arXiv:2506.21306v2 Announce Type: replace-cross Abstract: Functions that grow without bound on one side of the real line and decay to zero on the other cannot be…
arXiv:2607.07538v2 Announce Type: replace-cross Abstract: Training a model with noisy gradient descent can be idealized as overdamped Langevin dynamics, and a natural safety question is to…
arXiv:2607.13749v1 Announce Type: cross Abstract: Neural networks trained on modular arithmetic exhibit grokking, a delayed transition from memorisation to generalisation known to depend on model capacity:…
arXiv:2607.10936v2 Announce Type: replace-cross Abstract: We study the bandit-feedback version of online principal component analysis (Bandit PCA): in each round $t = 1,dots,T$, the adversary selects…
arXiv:2510.19283v2 Announce Type: replace-cross Abstract: We present a systematic analysis of estimation errors for a class of optimal transport based algorithms for filtering and data assimilation.…
arXiv:2603.21393v2 Announce Type: replace-cross Abstract: The widespread use of AI and ML models in sensitive areas raises significant concerns about fairness. While the research community has…
arXiv:2607.13841v1 Announce Type: cross Abstract: Heavy-tailed data arise in many domains where rare events carry disproportionate importance, such as imbalanced image datasets, financial returns, and weather…
arXiv:2606.11283v2 Announce Type: replace-cross Abstract: We study the problem of generating synthetic data under differential privacy. We establish fixed-parameter tractability (FPT) for this problem where the…
arXiv:2607.13883v1 Announce Type: cross Abstract: We formalize verification in causal graphical models: deciding whether a given observational formula identifies a target interventional distribution. This opens a…
arXiv:2602.10691v2 Announce Type: replace Abstract: We study the slice-matching scheme, an efficient iterative method for distribution matching based on sliced optimal transport. We investigate convergence to…