On the Robustness of the Successive Projection Algorithm
This work addresses noise robustness in a foundational algorithm for latent simplex learning, with incremental improvements in error bounds for specific scenarios.
The paper revisits the robustness of the Successive Projection Algorithm (SPA) and its variants to noise, proving tightness of existing error bounds for certain cases and providing improved bounds by factors related to vertex conditioning, with a new variant proposed to minimize conditioning.
The successive projection algorithm (SPA) is a workhorse algorithm to learn the $r$ vertices of the convex hull of a set of $(r-1)$-dimensional data points, a.k.a. a latent simplex, which has numerous applications in data science. In this paper, we revisit the robustness to noise of SPA and several of its variants. In particular, when $r \geq 3$, we prove the tightness of the existing error bounds for SPA and for two more robust preconditioned variants of SPA. We also provide significantly improved error bounds for SPA, by a factor proportional to the conditioning of the $r$ vertices, in two special cases: for the first extracted vertex, and when $r \leq 2$. We then provide further improvements for the error bounds of a translated version of SPA proposed by Arora et al. (''A practical algorithm for topic modeling with provable guarantees'', ICML, 2013) in two special cases: for the first two extracted vertices, and when $r \leq 3$. Finally, we propose a new more robust variant of SPA that first shifts and lifts the data points in order to minimize the conditioning of the problem. We illustrate our results on synthetic data.