Robust Vertex Enumeration for Convex Hulls in High Dimensions

02/05/2018
by   Pranjal Awasthi, et al.
0

Computation of the vertices of the convex hull of a set S of n points in R ^m is a fundamental problem in computational geometry, optimization, machine learning and more. We present "All Vertex Triangle Algorithm" (AVTA), a robust and efficient algorithm for computing the subset S of all K vertices of conv(S), the convex hull of S. If Γ_* is the minimum of the distances from each vertex to the convex hull of the remaining vertices, given any γ≤γ_* = Γ_*/R, R the diameter of S, AVTA computes S in O(nK(m+ γ^-2)) operations. If γ_* is unknown but K is known, AVTA computes S in O(nK(m+ γ_*^-2)) (γ_*^-1) operations. More generally, given t ∈ (0,1), AVTA computes a subset S^t of S in O(n | S^t|(m+ t^-2)) operations, where the distance between any p ∈ conv(S) to conv( S^t) is at most t R. Next we consider AVTA where input is S_ε, an ε perturbation of S. Assuming a bound on ε in terms of the minimum of the distances of vertices of conv(S) to the convex hull of the remaining point of S, we derive analogous complexity bounds for computing S_ε. We also analyze AVTA under random projections of S or S_ε. Finally, via AVTA we design new practical algorithms for two popular machine learning problems: topic modeling and non-negative matrix factorization. For topic models AVTA leads to significantly better reconstruction of the topic-word matrix than state of the art approaches arora2013practical, bansal2014provable. For non-negative matrix AVTA is competitive with existing methods arora2012computing. Empirically AVTA is robust and can handle larger amounts of noise than existing methods.

READ FULL TEXT
research
12/06/2017

An Efficient Algorithm for Non-Negative Matrix Factorization with Random Projections

Non-negative matrix factorization (NMF) is one of the most popular decom...
research
04/14/2019

Finding a latent k-simplex in O(k . nnz(data)) time via Subset Smoothing

The core problem in many Latent Variable Models, widely used in Unsuperv...
research
10/11/2021

Smoothed Separable Nonnegative Matrix Factorization

Given a set of data points belonging to the convex hull of a set of vert...
research
05/08/2017

Non-negative Matrix Factorization via Archetypal Analysis

Given a collection of data points, non-negative matrix factorization (NM...
research
08/09/2017

Maximum Volume Inscribed Ellipsoid: A New Simplex-Structured Matrix Factorization Framework via Facet Enumeration and Convex Optimization

Consider a structured matrix factorization scenario where one factor is ...
research
04/27/2020

A Geometric Algorithm for Solving Linear Systems

Based on the geometric Triangle Algorithm for testing membership of a p...
research
10/17/2018

An O(1/ε)-Iteration Triangle Algorithm for A Convex Hull Membership

A fundamental problem in linear programming, machine learning, and compu...

Please sign up or login with your details

Forgot password? Click here to reset