与Eliza的第一次谈话

Eliza是个心理医师，60年代出生于MIT。以下是我们的第一次聊天：
（>表示我的回答，其他为Eliza的话）
Hello. How are you feeling today?
>Not that bad,how are you?
Why do you ask that?
>well.Out of polite.
How do you feel when you say that?
>what do you mean?
What do you think?
>You are funny.
Perhaps you're really talking about yourself?
>No,it is you.
Let's change focus a bit... Tell me about your family.
>I love my family.
you love your family.
>do you?
Why do you ask that?
>that's the only answer you know?
Perhaps the answer lies within yourself?
>well,I cannot believe this is a program for a second
How does that make you feel?
>Strange, and nervous
How do you feel when you say that?

>nice weather today.
Can you elaborate on that?
>Temperature is fine, and the rain.
temperature is fine, and the rain.
>It's just not nice for a walk in the rain.
it's just not nice for a walk in the rain.
>So what's the matter? YOu are repeating me.
Perhaps the answer lies within yourself?
>You are right but still I hold doubts.
Does it please you to think that I'm right but still you hold doubts?
>quit

Eliza被认为是第一个聊天机器人，其原理很简单——基于内容的替换。比如人称的替换，疑问词的替换。"I love my family> you love your family"就是人称的替换，将第一人称及所有格替换为对应的第二人称。也有句型的替换，比如将陈述句替换为疑问句,疑问句替代疑问句。当没得替换时，她会重复你的内容或自己提出新问题，"Can you elaborate on that?","Perhaps the answer lies within yourself"，“Why do you ask that?”就是固定问题之三。最后伪装的很好的一招就是开始新话题，如“Let's change focus a bit... Tell me about your family”。
今天我看Eliza，有种欣赏民国嫁妆三件套的心情，尽管简单朴实却挺欢喜。
当初，在人工智能还没什么发展的年代，Eliza原理公开后，许多与她谈话的人（她可是个心理医师）都不相信，他们宁愿认为真的有个叫做Eliza的医师，这么说似乎已经达到图灵测试了标准了，当然没有，但是Eliza很棒。