Tag
This paper introduces specified-foil counterfactuals for temporal graphs, a method to find low-cost past-event interventions that lead to a specific alternative prediction, evaluated on continuous-time dynamic graphs and temporal knowledge graphs.
BERTilda is an explainable framework that tracks topic lifecycles in longitudinal text streams by constructing temporal graphs with similarity and flow signals to detect splits, merges, and other transitions, achieving high agreement rates on annotated datasets.
This paper presents a benchmark for predicting institutional equity holdings using temporal graph machine learning, framing it as node affinity prediction on a bipartite graph. The proposed NAVIS model achieves state-of-the-art NDCG of 0.9127 on a dataset from SEC Form 13F filings.
This paper explores predicting whether Lightning Network channels will close mutually or via forced closure using machine learning on gossip data. An MLP with temporal features outperforms graph-based models, and the dataset is publicly released.