Date/Time
Date(s) - 15/02/2024
4:30 pm - 5:30 pm

Location: BSB B142

Speaker: Anastasis Kratsios

Title: Digital Computers Break The Curse of Dimensionality: Adaptive Bounds via Discrete Geometry

Abstract: Many of the foundations of machine learning rely on the idealized premise that all input and output spaces are infinite, e.g.~R^d. This core assumption is systematically violated in practice due to digital computing limitations from finite machine precision, rounding, and limited RAM.  In short, digital computers operate on finite grids in R^d, not on all of R^d.  By exploiting these discrete structures, we show the curse of dimensionality in statistical learning is systematically broken when models are implemented on real computers.  Consequentially, we obtain new generalization bounds with dimension-free rates for kernel and deep ReLU MLP regressors, which are implemented on real-world machines.

Our results are derived using a new non-asymptotic concentration of measure result between a probability measure over any finite metric space and its empirical version associated with N i.i.d. samples when measured in the 1-Wasserstein distance.  Unlike standard concentration of measure results, the concentration rates in our bounds do not hold uniformly for all sample sizes $N$; instead, our rates can adapt to any given $N$.  This yields significantly tighter bounds for realistic sample sizes while achieving the optimal worst-case rate of O(N^{-1/2})$ for massive.  Our results are built on new techniques combining metric embedding theory with optimal transport.

Joint work with: Martina Neumann (TU Vienna) and Gudmund Pammer (ETH Zürich)