Differentially-private-topic-mining

This repository is the implementation of A Model-Agnostic Approach to Differentially Private Topic Mining (KDD 2022)

Usage

First, please run the file model_creation.py to generate the topics of documents

Second, please runt the file sensitivity_sampler_user_list.py to get the smooth sensitivity

Third, please run the noise_generation_distance.py to generate the noisy matrix

Citation

@inproceedings{wang2022model, title={A model-agnostic approach to differentially private topic mining}, author={Wang, Han and Sharma, Jayashree and Feng, Shuya and Shu, Kai and Hong, Yuan}, booktitle={Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining}, pages={1835--1845}, year={2022} }