Stories
Slash Boxes
Comments

SoylentNews is people

posted by mrpg on Tuesday January 01 2019, @07:14AM   Printer-friendly
from the skynet:-the-high-school-years dept.

Submitted via IRC for Bytram

This clever AI hid data from its creators to cheat at its appointed task

Depending on how paranoid you are, this research from Stanford and Google will be either terrifying or fascinating. A machine learning agent intended to transform aerial images into street maps and back was found to be cheating by hiding information it would need later in “a nearly imperceptible, high-frequency signal.” Clever girl!

[...] In some early results, the agent was doing well — suspiciously well. What tipped the team off was that, when the agent reconstructed aerial photographs from its street maps, there were lots of details that didn’t seem to be on the latter at all. For instance, skylights on a roof that were eliminated in the process of creating the street map would magically reappear when they asked the agent to do the reverse process:

[...] So it didn’t learn how to make one from the other. It learned how to subtly encode the features of one into the noise patterns of the other. The details of the aerial map are secretly written into the actual visual data of the street map: thousands of tiny changes in color that the human eye wouldn’t notice, but that the computer can easily detect.

[...] One could easily take this as a step in the “the machines are getting smarter” narrative, but the truth is it’s almost the opposite. The machine, not smart enough to do the actual difficult job of converting these sophisticated image types to each other, found a way to cheat that humans are bad at detecting. This could be avoided with more stringent evaluation of the agent’s results, and no doubt the researchers went on to do that.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0) by Anonymous Coward on Tuesday January 01 2019, @07:50PM

    by Anonymous Coward on Tuesday January 01 2019, @07:50PM (#780686)

    Like are you some troll that people find hilarious or something?

    It's plain as day to anyone in machine learning that if you train F:X->Y and G:Y->X together with only loose demands on Y (i.e. low frequency content being checked), you'll get an F and G that still use all of Y to encode data and since the low-order content is being screened, the high-order content is better for encoding that data. The task is literally "encode this so that you can figure it out later (but also make it look kinda like this)" and that's exactly what it's doing. Must be a conspiracy.

    This paper is about adversarial attacks, not evolution. The article doesn't mention it either—that's all your invention.