DP-VAE: Human-Readable Text Anonymization for Online Reviews with Differentially Private Variational Autoencoders

N/A

DP-VAE: Human-Readable Text Anonymization for Online Reviews with Differentially Private Variational Autoencoders Benjamin Weggenmann∗ Valentin Rublack∗† Michael Andrejczuk∗† SAP Security Research SAP Security Research University of Edinburgh Karlsruhe, Germany Karlsruhe, Germany Edinburgh, Scotland…

saved by

• Daniel R

related reading

• [1906.02691] An Introduction to Variational Autoencodersarxiv.org
• Difference between AutoEncoder (AE) and Variational AutoEncoder (VAE) | Towards Data Sciencetowardsdatascience.com
• 2409.02908arxiv.org
• Tutorial - What is a variational autoencoder? – Jaan Lı 李jaan.io
• Sparse Autoencoders Find Highly Interpretable Features in Language Modelsarxiv.org
• A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Modelsarxiv.org
• 1706.03762arxiv.org
• Tutorial on Variational Autoencodersarxiv.org
• [2602.16800] Large-scale online deanonymization with LLMsarxiv.org
• Diffusion Models as a kind of VAE | Angus Turnerangusturner.github.io
• [2602.23163] A Decision-Theoretic Formalisation of Steganography With Applications to LLM Monitoringarxiv.org
• arxiv.org/pdf/2511.08544arxiv.org