Education
No-regret Online Learning over Riemannian Manifolds
We establish upper bounds on the regrets of the problem with respect to time horizon, manifold curvature, and manifold dimension.
2 ProblemFormulationandRelatedWorks Intypical graph learning problems, givenasetoflabeled graphsDtr ={(Gi,yi)}N
In this paper, we first mathematically formulate thetwochallenging problems forgraph data andtake an initiative on tackling them under a unified probabilistic model.
Uncertainty-based Continual Learning with Adaptive Regularization
Hongjoon Ahn, Sungmin Cha, Donggyu Lee, Taesup Moon
Neural Information Processing Systems http://nips.cc/
Dataset Decomposition: Faster LLM Training with Variable Sequence Length Curriculum
Large language models (LLMs) are commonly trained on datasets consisting of fixed-length token sequences.
Multistep Distillation of Diffusion Models via Moment Matching
We present a new method for making diffusion models faster to sample.