Country
Convergence of Cubic Regularization for Nonconvex Optimization under KL Property
Yi Zhou, Zhe Wang, Yingbin Liang
Cubic-regularized Newton's method (CR) is a popular algorithm that guarantees
Fixing the train-test resolution discrepancy
Hugo Touvron, Andrea Vedaldi, Matthijs Douze, Herve Jegou
Neural Information Processing Systems http://nips.cc/
Sparsified SGD with Memory
Sebastian U. Stich, Jean-Baptiste Cordonnier, Martin Jaggi
Several papers consider approaches that limit the number of bits to represent floating point numbers [13, 24, 31]. Recent work proposes adaptive tuning of the compression ratio [7].
Splatter a Video: Video Gaussian Representation for Versatile Processing Y ang-Tian Sun 1 Yi-Hua Huang 1 Lin Ma Xiaoyang Lyu 1 Y an-Pei Cao 2
Existing research on video representation for processing has primarily focused on 2D/2.5D
b43a6403c17870707ca3c44984a2da22-Paper.pdf
We perform computational experiments that support our theoretical results.
Locally Private Learning without Interaction Requires Separation
Neural Information Processing Systems http://nips.cc/
Corruption-Robust Offline Reinforcement Learning with General Function Approximation
These authors contributed equally to this work.
Discovering Neural Wirings
Mitchell Wortsman, Ali Farhadi, Mohammad Rastegari
Figure 2: Anexampleofadynamic (left) andstatic (right) neuralgraph. Claim: Assume L is Lipschitzcontinuous.