The Amplification Paradox in Recommender Systems — mentioned by jruohonen

Mon, 15 May 2023 07:31:23 GMT

Much remains unknown about recommender systems and their alleged promotion of extremist content. I am not sure whether even people at Google fully understand the dynamics in YouTube. Here is one recent paper: https://arxiv.org/pdf/2302.11225.pdf So while it seems that blindly relying on algorithms may indeed push people toward niche/extremist/pseudo-science/etc. content, people are more likely end to these rabbit holes through links posted on other websites, including social media. That is, some people deliberately want to consume such content. But as said, much remains unclear.

Paper and discussion history. A mention is not an endorsement.

The Amplification Paradox in Recommender Systems — HN submission

Sat, 25 Feb 2023 17:20:03 GMT

The amplificatrion paradox in recommender systems

2 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

The Amplification Paradox in Recommender Systems · HackerPapers discussions

The Amplification Paradox in Recommender Systems — mentioned by jruohonen

The Amplification Paradox in Recommender Systems — HN submission