“并行 + 协同 + 模糊算法" 这种模式
下才可能有。。。呵呵
【在 g****e 的大作中提到】
: ft.of coz ah! computer borrows idea from brain,
: we borrow terminology from CS
: of coz not necessarily the same
: just a possible solution ah
|
g****e
发帖数: 222 | 5 ft. tong2 xue2 ah
implementation! implementation!
everyone can speak these sexy words like parrallal processing
or whatever
but tell how it's possible to work that way.
details!!!
xixi
【在 E*****y 的大作中提到】
: 我感觉大脑不应该跟笨蛋计算机一样的运作，它应该有自己的运算方式，
: 并行 + 协同 + 模糊算法 可能是大脑的主要运转方式。。。 计算机下
: 棋是很难有全局观念的，只有在“并行 + 协同 + 模糊算法" 这种模式
: 下才可能有。。。呵呵
|
E*****y
发帖数: 33 | 6
计算机的并行计算的数很多呀。。。
对计算机来说，就是把一个大任务分成一些小任务，而这些小任务在
不同节点(也是一些处理单元甚至是完整的计算机)上同时进行，完成
以后把结果送回集中处理。这里的关键问题是大问题怎么分成小问题，
各个节点又怎么样协调起来，个节点之间的信息传输什么的。。。
对于大脑来说，大脑也有很多处理单元。大脑在干什么事时，各处理
单元同时协调运行，然后再在大脑的某部分做最后处理。。。可能不
同类型的任务是在大脑的不同部分最后处理的。。。呵呵
【在 g****e 的大作中提到】
: ft. tong2 xue2 ah
: implementation! implementation!
: everyone can speak these sexy words like parrallal processing
: or whatever
: but tell how it's possible to work that way.
: details!!!
: xixi
|
E*****y
发帖数: 33 | 7
哦，对了，对于量子计算机，并行就没这么复杂了，量子计算
有内在的并行行。。。 呵呵，说不定大脑的运行方式就是类似
于量子计算机的方式。。。呵呵，量子大脑
【在 E*****y 的大作中提到】
:
: 计算机的并行计算的数很多呀。。。
: 对计算机来说，就是把一个大任务分成一些小任务，而这些小任务在
: 不同节点(也是一些处理单元甚至是完整的计算机)上同时进行，完成
: 以后把结果送回集中处理。这里的关键问题是大问题怎么分成小问题，
: 各个节点又怎么样协调起来，个节点之间的信息传输什么的。。。
: 对于大脑来说，大脑也有很多处理单元。大脑在干什么事时，各处理
: 单元同时协调运行，然后再在大脑的某部分做最后处理。。。可能不
: 同类型的任务是在大脑的不同部分最后处理的。。。呵呵
|
g****e
发帖数: 222 | 8
what diff about it?
whatever the process is, u need to provide a possiblity for
these processes to be able to generate those different patterns
of products
【在 E*****y 的大作中提到】
:
: 哦，对了，对于量子计算机，并行就没这么复杂了，量子计算
: 有内在的并行行。。。 呵呵，说不定大脑的运行方式就是类似
: 于量子计算机的方式。。。呵呵，量子大脑
|
E*****y
发帖数: 33 | 9
哦，传统计算机的操作都是 0,1 的操作，所有数据都是用 0,1 来表示。
而量子计算机的操作是对希尔伯特空间态的操作，0用态 |0>, 1用态|1>
来表示。。。 哦，对了，以前传过一篇科普到这的，自己看去吧
呵呵，不知道，反正量子计算机不会比图灵机更厉害。。。
【在 g****e 的大作中提到】
:
: what diff about it?
: whatever the process is, u need to provide a possiblity for
: these processes to be able to generate those different patterns
: of products
|
a*****t
发帖数: 81 | 10 个人认为大脑内内并不存在存在一个所谓的buffer. 看到一个熟悉的单词如 internet然后把
它说出来应该至少包括下面几个动作.
1.记忆的唤起 结果是存贮 internet 单词发音的大脑特定区域被激活, 从现在研究
成果来看, 这个存贮区域极有可能在前脑.
2. 语言的翻译 从记忆(假设记忆的是语言动作序列) 到语言中枢信息的传递
3. 动作序列的执行 motor cortex -> brain stem neclei -> muscle
这个问题的争论点其实就是我们说 How are you. 这三个单词是依次激活->翻译->执行
还是全部翻译好再一次执行. 从说话存在停顿来看, 我更偏向前者.
【在 g****e 的大作中提到】
: anyone intersted in how the brain processes some infomation? ;P
: here is one issue
: imagine you are asked to spell a word, like 'internet'
: you get the knowledge of the spelling of this from your
: memory of mental representation,
: load this into a temporary buffer (working memory, whatever
: we call it), then spell out letter by letter.
: does this plausible or make any sense?
: evidence convergent that in this buffer not only hold
: the elements(letters) and their orders, but also some
|
|
|
g****e
发帖数: 222 | 11 hehe....
are you saying that in 'how are you' case, when 'how' is activated
or translated or even articulated, 'are you' are still on the fly yet?
apparently that's not the case..
for two serial processes, like 'activation' or translation or whatever,
if one is parallel and one is serial, or one is faster than another,
a buffer is needed to hold the info processed by the first one. that's
a base logic of proposing buffer.
of coz if another mechanism can do the work w/out such thing, would also
b
【在 a*****t 的大作中提到】
: 个人认为大脑内内并不存在存在一个所谓的buffer. 看到一个熟悉的单词如 internet然后把
: 它说出来应该至少包括下面几个动作.
: 1.记忆的唤起 结果是存贮 internet 单词发音的大脑特定区域被激活, 从现在研究
: 成果来看, 这个存贮区域极有可能在前脑.
: 2. 语言的翻译 从记忆(假设记忆的是语言动作序列) 到语言中枢信息的传递
: 3. 动作序列的执行 motor cortex -> brain stem neclei -> muscle
: 这个问题的争论点其实就是我们说 How are you. 这三个单词是依次激活->翻译->执行
: 还是全部翻译好再一次执行. 从说话存在停顿来看, 我更偏向前者.
|
l**n
发帖数: 67 | 12
I don't know your way of doing science. To me i guess your questions
may not have a common solution for all human-beings. The answer may
depend on the learning process for different human-beings. We know
that for artificial neural network, the training phase takes a very
important role for the performance of the neural net. Diversity is
indeed a distinct character for human-being. And different individual
may acquire different abilities from their own experience. And your
may be in the trouble
【在 g****e 的大作中提到】
: hehe....
: are you saying that in 'how are you' case, when 'how' is activated
: or translated or even articulated, 'are you' are still on the fly yet?
: apparently that's not the case..
: for two serial processes, like 'activation' or translation or whatever,
: if one is parallel and one is serial, or one is faster than another,
: a buffer is needed to hold the info processed by the first one. that's
: a base logic of proposing buffer.
: of coz if another mechanism can do the work w/out such thing, would also
: b
|
g****e
发帖数: 222 | 13 thx.
you mean the process of learning speaking varies from baby
to baby?
and there's no general grammer or rule of information processing,
but different for each individual according to his/her experience
or exposure to learning enviorment?
that must be an interesting world ;)
-z
【在 l**n 的大作中提到】
:
: I don't know your way of doing science. To me i guess your questions
: may not have a common solution for all human-beings. The answer may
: depend on the learning process for different human-beings. We know
: that for artificial neural network, the training phase takes a very
: important role for the performance of the neural net. Diversity is
: indeed a distinct character for human-being. And different individual
: may acquire different abilities from their own experience. And your
: may be in the trouble
|
l**n
发帖数: 67 | 14
Yes, it must be different at least in some stage. no body can
develop speaking ability by himself. A ape-man may never develop
speaking ability at all. I really have a hard time learning english
and i am sure my english and chinese speaking abilities follow
different laws. Obviously, all these depends on the process of
learning.
【在 g****e 的大作中提到】
: thx.
: you mean the process of learning speaking varies from baby
: to baby?
: and there's no general grammer or rule of information processing,
: but different for each individual according to his/her experience
: or exposure to learning enviorment?
: that must be an interesting world ;)
: -z
|
S******y
发帖数: 39 | 15 I've been in the field of neuroscience for quite some years, but I have to admit
I don't know the answer to the question, and I doubt anyone has anything more
than educated guess.
The process of image recognition has been relatively well characterized. I'll
skip this part for now, but in principle it goes like this: a rectangle would
activated two horizontally tuned-cells and two vertically tuned cells in the
primary visual cortex, and then there are some cells that receive synapses (the
informa |
