Do Loyal Users Enjoy Better Recommendations?

Ji, Yitong; Sun, Aixin; Zhang, Jie; Li, Chenliang

doi:10.1145/3539813.3545124

Cited by 5 publications

(4 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…G (6) a (6) 0 1 1 0 1 0 0 1 1 0 0 1 0 1 1 0 possible hidden causal dependencies a (6) ! observation time t 0 = 6 permutation π = (3, 2, 4, 1) undirected = true…”

Section: (A) a Universal Family Of Causal Models For Graphsmentioning

confidence: 99%

“…We show the execution of an SCM as described in electronic supplementary material, appendix C where, together with the observation parameters t 0 = 6 and π = (3, 2, 4, 1), generate the observed graph G (6) with adjacency matrix a (6) . We colour the entries in a (6) according to the corresponding variables in the unobserved SCM that generated them. Grey nodes represent observed random variables, while white nodes represent unobserved ones.…”

Section: (A) a Universal Family Of Causal Models For Graphsmentioning

confidence: 99%

“…Encouraging girls to perform spatial tasks in playtime (an intervention) could improve their spatial reasoning skills under the path-dependent hypothesis but not under an innate factors hypothesis. Similarly, performing an intervention to probe into whether a consumer will buy a computer keyboard (through an ad or a top search ranking) [4,5], who then buys it, will lead to assuming that (a) maybe the consumer no longer needs (another) keyboard under a path-dependent model; or (b) the consumer will keep buying more keyboards (if we keep showing more ads and top search results) [6] under an innate factor hypothesis. Current causal models for link prediction operate under the assumption of innate factors [4,7].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Causal lifting and link prediction

Cotta,

Bevilacqua,

Ahmed

et al. 2023

Proc. R. Soc. A.

View full text Add to dashboard Cite

Existing causal models for link prediction assume an underlying set of inherent node factors—an innate characteristic defined at the node’s birth—that governs the causal evolution of links in the graph. In some causal tasks, however, link formation is path-dependent : the outcome of link interventions depends on existing links. Unfortunately, these existing causal methods are not designed for path-dependent link formation, as the cascading functional dependencies between links (arising from path dependence ) are either unidentifiable or require an impractical number of control variables. To overcome this, we develop the first causal model capable of dealing with path dependencies in link prediction. In this work, we introduce the concept of causal lifting, an invariance in causal models of independent interest that, on graphs, allows the identification of causal link prediction queries using limited interventional data. Further, we show how structural pairwise embeddings exhibit lower bias and correctly represent the task’s causal structure, as opposed to existing node embeddings, e.g. graph neural network node embeddings and matrix factorization. Finally, we validate our theoretical findings on three scenarios for causal link prediction tasks: knowledge base completion, covariance matrix estimation and consumer-product recommendations.

show abstract

“…G (6) a (6) 0 1 1 0 1 0 0 1 1 0 0 1 0 1 1 0 possible hidden causal dependencies a (6) ! observation time t 0 = 6 permutation π = (3, 2, 4, 1) undirected = true…”

Section: (A) a Universal Family Of Causal Models For Graphsmentioning

confidence: 99%

Section: (A) a Universal Family Of Causal Models For Graphsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Causal lifting and link prediction

Cotta,

Bevilacqua,

Ahmed

et al. 2023

Proc. R. Soc. A.

View full text Add to dashboard Cite

show abstract

“…In Proceedings of the 2022 ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR'22). Association for Computing Machinery, New York, NY, USA, 92-97 [184].…”

mentioning

confidence: 99%

Impact of temporal context on recommender systems along global timeline

View full text Add to dashboard Cite

Sun Aixin, for his continuous guidance and support throughout my research journey. Prof. Sun always provides valuable insights and inspiring ideas during our discussions. He inspires me to perceive recommender system from a new perspective, and motivates critical thinking in research problems. Beyond his expertise, Prof. Sun is also a friendly and kind person. My PhD study has been interesting and fruitful because of him. Furthermore, I would like to express my appreciation to Prof. Zhang Jie, for his continuous help during my research. Prof. Zhang provides valuable feedback and constructive comments during our discussions. Moreover, he provided opportunities for me to work in real-life recommendation project, which further broadens my knowledge in my research area. Lastly, I would like to thank my parents, my husband Zhao Yue and my friends, Zhang Yixiang, Shi Jiachen and Xia Fan, my pet friends Pepper, Bun and Button for accompanying me throughout the years. Their encouragement and belief in my abilities have been a source of strength throughout my research journey.

show abstract