“There are some things which cannot be learned quickly, and time, which is all we have, must be paid heavily for their acquiring. They are the very simplest things, and because it takes a man’s life to know them the little new that each man gets from life is very costly and the only heritage he has to leave.” - Ernest Hemingway (More…)
News #
I will be updating both good news, bad news and all kinds of news.
Posts #
Mathematics Lecture Notes
Mathematics MOOCS
Pre-print articles on Difference-of-Convex (DC) Programming
57. Stochastic Difference-of-Convex Optimization with Momentum # Authors: El Mahdi Chayti, Martin Jaggi Abstract: Stochastic difference-of-convex (DC) optimization is prevalent in numerous machine learning applications, yet its convergence properties under small batch sizes remain poorly understood. Existing methods typically require large batches or strong noise assumptions, which limit their practical use. In this work, we show that momentum enables convergence under standard smoothness and bounded variance assumptions (of the concave part) for any batch size. We prove that without momentum, convergence may fail regardless of stepsize, highlighting its necessity. Our momentum-based algorithm achieves provable convergence and demonstrates strong empirical performance.
Recent Advanced in Research on Difference-of-Convex (DC) Programming
Second-order Stochastic Optimization methods for Machine Learning
Analysis of the Hessian # 1. Empirical Analysis of the Hessian of Over-Parametrized Neural Networks # Year: 2017 Authors: Levent Sagun, Utku Evci, V. Ugur Guney, Yann Dauphin, Leon Bottou ArXiv ID: arXiv:1706.04454 URL: https://arxiv.org/abs/1706.04454 Abstract: We study the properties of common loss surfaces through their Hessian matrix. In particular, in the context of deep learning, we empirically show that the spectrum of the Hessian is composed of two parts: (1) the bulk centered near zero, (2) and outliers away from the bulk. We present numerical evidence and mathematical justifications to the following conjectures laid out by Sagun et al. (2016): Fixing data, increasing the number of parameters merely scales the bulk of the spectrum; fixing the dimension and changing the data (for instance adding more clusters or making the data less separable) only affects the outliers. We believe that our observations have striking implications for non-convex optimization in high dimensions. First, the flatness of such landscapes (which can be measured by the singularity of the Hessian) implies that classical notions of basins of attraction may be quite misleading. And that the discussion of wide/narrow basins may be in need of a new perspective around over-parametrization and redundancy that are able to create large connected components at the bottom of the landscape. Second, the dependence of small number of large eigenvalues to the data distribution can be linked to the spectrum of the covariance matrix of gradients of model outputs. With this in mind, we may reevaluate the connections within the data-architecture-algorithm framework of a model, hoping that it would shed light into the geometry of high-dimensional and non-convex spaces in modern applications. In particular, we present a case that links the two observations: small and large batch gradient descent appear to converge to different basins of attraction but we show that they are in fact connected through their flat region and so belong to the same basin.