My YOShInOn RSS reader uses an SBERT model for classification (will I upvote this or not?) and large-scale clustering (20 k-means clusters and show me the top N in each cluster so I get a diversity of different articles.)

For duplicate detection I am using DBSCAN

https://scikit-learn.org/stable/modules/generated/sklearn.cl...

and found some parameters where I get almost no false positives but a lot of duplicates get missed when I lowered the threshold to make clusters I started getting false positives fast. I don't find duplicates are a big problem in my system with the 110 feeds I have and the subjects I am interested in, but insofar as they are a problem there tend to be structured relationships between articles: that is, site A syndicates articles from site B but for some reason articles from site A usually get selected and site B articles don't. An article from Site A often links to one or more articles, often that I don't have a feed for, and it would be nice if the system looked at the whole constellation. Stuff like that.

Effective clustering is the really interesting technology Google News has had for a long time.

That was partially the original promise of Fever, which is the API many RSS services still support and that somehow lives on.

Nuzzle did something similar for Twitter but shut down (https://daringfireball.net/linked/2021/05/05/nuzzel).

That would be a good addition to feed readers, especially for news feeds.

You should try Scour (https://scour.ing)!

You specify your interests as free form text, it ranks articles by how closely they match, and you can consume your Scour feed as an RSS feed to read it in NNW.

Disclaimer: I’m the developer

I haven't used it much but I think Iconfactory's Tapestry[0] does some of this.

[0]: https://usetapestry.com/