2 Hits in 3.1 sec

Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issue [article]

Wenjie Wang, Fuli Feng, Xiangnan He, Hanwang Zhang, Tat-Seng Chua
2021 arXiv   pre-print
In this paper, we formulate the recommendation models as a causal graph that reflects the cause-effect factors in recommendation, and address the clickbait issue by performing counterfactual inference  ...  By estimating the click likelihood of a user in the counterfactual world, we are able to reduce the direct effect of exposure features and eliminate the clickbait issue.  ...  process. • We introduce counterfactual inference into recommendation to mitigate the clickbait issue, and propose a counterfactual recommendation framework which can be applied to any recommender models  ... 
arXiv:2009.09945v3 fatcat:5m3ikhacxfci5ic3g63jatunyu

Hidden Incentives for Auto-Induced Distributional Shift [article]

David Krueger, Tegan Maharaj, Jan Leike
2020 arXiv   pre-print
To address this issue, we introduce 'unit tests' and a mitigation strategy for HI-ADS, as well as a toy environment for modelling real-world issues with HI-ADS in content recommendation, where we demonstrate  ...  We demonstrate that changes to the learning algorithm, such as the introduction of meta-learning, can cause hidden incentives for auto-induced distributional shift (HI-ADS) to be revealed.  ...  Thanks to Valentin Dalibard for help with Population Based Training, and Toby Pohlen for help with using Google infrastructure.  ... 
arXiv:2009.09153v1 fatcat:xpduninmzzdrrn5slq3lbr7jfy