/SBO

This repository contains the source code for the paper: "Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models" by Gustavo Escobedo, Marta Moscati, Peter Muellner, Simone Kopeinik, Dominik Kowald, Elisabeth Lex, Markus Schedl, ECML-PKDD, 2024 Research Track

Primary LanguagePython

Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models

This repository contains the source code for paper:

Escobedo, G., Moscati, M., Müllner, P., Kopeinik, S., Kowald, D., Lex, E., and Schedl M. Making Alice Appear Like Bob: A Probabilistic Preference Obfuscation Method For Implicit Feedback Recommendation Models, Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2024), Vilnius, Lithuania, September 2024.

Installation and configuration

# Create environment
conda env create -f environment.yml
# Activate environment
conda activate sbo

Data preparation

  • Set the root directory for all datasets by changing the variable the variable ROOT_DIR_STR in constants.py. The downloaded datasets should be in the folder ROOT_DIR_STR (i.e. ../<ROOT_DIR_STR>/ml-1m/..).
  • Modify the obfuscation parameters in constants.py, an obfuscated dataset will be generated for each combination of these parameters inside the previously set root directory. Moreover, all the datasets are in RecBole's atomic format.
# Pre-process and generate standard format
python preprocess.py

Run

  • Execute the following command to generate the obfuscated datasets this will generate several parallel processes for generating the different datasets.
# Generate the obfuscated dataset files
python run_obfuscation.py
  • To train the attacker networks on the obfuscated datasets run/adapt the script scripts\run_train_attack.sh.
cd scripts
sh run_train_attack.sh <DATASET_CONFIG> <DATASETS_FILES> 

# Example
# sh run_train_attack.sh ml-1m ../datasets_files/datasets_sample.json
  • To train the RecBole models on the obfuscated datasets run/adapt the script scripts\run_recbole.sh. The script executes pool_recs.py code which is meant to run multiple jobs. Several csv files will be generated containing test results of the trained models. Additionally, given that our datasets are RecBole ready, individual models can be trained following the documentation
cd scripts
sh run_recbole.sh <DATASETS_FILES> <RECOMMENDER_MODEL> 

# Example
# sh run_recbole.sh ../datasets_files/datasets_sample.json BPR

Acknowledgement

This research was funded in whole or in part by the FFG COMET center program, by the Austrian Science Fund (FWF): P36413, P33526, and DFH-23, and by the State of Upper Austria and the Federal Ministry of Education, Science, and Research, through grant LIT-2021-YOU-215 and LIT-2020-9-SEE-113.