Primary supervisor

Wray Buntine

Co-supervisors

  • Le Duy Dung, VinUniversity

Bundle recommendation systems enhance user experience and increase sales by recommending a set of items as a bundle rather than individual items [1].   The understanding of items in bundles is that they should be complementary some how.  In this project, we will explore the relationship between causal reasoning on items purchased and bundles.  Causal reasoning could be used to infer if two purchases are related, and, moreover, language language models can be used to assess the plausability of this, to help create an argument using the Bradford-Hill criteria.  Some recent research has considered the use of causallity in the recommendor systems context [2,3].  This research will work jointly with VinUniversity recommender system expert Dr. Le Duy Dung and his PhD students. 

Aim/outline

  1. Develop datasets suitable for causal reasoning from purchase history and bundles.
  2. Prepare literature review.
  3. Apply advanced causal reasoning systems to bundles and purchase histories.
  4. Employ LLMs to support assessment of methods, or data augmentation.

 

URLs/references

[1]  "A Survey on Bundle Recommendation: Methods, Applications, and Challenges", ACM Computing Surveys, Volume 58, Issue 11, Sun et al. https://dl.acm.org/doi/10.1145/3802820 

[2]  "Data-Augmented Counterfactual Learning for Bundle Recommendation", Shixuan Zhu, Qi Shen, Chuan Cui, Yu Ji, Yiming Zhang, Zhenwei Dong, Zhihua Wei, Database Systems for Advanced Applications. DASFAA 2023, https://dl.acm.org/doi/10.1007/978-3-031-35415-1_22

[3]   "Bundle Recommendation with Item-Level Causation-Enhanced Multi-view Learning",  Huy-Son Nguyen, Tuan-Nghia Bui, Long-Hai Nguyen, Hung Hoang, Cam-Van Thi Nguyen, Hoang-Quynh Le, Duc-Trong, ECML-PKDD 2024, https://dl.acm.org/doi/10.1007/978-3-031-70371-3_19

Required knowledge

Knowledge of recommender systems.  Good Python programming.  Good experience with LLMs.