Statistical Learning
GADBench: Revisiting and Benchmarking Supervised Graph Anomaly Detection
We shed light on the current progress of GAD, setting a robust groundwork for subsequent investigations in this domain.
Riemannian stochastic optimization methods avoid strict saddle points
Informally, our main result may be stated as follows: Under any stochastic Riemannian Robbins-Monro method, the probability of converging to a strict saddle point (or a submanifold thereof) is zero.
Anytime Model Selection in Linear Bandits
M different samples in parallel from the environment so that the reward for each agent is realized.
S)GD over Diagonal Linear Networks Implicit Bias Large and Edge of Stability
Currently, most theoretical works on implicit regularisation have primarily focused on continuous time approximations of (S)GD where the impact of crucial hyperparameters such as the stepsize and the minibatch size are ignored. One such common simplification is to analyse gradient flow, which is a continuous time limit of GD and minibatch SGD with an infinitesimal stepsize. By definition, this analysis does not capture the effect of stepsize or stochasticity.
Rethinking Semi-Supervised Imbalanced Node Classification from Bias-Variance Decomposition
This paper introduces a new approach to address the issue of class imbalance in graph neural networks (GNNs) for learning on graph-structured data.