The topic is that poems written by Google's artificial intelligence are cowardly

ByVictoria Nevland

Google develops AI (artificial intelligence) that speaks natural language using deep learning, but poem actually spelled out in the research has a unique atmosphere and even a little disgusting It is talked about.

"I want to talk to you": See the creepy, romantic poetry that came out of a Google AI system - Quartz
http://qz.com/682814/i-want-to-talk-to-you-see-the-creepy-romantic-poetry-that-came-out-of-a-google-ai-system/


An example of poem generated by AI is kore. Although it has some atmosphere when it looks like it, in fact it seems to be said that it is quite surreal content.

"I want to talk to you."
"I want to be with you."
"I do not want to be with you."
I do not want to be with you.
She did not want to be with him.

(Translation)
"I want to talk with you"
"I want to stay with you"
"I do not want to be with you"
I do not want to be with you.
She did not want to be with him.

What a strange poem, this is Google's AI development project team "Google BrainWhat AI made of. The team is made up of linguists and computer scientists and seems to have AI loading 11,000 unpublished books for learning. Three hundred thousand romance series works and 1500 fantasy works are included in books, and by learning from them, an attempt to reproduce the network (neural network) of the same cranial nerve cells as human beings on a computer It is done.

Based on the content learned as described above, AI reads his poems, but the instruction given at that time is "I'm fine (I am OK)" to "But you need to talk I'd like to write 13 lines of poetry connecting two of "to me now." In accordance with this instruction, AI creates poetry that is clean from start to goal, each sentence is connected properly with the front and back, and grammatical and syntactic problems are also created by himself.

The ultimate goal is to create completely original poetry by "thinking" by myself based on what you learned, but at this stage it is quite an interesting work created as follows It seems to be. As mentioned earlier, it is said that instructions are given to designate the first and last lines, and to make a poem that connects between them.

No.
He said.
"No," he said.
"No," i said.
"I know," she said.
"Thank you," she said.
"Come with me," she said.
"Talk to me," she said.
"Do not worry about it," she said.

(Translation)
No.
He said.
"No," he said.
"No," I said.
"I know," she said.
"Thank you," she said.
"Come with us," she said.
"Talk," she said.
"You do not have to worry about that thing," she said.

It made me want to cry.
No one had seen him since.
It made me feel uneasy.
No one had seen him.
The thought made me smile.
The pain was unbearable.
The crowd was silent.
The man called out.
The old man said.
The man asked.

(Translation)
I felt like crying.
Ever since that nobody has seen him.
I felt uneasy.
Nobody is looking at him.
The idea that crossed my head made me smile.
I can not put up with this suffering.
A group of people was silent.
A man shouted.
The old man's man said.
The man asked.

He was silent for a long moment.
He was silent for a moment.
It was quiet for a moment.
It was dark and cold.
There was a pause.
It was my turn.

(Translation)
He kept silent for a long time.
I kept silent for a while.
For a while, it was quiet.
It was dark and cold.
There was a moment.
It's my turn.

There is no one else in the world.
There is no one else in sight.
They were the ones who mattered.
They were the only ones left.
He had to be with me.
She had to be with him.
I had to do this.
I wanted to kill him.
I started to cry.
I turned to him.

(Translation)
There is no one else in the world.
There is no one else in sight.
Only they are important to me.
Only they were left.
He must stay with me.
She must stay with him.
I have to do this.
I wanted to kill him.
I began to cry.
I turned toward him.

It was a number of works by AI who spontaneously blew out and felt the spirit freezing in fear. The research team held an international conference "From May 2 to May 2016"International Conference on Learning Representations"We have published research content at the time, you can read the report from the following pages.

(PDF)Generating Sentences from a Continuous Space