27
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 11 Sep 2023
27 points (88.6% liked)
Fediverse
17788 readers
2 users here now
A community dedicated to fediverse news and discussion.
Fediverse is a portmanteau of "federation" and "universe".
Getting started on Fediverse;
- What is the fediverse?
- Fediverse Platforms
- How to run your own community
founded 5 years ago
MODERATORS
I think it will be pretty much I possible to create an unbiased recommendation algorithm without creating echo chambers.
Do you mean "impossible"?
I think it doesn't have to be. The crucial thing is that you have the option to simply change instances if you don't like the algorithm that comes with it. The scholar Katharina Zweig introduced danger-levels of algorithms. The danger level is determined by A.) whether you can easily get an alternative for the given algorithm without investing too much and B.) which has low negative impact in case the algorithm is bad.
She would categorizes curating-algorithms in social media therefore as low as long as there are enough alternatives that you can easily switch to and I would agree with that.
Edit: The categorization was about algorithm-decision-systems but it can be applied for algorithms in general
I doubt you would find an algorithm per instance. The question is what kind of recommendations would like? What is the kind of content you want to push and how do you decide to push it. I am genuinely wondering because even when I used Twitter, I always used the chronological view so I'm wondering what you are missing.