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

图灵测试:计算机能假装人类通过测试吗? – 译学馆
未登陆,请登陆后再发表信息
最新评论 (0)
播放视频

图灵测试:计算机能假装人类通过测试吗?

The Turing test: Can a computer pass for a human? - Alex Gendler

什么是意识?
What is consciousness?
人造的机器真的能思考吗?
Can an artificial machine really think?
人类的大脑到底仅仅是一个神经元的集体
Does the mind just consist of neurons in the brain,
还是是一种神秘意志的存在?
or is there some intangible spark at its core?
对很多人来说 这些都是关于
For many, these have been vital considerations
未来人工智能的一些重要考虑因素
for the future of artificial intelligence.
但是对于一位英国电脑科学家 阿兰-图灵
But British computer scientist Alan Turing decided to disregard all these questions
而言他关心一个更简单的问题:
in favor of a much simpler one:
电脑可以如人类一样地交流吗?
can a computer talk like a human?
这个问题让他有了测量人工智能的想法之后
This question led to an idea for measuring aritificial intelligence
这便演变为著名的“图灵测验”
that would famously come to be known as the Turing test.
在1950年 图灵在“计算机器与智能”的报告里
In the 1950 paper, “Computing Machinery and Intelligence,”
提出了一个游戏:
Turing proposed the following game.
一个判官和看不到的选手用短信交流
A human judge has a text conversation with unseen players
然后评价他们的回答
and evaluates their responses.
要通过考试 电脑必须能在没有改变结果的情况下
To pass the test, a computer must be able to replace one of the players
代替其中一位选手
without substantially changing the results.
换句话说 如果无轻易法分辨一台电脑与一个人的区别
In other words, a computer would be considered intelligent
这台电脑就是“聪明”的
if its conversation couldn’t be easily distinguished from a human’s.
图灵预计在2000年
Turing predicted that by the year 2000,
拥有100兆字节内存的机器会轻易地通过图灵测试
machines with 100 megabytes of memory would be able to easily pass his test.
但是图灵预计错了
But he may have jumped the gun.
虽然现代的电脑具备更多的内存
Even though today’s computers have far more memory than that,
没有几个通过了图灵测试
few have succeeded,
那些通过图灵测试的电脑
and those that have done well
并不是用了压倒性的计算能力
focused more on finding clever ways to fool judges
而是用了巧妙的手段来迷惑判官
than using overwhelming computing power.
虽然没有经过正式的考试
Though it was never subjected to a real test,
ELIZA成了历史上第一个有资格成功的程序
the first program with some claim to success was called ELIZA.
仅用了一个十分简短的脚本
With only a fairly short and simple script,
ELIZA成功地迷惑了很多人
it managed to mislead many people by mimicking a psychologist,
模仿心理专家 鼓励他们多说话
encouraging them to talk more
同时也发问他们的问题
and reflecting their own questions back at them.
另一个早期程式脚本 PARRY 运用相反的方式
Another early script PARRY took the opposite approach
模仿了偏执的精神分裂症患者
by imitating a paranoid schizophrenic
一直将话题转移回他自己预设的困扰
who kept steering the conversation back to his own preprogrammed obsessions.
它们玩弄人们的成功凸显出测试的缺点
Their success in fooling people highlighted one weakness of the test.
人类常常把很多并不聪明的事物
Humans regularly attribute intelligence to a whole range of things
归类于“聪明”
that are not actually intelligent.
尽管如此 年度竞赛比如洛伯纳奖
Nonetheless, annual competitions like the Loebner Prize,
使测试变得更规范
have made the test more formal
让判官们事先知道
with judges knowing ahead of time
有些对话选手是机器
that some of their conversation partners are machines.
虽然总体质量上升了
But while the quality has improved,
很多聊天机器人还是局用了与ELIZA和PARRY类似的技巧
many chatbot programmers have used similar strategies to ELIZA and PARRY.
1997年的获胜者
1997’s winner Catherine
凯瑟琳能够进行惊人般的聪慧会话
could carry on amazingly focused and intelligent conversation,
但是这是在讨论与比尔·克林顿
but mostly if the judge wanted to talk about Bill Clinton.
相关的话题最近的获奖者 尤金·古斯特曼
And the more recent winner Eugene Goostman
拥有13岁的乌克兰少年的个性
was given the persona of a 13-year-old Ukrainian boy,
使判官将他的别扭语法
so judges interpreted its nonsequiturs and awkward grammar
理解为语言及文化的阻碍
as language and culture barriers.
在这同时 其他的程式如 Cleverbot 采取了另一个方式
Meanwhile, other programs like Cleverbot have taken a different approach
通过分析与统计巨大的真实对话数据
by statistically analyzing huge databases of real conversations
决定最好的回答方式
to determine the best responses.
有一些还存储先前对话的记忆
Some also store memories of previous conversations
以便长期改善
in order to improve over time.
但尽管Cleverbot 自己的答复听起来非常像人类
But while Cleverbot’s individual responses can sound incredibly human,
对始终如一的个性的缺乏
its lack of a consistent personality
及无法回答崭新的话题
and inability to deal with brand new topics
则完全暴露了它
are a dead giveaway.
在图灵的时代里 谁可能预料到今日的电脑
Who in Turing’s day could have predicted that today’s computers
能够驾驶宇宙飞船
would be able to pilot spacecraft,
能操控精巧的手术
perform delicate surgeries,
还能解答大量的方程
and solve massive equations,
可仍与最基本的对话斗争?
but still struggle with the most basic small talk?
人类语言是样如何巧妙及复杂的现象
Human language turns out to be an amazingly complex phenomenon
连最大最广泛的字典都
that can’t be captured by even the largest dictionary.
无法记录 聊天机器人会被简单的停顿 如“额…”
Chatbots can be baffled by simple pauses, like “umm…”
或被没有正确回答的问题而弄得团团转
or questions with no correct answer.
一个简单的对话句子
And a simple conversational sentence,
比如“我从冰箱里拿出了果汁
like, “I took the juice out of the fridge and gave it to him,
然后给了他 但忘了查明日子
but forgot to check the date,”
“需要丰富的潜在知识与直觉来解析
requires a wealth of underlying knowledge and intuition to parse.
事件证明若要模仿人类对话
It turns out that simulating a human conversation
比单纯地增加记忆力和运算量
takes more than just increasing memory and processing power,
更加复杂在接近图灵的目标的同时
and as we get closer to Turing’s goal,
我们也终究需考虑下与“意识”相关的重大问题
we may have to deal with all those big questions about consciousness after all.

发表评论

译制信息
视频概述
听录译者

收集自网络

翻译译者

收集自网络

审核员

自动通过审核

视频来源

https://www.youtube.com/watch?v=3wLqsRLvV-c

相关推荐