8 ms·
Welcome to the spammer vs machine learning arms race, Wikipedia. :-) (Former YouTube Abuse team member)
by zellyn 11y ago
Welcome to the spammer vs machine learning arms race, Wikipedia. :-)
(Former YouTube Abuse team member)
- MollyR 11y agoCan't spammers also use machine learning, or maybe I misunderstood something.
- zellyn 11y agoWhat I meant is that I expect a huge initial win, as this catches 80-90% of spammy edits, until the spammers figure out how to get around it. Hopefully the constant stream of human-created training data (in the form of reverted or non-reverted edits) helps with this. But undoubtedly there will need to be some continual feature engineering to stay ahead.
- halfak_wmf 11y agoYeah. We're actively working with patrollers to chase down mistakes and retrain the models. This is a system that will require continual improvement. Luckily, there's a lot of researchers working on this problem so we're drawing from a lot of brilliance beyond our little team at the Wikimedia Foundation.
- kuschku 11y agoWell, Wikipedia at least has humans working on it, who can catch issues and improve it. It can’t be worse than YouTube/Google’s systems anyway.
- titanomachy 11y agoI don't see that many bot-generated comments anymore on YouTube. The genuine user-generated comments are among the lowest-quality on any site that I frequent, but I suspect there isn't much that Google can do about that.
- kuschku 11y agoI’ve seen tons of spam, "buy viagra here", and I’ve seen hundreds of videos that have just an empty image, no sound, and a full-screen annotation (and a description) that link to another video site.
- Coding_Cat 11y agoI see plenty, all following a very simple formula & mostly spoofing the original uploader. And those comments stay around for quite a while. They really need to improve the system. Not that there is much value in the YT comments, as you say, I just think it reflects badly on Google that they can't keep these comments out.