SoylentNews is people
SoylentNews
New research has shown just how bad AI is at dealing with online trolls.
Such systems struggle to automatically flag nudity and violence, don’t understand text well enough to shoot down fake news and aren’t effective at detecting abusive comments from trolls hiding behind their keyboards.
A group of researchers from Aalto University and the University of Padua found this out when they tested seven state-of-the-art models used to detect hate speech. All of them failed to recognize foul language when subtle changes were made, according to a paper [PDF] on arXiv.
Adversarial examples can be created automatically by using algorithms to misspell certain words, swap characters for numbers or add random spaces between words or attach innocuous words such as ‘love’ in sentences.
The models failed to pick up on adversarial examples and successfully evaded detection. These tricks wouldn’t fool humans, but machine learning models are easily blindsighted. They can’t readily adapt to new information beyond what’s been spoonfed to them during the training process.
(Score: 2) by requerdanos on Saturday September 01 2018, @02:18PM (5 children)
False premise:
Supported but poorly by clumsy strawman:
I want the people whose speech I don't want to see in a certain place (which you inaccurately call "anything deemed objectionable") to put the speech somewhere else. There's room for everybody; just because you are all holy-roller on wanting people to see objectionable things doesn't give them the right to clutter up any random (or specific third-party) forum or news feed with them.
There is a big difference between finding something objectionable because I disagree with it (such as your nonsense here) and finding something objectionable because it's inappropriate for the context (such as the "déck herders we dón't type no young words"* and "murder snuff spam" we've seen on discussions in this site).
Your opinions here, though vapid, advance discussion, and so contribute something positive, however slight.
The spams I mention, here, and false or invented stories on a supposedly trustworthy news site, or hateful attacks in a community discussion forum, don't.
-------
* Mildly disappointed that this didn't get flagged as a spam pattern. Poor AI!
(Score: -1, Flamebait) by Anonymous Coward on Saturday September 01 2018, @02:41PM (2 children)
1) You are wrong.
1) a) You are an arrogant prick.
2) when the time comes it will be my pleasure to fight against people like you and wipe you off the face of the earth in order to preserve freedom.
(Score: 2) by requerdanos on Saturday September 01 2018, @02:50PM
The irony-troll force is strong with this one. :)
(Score: 0) by Anonymous Coward on Sunday September 02 2018, @01:30AM
"2) when the time comes it will be my pleasure to fight against people like you and wipe you off the face of the earth in order to preserve freedom."
.
That time has passed and you did nothing. Now live with the shame for your entire life!
(Score: 1) by realDonaldTrump on Saturday September 01 2018, @09:47PM (1 child)
When you disagree you say, "mod parent to oblivion!"
(Score: 2) by requerdanos on Saturday September 01 2018, @11:23PM
You've picked the wrong side: When I disagree I reply, or let it go. When it's junk that clutters up the thread, I mod down; when just spam, I mod spam. For repeat offender trolls that others are overly tempted to feed, you'll get a call to mod parent to oblivion.
Thanks for noticing.