Differentially Private Synthetic Data via Foundation Model APIs 2: Text — HN submission

Tue, 15 Apr 2025 12:34:37 GMT

Differentially Private Synthetic Data via Foundation Model APIs

3 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Differentially Private Synthetic Data via Foundation Model APIs 2: Text — mentioned by jsenn

Tue, 15 Apr 2025 12:15:42 GMT

yeah, the linked paper [1] has more detail--basically they seem to start with a seed set of "class labels" and subcategories (e.g. "restaurant review" + "steak house"). They ask an LLM to generate lots of random texts incorporating those labels. They make a differentially private histogram of embedding similarities from those texts with the private data, then use that histogram to resample the texts, which become the seeds for the next iteration, sort of like a Particle Filter. I'm still unclear on how you create that initial set of class labels used to generate the random seed texts, and how…

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

Differentially Private Synthetic Data via Foundation Model APIs 2: Text · HackerPapers discussions

Differentially Private Synthetic Data via Foundation Model APIs 2: Text — HN submission

Differentially Private Synthetic Data via Foundation Model APIs 2: Text — mentioned by jsenn