Unique on Facebook: Formulation and Evidence of (Nano)targeting Individual Users with non-PII Data · HackerPapers discussions

Unique on Facebook: Formulation and Evidence of (Nano)targeting Individual Users with non-PII Data — HN submission

Tue, 19 Oct 2021 07:41:49 GMT

Unique on Facebook: (Nano)Targeting Individual Users with Non-PII Data

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

Unique on Facebook: Formulation and Evidence of (Nano)targeting Individual Users with non-PII Data — mentioned by nemoniac

Sun, 17 Oct 2021 15:04:44 GMT

Reading the article on which this is based (https://arxiv.org/abs/2110.06636), the researchers don't appear to specify how they randomly select the 25 interests from the interests assigned to a user. This could have a considerable effect on their results, leading to a number somewhat higher than 22. The random selection could be uniform or it could be weighted by the global popularity of the interests. Presumably it's uniform. But if so, you could challenge the reasonableness of that assumption since it would be more likely to know a more common interest of a user that one in the tail of the…

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

Unique on Facebook: Formulation and Evidence of (Nano)targeting Individual Users with non-PII Data — mentioned by yellow_postit

Sat, 16 Oct 2021 17:37:58 GMT

Here’s the link to the paper on arXiv: https://arxiv.org/abs/2110.06636 as I didn’t find it in the TechCrunch article.

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

Unique on Facebook: Formulation and Evidence of (Nano)targeting Individual Users with non-PII Data — HN submission

Thu, 14 Oct 2021 15:31:57 GMT

Unique on Facebook: Formulation and Evidence of (Nano)Targeting Individual Users

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