Evaluating Recommender System Stability with Influence-Guided Fuzzing
Abstract
Recommender systems help users to find products or services they may like when lacking personal experience or facing an overwhelming set of choices. Since unstable recommendations can lead to distrust, loss of profits, and a poor user experience, it is important to test recommender system stability. In this work, we present an approach based on inferred models of influence that underlie recommender systems to guide the generation of dataset modifications to assess a recommender’s stability. We implement our approach and evaluate it on several recommender algorithms using the MovieLens dataset. We find that influence-guided fuzzing can effectively find small sets of modifications that cause significantly more instability than random approaches.
Cite
Text
Shriver et al. "Evaluating Recommender System Stability with Influence-Guided Fuzzing." AAAI Conference on Artificial Intelligence, 2019. doi:10.1609/AAAI.V33I01.33014934Markdown
[Shriver et al. "Evaluating Recommender System Stability with Influence-Guided Fuzzing." AAAI Conference on Artificial Intelligence, 2019.](https://mlanthology.org/aaai/2019/shriver2019aaai-evaluating/) doi:10.1609/AAAI.V33I01.33014934BibTeX
@inproceedings{shriver2019aaai-evaluating,
title = {{Evaluating Recommender System Stability with Influence-Guided Fuzzing}},
author = {Shriver, David and Elbaum, Sebastian G. and Dwyer, Matthew B. and Rosenblum, David S.},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2019},
pages = {4934-4942},
doi = {10.1609/AAAI.V33I01.33014934},
url = {https://mlanthology.org/aaai/2019/shriver2019aaai-evaluating/}
}