Breaking Feedback Loops in Recommender Systems with Causal Inference

Krauth, Karl; Wang, Yixin; Jordan, Michael I.

doi:10.48550/arxiv.2207.01616

Cited by 4 publications

(6 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to debiasing, some IPS-based methods are dedicated to addressing other issues that abound in RS. 20 , 83 , 84 For example, Mehrotra et al. 20 propose an unbiased estimator of user satisfaction based on IPS to jointly optimize for supplier fairness and consumer relevance.…”

Section: Po-based Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A survey on causal inference for recommendation

Luo,

Zhuang,

Xie

et al. 2024

The Innovation

View full text Add to dashboard Cite

Section: Po-based Methodsmentioning

confidence: 99%

“…Besides, the CBDF (counterfactual bandit with delayed feedback) 83 re-weights the observed feedback with importance sampling, which is determined by a survival model to deal with delayed feedback. The CAFL (causal adjustment for feedback loops) 84 extends the IPS estimator to break feedback loops.…”

Section: Po-based Methodsmentioning

confidence: 99%

A survey on causal inference for recommendation

Luo,

Zhuang,

Xie

et al. 2024

The Innovation

View full text Add to dashboard Cite

“…• Training a recommendation model under an unfair and biased empirical distribution of the exposures p (𝑜 𝑖 |z 𝑢 , z 𝑖 ) instead of 𝑝 uni (𝑜 𝑖 |z 𝑢 , z 𝑖 ) would inevitably result in biased relevance scores due the feedback loop [17,30], i.e. unfair up or down weighting for specific items.…”

Section: Balanced Fairness Objectivementioning

confidence: 99%

“…The unfairness problem from training recommendation models on biased logged data is well-known [24,30]. Building a model based on data that was generated by one's own recommendation systems will inevitably create a feedback loop which amplify biases [9,17] and where the true relevance score is harder to determine.…”

Section: Introductionmentioning

confidence: 99%

Rectifying Unfairness in Recommendation Feedback Loop

Yang

Wang

Ton³

2023

Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval

View full text Add to dashboard Cite

The issue of fairness in recommendation systems has recently become a matter of growing concern for both the academic and industrial sectors due to the potential for bias in machine learning models. One such bias is that of feedback loops, where the collection of data from an unfair online system hinders the accurate evaluation of the relevance scores between users and items. Given that recommendation systems often recommend popular content and vendors, the underlying relevance scores between users and items may not be accurately represented in the training data. Hence, this creates a feedback loop in which the user is not longer recommended based on their true relevance score but instead based on biased training data. To address this problem of feedback loops, we propose a two-stage representation learning framework, B-FAIR, aimed at rectifying the unfairness caused by biased historical data in recommendation systems. The framework disentangles the context data into sensitive and non-sensitive components using a variational autoencoder and then applies a novel Balanced Fairness Objective (BFO) to remove bias in the observational data when training a recommendation model. The efficacy of B-FAIR is demonstrated through experiments on both synthetic and real-world benchmarks, showing improved performance over state-of-the-art algorithms.

show abstract

“…User interactions generate data, and therefore promoting items influences users' choices. This interaction between the user and RS is a closed feedback loop since historical interactions are the backbone of updating or retraining the mathematical model [65]. Promoting random items affects this feedback loop.…”

Section: Introductionmentioning

confidence: 99%

Data-Driven Strategies in Hospitality

van Leeuwen

View full text Add to dashboard Cite

This dissertation focuses on data-driven strategies for hospitality to increase the adaptability of predictive and prescriptive models. The study emphasizes the incorporation of business context and the utilization of a white-box approach. Diverse methods are applied and examined in this dissertation, and their main contributions are introduced. The research aims to bolster the effectiveness and transparency of these models, ensuring they align with the specific needs and challenges faced by the hospitality sector.

show abstract

Breaking Feedback Loops in Recommender Systems with Causal Inference

Cited by 4 publications

References 0 publications

A survey on causal inference for recommendation

A survey on causal inference for recommendation

Rectifying Unfairness in Recommendation Feedback Loop

Data-Driven Strategies in Hospitality

Contact Info

Product

Resources

About