Statistical Learning
The Lingering of Gradients: How to Reuse Gradients Over Time
Zeyuan Allen-Zhu, David Simchi-Levi, Xinshang Wang
Classically, the time complexity of a first-order method is estimated by its number of gradient computations. In this paper, we study a more refined complexity by taking into account the "lingering" of gradients: once a gradient is computed at x
A Structured Prediction Approach for Label Ranking
Anna Korba, Alexandre Garcia, Florence d'Alché-Buc
Neural Information Processing Systems http://nips.cc/
A theory on the absence of spurious solutions for nonconvex and nonsmooth optimization
Cedric Josz, Yi Ouyang, Richard Zhang, Javad Lavaei, Somayeh Sojoudi
We study the set of continuous functions that admit no spurious local optima (i.e.
Statistical mechanics of low-rank tensor decomposition
Jonathan Kadmon, Surya Ganguli
More complex data consisting of measurements across multiple modalities can be organized as higher dimensional data arrays, or higher order tensors.
Reparameterization Gradient for Non-differentiable Models
Wonyeol Lee, Hangyeol Yu, Hongseok Yang
Neural Information Processing Systems http://nips.cc/
Online Adaptive Methods, Universality and Acceleration
Kfir Y. Levy, Alp Yurtsever, Volkan Cevher
In this work we propose such a method.
Fast Estimation of Causal Interactions using Wold Processes
Flavio Figueiredo, Guilherme Resende Borges, Pedro O.S. Vaz de Melo, Renato Assunção
The main reason to consider the Wold process as an alternative to the Hawkes process is its Markovian structure. Equally important, our results show significant improvements over the state of the art methods.
Dual Principal Component Pursuit: Improved Analysis and Efficient Algorithms
Zhihui Zhu, Yifan Wang, Daniel Robinson, Daniel Naiman, René Vidal, Manolis Tsakiris
The classical formulation of PCA, dating back to Carl F. Gauss, is based on minimizing the sum of
Masking: A New Perspective of Noisy Supervision
Bo Han, Jiangchao Yao, Gang Niu, Mingyuan Zhou, Ivor Tsang, Ya Zhang, Masashi Sugiyama
Neural Information Processing Systems http://nips.cc/
Stochastic Expectation Maximization with Variance Reduction
However, sEM has a slow asymptotic convergence rate due to the high variance of each update.