Storytelling software learns how to tell a good tale
"MY, WHAT a big mouth you have, Grandma," says Little Red Riding Hood, with just a hint of suspicion. The wolf sneezes. "Bless you," says the little girl. Sound odd? That's because this snippet of Little Red Riding Hood was written not by a person but by a piece of software called Xapagy.
Hal Hodson
Picture
AI creation or just riffs on an old story?
Mines the a new software development in artificial intelligence (AI) called Xapagy, a computer that is fed coded stories, which the the software remembers. The code is a hybrid of English and computer language called Xapi.

Notes that Xapagy is a significant move towards AI being capable of unique storytelling because the tales, though manually fed, are not broken down by specific rules and computer logic. Instead, the software learns more grammar and vocabulary, and each new story remains whole within the memory.

The system, then, is more flexible than previous attempts, and when producing stories, it first matches the closest equivalents from its memory, but fills in grammatical language where there are gaps in weaving a story together. A story such as the one above, is a significant move toward an AI with the ability to invent unique, creative stories. 

The breakthrough for AI experts is that the computer generates language similar to human processes, not by following strict rules and algorithms, but more organically. The stories are remembered by the computer just as they are (in Xapi of course), as a whole story beginning to end rather than breaking it down into rigid, mechanical pieces.

When learning new stories, the software actually perceives the intersections with other tales in its memory. When it can find connections, Xapagy uses the previous experience to predict what will occur in the rest of the story. In this way, the computer is adding new material to the story based upon predictions and memories. This is the closest they have come to instilling creativity in AI.

When Xapagy is confronted with missing words, it fills them in with its own language based upon what makes grammatical and contextual sense.  Researchers in AI see this as a tremendous advance, and with enough stories in its memory, they believe Xapagy will be able to generate unique stories of its own.

Darin's note: It's difficult to see, at first, the tremendous abilities of this computer because our own language is so complex and Xapagy inferior. However, if you think about AI being in its infant stage, literally, like a human baby learning to tell stories, the science is more profound.

I believe this is the way scientists view AI, as analogous to human evolution and growth. With each small advance they make, they are maturing the artificial brain, just as a baby's brain learns, grows, matures, and creates. With this analogy in mind, the AI baby is growing up really fast.

Interested? Click the title or image to read on.

Source is NewScientist.com

You might also enjoy:

 


Comments




Leave a Reply

    ZipMinis Extras!
    Like ZipMinis? Please let us know below.



    On This Page

    ZipMinis by Topic

    All
    Amazon Video
    Apple
    Artificial Intelligence
    Batman Slayings
    Black Friday
    Brain
    Chaos
    Cloud
    Cognitive Science
    Computer Crash
    Computer Meltdown
    Curation
    Design
    Dreams
    Education
    Emotional Intelligence
    Empathy
    Evernote
    Facebook
    Genius
    Gooogle
    Introverts
    Inventions
    Ipad
    Iphone
    Kindle
    Marissa Mayer
    Materialism
    Mobile
    Mobile Advertising
    Monopoly
    Mountain Lion
    Psychological Help
    Psychology Technology
    Rage Against The Machine
    Reading
    Social Media
    Startups
    Supervillains
    Technology Trash
    Twitter
    Value
    Video
    Video Games
    Writing Books

    Author

    Darin L. Hammond

    ZipMinis by Date

    June 2013
    May 2013
    April 2013
    March 2013
    February 2013
    January 2013
    December 2012
    November 2012
    October 2012
    September 2012
    August 2012
    July 2012