Georgia Institute of Technology College of Sciences

I describe my ongoing work using tools from computational and combinatorial algebraic geometry to classify minimal problems and identify which can be solved efficiently. I will not assume any background in algebraic geometry or computer vision.

Structure-from-motion algorithms reconstruct a 3D scene from many images, often by matching features (such as point and lines) between the images. Matchings lead to constraints, resulting in a nonlinear system of polynomial equations that recovers the 3D geometry. Since many matches are outliers, these methods are used in an iterative framework for robust estimation called RANSAC (RAndom SAmpling And Consensus), whose efficiency hinges on using a small number of correspondences in each iteration. As a result, there is a big focus on constructing polynomial solvers for these "minimal problems" that run as fast as possible. Our work classifies these problems in cases of practical interest (calibrated cameras, complete and partial visibility.) Moreover, we identify candidates for practical use, as quantified by "algebraic complexity measures" (degree, Galois group.)

joint w/ Anton Leykin, Kathlen Kohn, Tomas Pajdla arxiv1903.10008 arxiv2003.05015+ Viktor Korotynskiy, TP, and Margaret Regan (ongoing.)

Optimization and machine learning algorithms often use real-world data that has been generated through complex socio-economic and behavioral processes. This data, however, is noisy, and naturally encodes difficult-to-quantify systemic biases. In this work, we model and address bias in the secretary problem, which has applications in hiring. We assume that utilities of candidates are scaled by unknown bias factors, perhaps depending on demographic information, and show that bias-agnostic algorithms are suboptimal in terms of utility and fairness. We propose bias-aware algorithms that achieve certain notions of fairness, while achieving order-optimal competitive ratios in several settings.

Motivated by the Dikin walk, we develop aspects of an interior-point

theory for sampling in high dimensions. Specifically, we introduce symmetric

and strong self-concordance. These properties imply that the corresponding

Dikin walk mixes in $\tilde{O}(n\bar{\nu})$ steps from a warm start

in a convex body in $\mathbb{R}^{n}$ using a strongly self-concordant barrier

with symmetric self-concordance parameter $\bar{\nu}$. For many natural

barriers, $\bar{\nu}$ is roughly bounded by $\nu$, the standard

self-concordance parameter. We show that this property and strong

self-concordance hold for the Lee-Sidford barrier. As a consequence,

we obtain the first walk to mix in $\tilde{O}(n^{2})$ steps for an

arbitrary polytope in $\mathbb{R}^{n}$. Strong self-concordance for other

barriers leads to an interesting (and unexpected) connection ---

for the universal and entropic barriers, it is implied by the KLS

conjecture.

We give an efficient algorithm for robustly clustering of a mixture of two arbitrary Gaussians, a central open problem in the theory of computationally efficient robust estimation, assuming only that the the means of the component Gaussian are well-separated or their covariances are well-separated. Our algorithm and analysis extend naturally to robustly clustering mixtures of well-separated logconcave distributions. The mean separation required is close to the smallest possible to guarantee that most of the measure of the component Gaussians can be separated by some hyperplane (for covariances, it is the same condition in the second degree polynomial kernel). Our main tools are a new identifiability criterion based on isotropic position, and a corresponding Sum-of-Squares convex programming relaxation.