Correcting Exposure Bias for Link Recommendation

Shantanu Gupta, Hao Wang, Zachary Lipton, Yuyang Wang
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:3953-3963, 2021.

Abstract

Link prediction methods are frequently applied in recommender systems, e.g., to suggest citations for academic papers or friends in social networks. However, exposure bias can arise when users are systematically underexposed to certain relevant items. For example, in citation networks, authors might be more likely to encounter papers from their own field and thus cite them preferentially. This bias can propagate through naively trained link predictors, leading to both biased evaluation and high generalization error (as assessed by true relevance). Moreover, this bias can be exacerbated by feedback loops. We propose estimators that leverage known exposure probabilities to mitigate this bias and consequent feedback loops. Next, we provide a loss function for learning the exposure probabilities from data. Finally, experiments on semi-synthetic data based on real-world citation networks, show that our methods reliably identify (truly) relevant citations. Additionally, our methods lead to greater diversity in the recommended papers’ fields of study. The code is available at github.com/shantanu95/exposure-bias-link-rec.

Cite this Paper

BibTeX
@InProceedings{pmlr-v139-gupta21c, title = {Correcting Exposure Bias for Link Recommendation}, author = {Gupta, Shantanu and Wang, Hao and Lipton, Zachary and Wang, Yuyang}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {3953--3963}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/gupta21c/gupta21c.pdf}, url = {https://proceedings.mlr.press/v139/gupta21c.html}, abstract = {Link prediction methods are frequently applied in recommender systems, e.g., to suggest citations for academic papers or friends in social networks. However, exposure bias can arise when users are systematically underexposed to certain relevant items. For example, in citation networks, authors might be more likely to encounter papers from their own field and thus cite them preferentially. This bias can propagate through naively trained link predictors, leading to both biased evaluation and high generalization error (as assessed by true relevance). Moreover, this bias can be exacerbated by feedback loops. We propose estimators that leverage known exposure probabilities to mitigate this bias and consequent feedback loops. Next, we provide a loss function for learning the exposure probabilities from data. Finally, experiments on semi-synthetic data based on real-world citation networks, show that our methods reliably identify (truly) relevant citations. Additionally, our methods lead to greater diversity in the recommended papers’ fields of study. The code is available at github.com/shantanu95/exposure-bias-link-rec.} }
Endnote
%0 Conference Paper %T Correcting Exposure Bias for Link Recommendation %A Shantanu Gupta %A Hao Wang %A Zachary Lipton %A Yuyang Wang %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-gupta21c %I PMLR %P 3953--3963 %U https://proceedings.mlr.press/v139/gupta21c.html %V 139 %X Link prediction methods are frequently applied in recommender systems, e.g., to suggest citations for academic papers or friends in social networks. However, exposure bias can arise when users are systematically underexposed to certain relevant items. For example, in citation networks, authors might be more likely to encounter papers from their own field and thus cite them preferentially. This bias can propagate through naively trained link predictors, leading to both biased evaluation and high generalization error (as assessed by true relevance). Moreover, this bias can be exacerbated by feedback loops. We propose estimators that leverage known exposure probabilities to mitigate this bias and consequent feedback loops. Next, we provide a loss function for learning the exposure probabilities from data. Finally, experiments on semi-synthetic data based on real-world citation networks, show that our methods reliably identify (truly) relevant citations. Additionally, our methods lead to greater diversity in the recommended papers’ fields of study. The code is available at github.com/shantanu95/exposure-bias-link-rec.
APA
Gupta, S., Wang, H., Lipton, Z. & Wang, Y.. (2021). Correcting Exposure Bias for Link Recommendation. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:3953-3963 Available from https://proceedings.mlr.press/v139/gupta21c.html.

Related Material