ADM-201 dump PMP dumps pdf SSCP exam materials CBAP exam sample questions

【人工智能实验系列】多维空间的可视化 – 译学馆
未登录,请登录后再发表信息
最新评论 (0)
播放视频

【人工智能实验系列】多维空间的可视化

A.I. Experiments: Visualizing High-Dimensional Space

[Music Play]
[音乐]
Hi, I’m Daniel.
大家好 我是丹尼尔
Hi, I’m Martin.
大家好 我是马丁
Hi, I’m Fernanda.
大家好 我是费尔南达
“Small team of data visualizers”
“数据视觉小分队”
Machine learning is pretty complex,
计算机学习是十分困难的
So we’ve been experimenting with ways to
所以 我们一直在探索一种方法
visualize what’s happening.
去想象发生了什么
[Music]
[音乐]
There’s a core concept in Machine Learning,
计算机学习中有一个核心观念
called high dimensional space.
叫做:多思维空间
Here’s one way to wrap your head around this concept.
有一种关于这个概念的方法让你头疼
You can think about people as being high dimensional.
你可以想想那些多思维的人
For example,take female scientists
例如 女科学家
you can think about
你也可以想想
when they were born
他们的出生日期
where they were born
他们出生的地点
their feels of the study
他们如何学习的
each of these is like a dimension of that person
所有这些都属于人的多思维表现
These dimensions become difficult to tangle
这时候 散发思维就会由困难变成混乱
when you think about different people.
当你考虑不同人的时候
Because someone might be similar in someways
因为有些人在某些方面可能是相似的
but very different in others
但又不同于他人
But this is a kind of thing
对于这样的一种情况
you can use Machine Learning for
你就可以使用机械学习方法了
With Machine Learning
用机器学习
the computer is told the meaning of these dimensions
电脑就会被告知这些思维的意思
just sees them as numbers
就像数字那样
And it sees each set of numbers as data point
它就会把每组数字看成数据点
but by looking across all of these dimensions of ones
但是仔细展望这些思维
It’s able to play through related point closer together in high dimensional space
在多思维空间可以把相关的数据点联系在一起
Here is a concrete example
举一个具体的例子
Where words treated in high dimensional data points
在一个单词里可以散发思维考虑
The important thing to remember
记住一件重要的事
is that we haven’t told the computer the meaning of words
我们没有告诉电脑单词的意思
Instead,we have showing millions of sentences as a example so how words get used
可是 我们却看到了成千上万关于这个单词的例子
Here is a visualization of results
有一个可视的例子
We are looking at a subset of words
我们在寻找一个单词的子集
the computer has learned about each dot represents one word
电脑知道每点代表一个单词
each word is a data point with two hundred the dimensions
一个单词包含两百种思维的数据点
using technique called T-SNE
使用一种叫降维的方法
the computer cluster words together that consider words releated
电脑把那些相关的单词聚集在一起
and clusters form based meaning
相似意思的单词也会被聚集在一起
even though we never talk their the meaning of words
即使我们从未讨论这些单词的意思
Here is a cluster of numbers
这有一组集合的数字
months of the year
月份
words related to space
与空间相关的单词
people’s names
人名
cities and so on
城市 等等
we can also look closely at smaller set of words
我们可以更紧密发现在一组单词中
if we search “piano”
如果我们查找钢琴
we can run T-SNE only on words relate to piano
我们能使用降维工具帮我们寻找与钢琴相关的单词
we get clusters of composers, genres
我们集合作曲流派
musical, instruments and more
乐具
and this approach doesn’t just work for words
此时 这种方法并不适用了
For example
例如
you can also treat an image as a high-dimensional data point
以多思维思考 你可以看做是一幅图案
Here is a data set
这有一组数据
a lot of people wrote digit between zero and nine ,people write all kinds of ways
许多人用各种方式从一写到九
so the question is
所以 问题就是
Instead of us needing to manually code the rule
而不是随意操控密码规则
for all the ways people write
对于人们写的方式
could a machine figure it out itself use Machine Learning?
使用机器学习真的可以弄懂它本身嘛
each images seven hundreds data for pixels
每个图像都对应时轴
the computer treats each pixel as a dimension
电脑把每个像素看成一种思维
again,using T-SNE
然后 使用降维技术
it cluster these images in high-dimensional space
用多思维空间聚集这些图案
we’ve color coded them
我们按颜色区分好
so it is easier for us to see what’s going on
所以我们能更好地看见发生了什么
and you can see groups of digit clustering together
我们可以看见成组的数据集合在一起
if learn something about the meaning about these digits
如果我们知道了这些数据的意思
This visualization techniques we have been exploring
我们一直在探寻的可视化技术
can be use for all kinds of things
可在很多方面应用
That’s why we are working on open sourcing
那就是我们让它们开源的原因
all of these as spread of tensorflow
所有这些都属于Tensorflow的延展
So that anyone can use these tool to explore their data
所以不论谁都可以应用这项技术去探索他们在研究的

发表评论

译制信息
视频概述

介绍了高维空间的可视化,什么时候可以使用以及使用的范围。。

听录译者

小静子

翻译译者

小静子

审核员

自动通过审核

视频来源

https://www.youtube.com/watch?v=wvsE8jm1GzE

相关推荐