# Debiasing the Cloze Task in Sequential Recommendation with Bidirectional Transformers

arXiv: https://arxiv.org/abs/2301.09210 · HTML: https://mlsanity.com/paper/2301.09210/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 42 references that are in the map.
- First posted on arXiv 22 Jan 2023.

## Builds on (most cited first, 23 of 23)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation](https://mlsanity.com/paper/1406.1078/) (2014)
- [Self-Attentive Sequential Recommendation](https://mlsanity.com/paper/1808.09781/) (2018)
- [BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer](https://mlsanity.com/paper/1904.06690/) (2019)
- [Image-based Recommendations on Styles and Substitutes](https://mlsanity.com/paper/1506.04757/) (2015)
- [Session-based Recommendations with Recurrent Neural Networks](https://mlsanity.com/paper/1511.06939/) (2015)
- [Sequence Level Training with Recurrent Neural Networks](https://mlsanity.com/paper/1511.06732/) (2015)
- [Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding](https://mlsanity.com/paper/1809.07426/) (2018)
- [Recommendations as Treatments: Debiasing Learning and Evaluation](https://mlsanity.com/paper/1602.05352/) (2016)
- [Bias and Debias in Recommender System: A Survey and Future Directions](https://mlsanity.com/paper/2010.03240/) (2020)
- [A Critical Review of Recurrent Neural Networks for Sequence Learning](https://mlsanity.com/paper/1506.00019/) (2015)
- [Recurrent Neural Networks with Top-k Gains for Session-based Recommendations](https://mlsanity.com/paper/1706.03847/) (2017)
- [Fast Matrix Factorization for Online Recommendation with Implicit Feedback](https://mlsanity.com/paper/1708.05024/) (2017)
- [A Survey on Session-based Recommender Systems](https://mlsanity.com/paper/1902.04864/) (2019)
- [Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedback](https://mlsanity.com/paper/1909.03601/) (2019)
- [Modeling User Exposure in Recommendation](https://mlsanity.com/paper/1510.07025/) (2015)
- [Reinforced Negative Sampling over Knowledge Graph for Recommendation](https://mlsanity.com/paper/2003.05753/) (2020)
- [Sequential Recommendation with Self-Attentive Multi-Adversarial Network](https://mlsanity.com/paper/2005.10602/) (2020)
- [Fast Adaptively Weighted Matrix Factorization for Recommendation with Implicit Feedback](https://mlsanity.com/paper/2003.01892/) (2020)
- [Dynamic Matrix Factorization with Priors on Unknown Values](https://mlsanity.com/paper/1507.06452/) (2015)
- [Exploring Longitudinal Effects of Session-based Recommendations](https://mlsanity.com/paper/2008.07226/) (2020)
- [Debiased Explainable Pairwise Ranking from Implicit Feedback](https://mlsanity.com/paper/2107.14768/) (2021)

## Built on by (most cited first, 3 of 3)

- [Sequential Recommendation with Diffusion Models](https://mlsanity.com/paper/2304.04541/) (2023)
- [Deep Stable Multi-Interest Learning for Out-of-distribution Sequential Recommendation](https://mlsanity.com/paper/2304.05615/) (2023)
- [Facets of Disparate Impact: Evaluating Legally Consistent Bias in Machine Learning](https://mlsanity.com/paper/2505.05471/) (2025)

## Abstract (arXiv)

Bidirectional Transformer architectures are state-of-the-art sequential recommendation models that use a bi-directional representation capacity based on the Cloze task, a.k.a. Masked Language Modeling. The latter aims to predict randomly masked items within the sequence. Because they assume that the true interacted item is the most relevant one, an exposure bias results, where non-interacted items with low exposure propensities are assumed to be irrelevant. The most common approach to mitigating exposure bias in recommendation has been Inverse Propensity Scoring (IPS), which consists of down-weighting the interacted predictions in the loss function in proportion to their propensities of exposure, yielding a theoretically unbiased learning. In this work, we argue and prove that IPS does not extend to sequential recommendation because it fails to account for the temporal nature of the problem. We then propose a novel propensity scoring mechanism, which can theoretically debias the Cloze task in sequential recommendation. Finally we empirically demonstrate the debiasing capabilities of our proposed approach and its robustness to the severity of exposure bias.