Google's artificial intelligence software when shown a photo of a scene from an ultimate Frisbee game – a disc is thrown by a player who is being marked with a forehand force, the software responded with the description: A group of young people playing a game of frisbee. Compare the computer model to a human's depiction: A group of men playing Frisbee in the park. [Here's a hypothetical guess on how ESPN3 would probably describe the scene: An entertaining game.]
 |
| screengrab via NYT |
When the two networks - process images into a mathematical representation of their content and generate full English sentences as part of automated translation software - are combined, the first can "look" at an image and then feed the mathematical description of what it "sees" into the second, which uses that information to generate a human-readable sentence. The combined network was trained to generate more accurate descriptions by showing it tens of thousands of images with descriptions written by humans.
More Google resources toward games of Ultimate (Frisbee), please!
No comments:
Post a Comment