Beyond Isolated Clients: Integrating Graph-Based Embeddings into Event Sequence Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915929850380288 |
|---|---|
| author | Proshian, Harry Severin, Nikita Nikolenko, Sergey Ivan, Kireev Savchenko, Andrey Sergeev, Ivan Postnova, Maria Makarov, Ilya |
| author_facet | Proshian, Harry Severin, Nikita Nikolenko, Sergey Ivan, Kireev Savchenko, Andrey Sergeev, Ivan Postnova, Maria Makarov, Ilya |
| contents | Large-scale digital platforms generate billions of timestamped user-item interactions (events) that are crucial for predicting user attributes in, e.g., fraud prevention and recommendations. While self-supervised learning (SSL) effectively models the temporal order of events, it typically overlooks the global structure of the user-item interaction graph. To bridge this gap, we propose three model-agnostic strategies for integrating this structural information into contrastive SSL: enriching event embeddings, aligning client representations with graph embeddings, and adding a structural pretext task. Experiments on four financial and e-commerce datasets demonstrate that our approach consistently improves the accuracy (up to a 2.3% AUC) and reveals that graph density is a key factor in selecting the optimal integration strategy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_09085 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Beyond Isolated Clients: Integrating Graph-Based Embeddings into Event Sequence Models Proshian, Harry Severin, Nikita Nikolenko, Sergey Ivan, Kireev Savchenko, Andrey Sergeev, Ivan Postnova, Maria Makarov, Ilya Machine Learning Artificial Intelligence Large-scale digital platforms generate billions of timestamped user-item interactions (events) that are crucial for predicting user attributes in, e.g., fraud prevention and recommendations. While self-supervised learning (SSL) effectively models the temporal order of events, it typically overlooks the global structure of the user-item interaction graph. To bridge this gap, we propose three model-agnostic strategies for integrating this structural information into contrastive SSL: enriching event embeddings, aligning client representations with graph embeddings, and adding a structural pretext task. Experiments on four financial and e-commerce datasets demonstrate that our approach consistently improves the accuracy (up to a 2.3% AUC) and reveals that graph density is a key factor in selecting the optimal integration strategy. |
| title | Beyond Isolated Clients: Integrating Graph-Based Embeddings into Event Sequence Models |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2604.09085 |