Facebook is rating the trustworthiness of its users on a scale from zero to one

Re:Facebook is rating the trustworthiness of its users on a ... by Tarq0n on Tuesday August 21, 2018 @11:58AM (#17810754)

Do you check your spam filter to see what it's been filtering still? Text classification has gotten very reliable over the last decade, the process just needs to be transparent.

Valve also does something similar with their anti-cheat system called overwatch. A machine learning system selects which replays are most likely to result in a conviction, then those replays are shown to human peer-reviewers who must come to a consensus on whether the replay shows cheating or not. The pre-selection stage greatly increases the efficiency of the system.

[ Parent | on HN ]

Re:Facebook is rating the trustworthiness of its users on a ... by xienze on Tuesday August 21, 2018 @12:03PM (#17810809)

There’s a difference between filtering for spam content and the much more complicated job of determining “truth”, especially given that what is “true” in political matters is highly subjective.

[ Parent | on HN ]

Re:Facebook is rating the trustworthiness of its users on a ... by nradov on Tuesday August 21, 2018 @01:36PM (#17811847)

Yes I check my spam folder frequently. The Gmail spam filter has many false positives. In the real world text classification is still too inaccurate to rely upon for anything important.

[ Parent | on HN ]