{"type":"ScholarlyArticle","title":"Learnable Data Augmentation and Contrastive Pre-training for Temporal Link Prediction","authors":[{"name":"Canghong Jin","slug":"canghong-jin","given":"Canghong","family":"Jin","affiliation":""},{"name":"Jiafeng Zhao","slug":"jiafeng-zhao","given":"Jiafeng","family":"Zhao","affiliation":""},{"name":"Feng Xu","slug":"feng-xu","given":"Feng","family":"Xu","affiliation":""},{"name":"Tongya Zheng","slug":"tongya-zheng","given":"Tongya","family":"Zheng","affiliation":""},{"name":"Zemin Liu","slug":"zemin-liu","given":"Zemin","family":"Liu","affiliation":""},{"name":"Lina Wei","slug":"lina-wei","given":"Lina","family":"Wei","affiliation":""},{"name":"Mingli Song","slug":"mingli-song","given":"Mingli","family":"Song","affiliation":""}],"abstract":"Link prediction is a foundational task in temporal graphs. While temporal graph neural networks exhibit commendable performance, they are often criticized for providing inadequate representations, especially under limited data. Contrastive learning has been introduced as a solution for graph pre-training to mitigate data scarcity. However, the data augmentation techniques on which they depend frequently lead to suboptimal augmentations, manifesting either as over-augmentation or under-augmentation.\r\nIn light of these challenges, we explore the under-explored domain of contrastive learning of temporal graph transformers and propose a novel model, ContraTGT, which employs a dual-view graph transformer. This transformer is fine-tuned to extract sequences tailored for specific target nodes, encapsulating both spatial and temporal perspectives. To optimize the contrastive learning of this dual-view transformer, we put forth an innovative, learnable data augmentation method. This technique, which involves selective masking of elements within dual-view sequences, generates superior augmentations, thereby amplifying the potency of the contrastive learning approach.\r\nExtensive experiments on five public temporal network datasets demonstrate that our model can consistently outperform all baselines, especially in small training set conditions.","doi":"10.24963/ijcai.2026/303","url":"https://ijcai.org/proceedings/2026/303","pdf":"https://ijcai.org/proceedings/2026/0303.pdf","markdown":"https://ijcai.org/proceedings/2026/303.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/303","volume":{"slug":"2026","title":"Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2026","url":"https://ijcai.org/proceedings/2026/","year":2026,"location":"Bremen, Germany","dates":"15-21 August 2026","editors":"Diego Calvanese","isbn":"978-1-956792-09-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2026"},"section":"Data Mining","track":"Main Track","number":303,"pages":{"first":2726,"last":2734},"keywords":["Data Mining: Mining graphs","Data Mining: Mining heterogenous data"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}