{"type":"ScholarlyArticle","title":"FLASH: Flexible Learning of Adaptive Sampling from History in Temporal Graph Neural Networks","authors":[{"name":"Or Feldman","slug":"or-feldman","given":"Or","family":"Feldman","affiliation":""},{"name":"Krishna Sri Ipsit Mantri","slug":"krishna-sri-ipsit-mantri","given":"Krishna Sri Ipsit","family":"Mantri","affiliation":""},{"name":"Carola-Bibiane Schönlieb","slug":"carola-bibiane-schonlieb","given":"Carola-Bibiane","family":"Schönlieb","affiliation":""},{"name":"Chaim Baskin","slug":"chaim-baskin","given":"Chaim","family":"Baskin","affiliation":""},{"name":"Moshe Eliasof","slug":"moshe-eliasof","given":"Moshe","family":"Eliasof","affiliation":""}],"abstract":"Aggregating temporal signals from historic interactions is a key step in future link prediction on dynamic graphs. However, incorporating long histories is resource-intensive. Hence, temporal graph neural networks (TGNNs) often rely on historical neighbors sampling heuristics such as uniform sampling or recent neighbors selection. These heuristics are static and fail to adapt to the underlying graph structure. We introduce FLASH, a learnable and graph-adaptive neighborhood selection mechanism that generalizes existing heuristics. FLASH integrates seamlessly into TGNNs and is trained end-to-end using a self-supervised ranking loss. We provide theoretical evidence that commonly used heuristics hinder TGNNs performance, motivating our design. Extensive experiments across multiple benchmarks demonstrate consistent and significant performance improvements for TGNNs equipped with FLASH","doi":"10.24963/ijcai.2026/476","url":"https://ijcai.org/proceedings/2026/476","pdf":"https://ijcai.org/proceedings/2026/0476.pdf","markdown":"https://ijcai.org/proceedings/2026/476.md","bibtex":"https://ijcai.org/proceedings/2026/bibtex/476","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":"Machine Learning","track":"Main Track","number":476,"pages":{"first":4278,"last":4285},"keywords":["Machine Learning: Deep learning architectures","Machine Learning: Geometric learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}