Simple recency and popularity memory heuristics match or beat neural temporal graph models on TGB and BenchTemp, but only after per-dataset heuristic selection that is not disclosed clearly.
Do we really need complicated model architectures for temporal networks? In The Eleventh International Conference on Learning Representations, 2023
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
On the Power of Heuristics in Temporal Graphs
Simple recency and popularity memory heuristics match or beat neural temporal graph models on TGB and BenchTemp, but only after per-dataset heuristic selection that is not disclosed clearly.