arXiv subjects

Steven J. Gortler

Publications and source records attributed to Steven J. Gortler.

At least 19 recordsLinked to original sources

arXiv2026

Higher Order Rigidity and Energy

In this paper, we revisit the notion of higher-order rigidity of a bar-and-joint framework. In particular, we provide a link between the rigidity properties of a framework, and the growth order of an energy function defined on that framework. Using our approach, we propose a general definition for the rigidity order of a framework, and we show that this definition does not depend on the details of the chosen energy function. Then we show how this order can be studied using higher order derivative tests. Doing so, we obtain a new proof that the lack of a second order flex implies rigidity. Our proof relies on our construction of a fourth derivative test, which may be applied to a critical point when the second derivative test fails. We also obtain a new proof that when the dimension of non-trivial first-order flex coefficients $\p'$ equals $1$, then the lack of a $k$th order flex for some $k$ implies a framework is rigid. The higher order derivative tests that we study here may have applications in more general optimization problems.

math.MG↗

arXiv2026

Maximum likelihood thresholds of generic linear concentration models

The maximum likelihood threshold of a statistical model is the minimum number of datapoints required to fit the model via maximum likelihood estimation. In this paper we determine the maximum likelihood thresholds of generic linear concentration models. This turns out to be the number that one might expect from a naive dimension count, which is nontrivial to prove given that the maximum likelihood threshold is a semi-algebraic concept. We also describe geometrically how a linear concentration model can fail to exhibit this generic behavior.

math.ST↗

arXiv2024

Reconstruction in one dimension from unlabeled Euclidean lengths

Let $G$ be a $3$-connected ordered graph with $n$ vertices and $m$ edges. Let $\mathbf{p}$ be a randomly chosen mapping of these $n$ vertices to the integer range $\{1, 2,3, \ldots, 2^b\}$ for $b\ge m^2$. Let $\ell$ be the vector of $m$ Euclidean lengths of $G$'s edges under $\mathbf{p}$. In this paper, we show that, with high probability over $\mathbf{p}$, we can efficiently reconstruct both $G$ and $\mathbf{p}$ from $\ell$. This reconstruction problem is NP-HARD in the worst case, even if both $G$ and $\ell$ are given. We also show that our results stand in the presence of small amounts of error in $\ell$, and in the real setting, with sufficiently accurate length measurements. Our method combines lattice reduction, which has previously been used to solve random subset sum problems, with an algorithm of Seymour that can efficiently reconstruct an ordered graph given an independence oracle for its matroid.

math.MG↗

arXiv2024

The Stress-Flex Conjecture

Recently, it has been proven that a tensegrity framework that arises from coning the one-skeleton of a convex polytope is rigid. Since such frameworks are not always infinitesimally rigid, this leaves open the question as to whether they are at least prestress stable. We prove here that this holds subject to an intriguing new conjecture about coned polytope frameworks, that we call the stress-flex conjecture. Multiple numerical experiments suggest that this conjecture is true, and most surprisingly, seems to hold even beyond convexity and also for higher genus~polytopes.

math.CO↗

arXiv2024

Complete Neural Networks for Complete Euclidean Graphs

Neural networks for point clouds, which respect their natural invariance to permutation and rigid motion, have enjoyed recent success in modeling geometric phenomena, from molecular dynamics to recommender systems. Yet, to date, no model with polynomial complexity is known to be complete, that is, able to distinguish between any pair of non-isomorphic point clouds. We fill this theoretical gap by showing that point clouds can be completely determined, up to permutation and rigid motion, by applying the 3-WL graph isomorphism test to the point cloud's centralized Gram matrix. Moreover, we formulate an Euclidean variant of the 2-WL test and show that it is also sufficient to achieve completeness. We then show how our complete Euclidean WL tests can be simulated by an Euclidean graph neural network of moderate size and demonstrate their separation capability on highly symmetrical point clouds.

cs.LG↗

arXiv2023

Maximum likelihood thresholds via graph rigidity

The maximum likelihood threshold (MLT) of a graph $G$ is the minimum number of samples to almost surely guarantee existence of the maximum likelihood estimate in the corresponding Gaussian graphical model. We give a new characterization of the MLT in terms of rigidity-theoretic properties of $G$ and use this characterization to give new combinatorial lower bounds on the MLT of any graph. We use the new lower bounds to give high-probability guarantees on the maximum likelihood thresholds of sparse Erd{ö}s-Rényi random graphs in terms of their average density. These examples show that the new lower bounds are within a polylog factor of tight, where, on the same graph families, all known lower bounds are trivial. Based on computational experiments made possible by our methods, we conjecture that the MLT of an Erd{ö}s-Rényi random graph is equal to its generic completion rank with high probability. Using structural results on rigid graphs in low dimension, we can prove the conjecture for graphs with MLT at most $4$ and describe the threshold probability for the MLT to switch from $3$ to $4$. We also give a geometric characterization of the MLT of a graph in terms of a new "lifting" problem for frameworks that is interesting in its own right. The lifting perspective yields a new connection between the weak MLT (where the maximum likelihood estimate exists only with positive probability) and the classical Hadwiger-Nelson problem.

math.CO↗

arXiv2023

Low Dimensional Invariant Embeddings for Universal Geometric Learning

This paper studies separating invariants: mappings on $D$ dimensional domains which are invariant to an appropriate group action, and which separate orbits. The motivation for this study comes from the usefulness of separating invariants in proving universality of equivariant neural network architectures. We observe that in several cases the cardinality of separating invariants proposed in the machine learning literature is much larger than the dimension $D$. As a result, the theoretical universal constructions based on these separating invariants is unrealistically large. Our goal in this paper is to resolve this issue. We show that when a continuous family of semi-algebraic separating invariants is available, separation can be obtained by randomly selecting $2D+1 $ of these invariants. We apply this methodology to obtain an efficient scheme for computing separating invariants for several classical group actions which have been studied in the invariant learning literature. Examples include matrix multiplication actions on point clouds by permutations, rotations, and various other linear groups. Often the requirement of invariant separation is relaxed and only generic separation is required. In this case, we show that only $D+1$ invariants are required. More importantly, generic invariants are often significantly easier to compute, as we illustrate by discussing generic and full separation for weighted graphs. Finally we outline an approach for proving that separating invariants can be constructed also when the random parameters have finite precision.

cs.LG↗

arXiv2023

Neural Injective Functions for Multisets, Measures and Graphs via a Finite Witness Theorem

Injective multiset functions have a key role in the theoretical study of machine learning on multisets and graphs. Yet, there remains a gap between the provably injective multiset functions considered in theory, which typically rely on polynomial moments, and the multiset functions used in practice, which rely on $\textit{neural moments}$ $\unicode{x2014}$ whose injectivity on multisets has not been studied to date. In this paper, we bridge this gap by showing that moments of neural networks do define injective multiset functions, provided that an analytic non-polynomial activation is used. The number of moments required by our theory is optimal essentially up to a multiplicative factor of two. To prove this result, we state and prove a $\textit{finite witness theorem}$, which is of independent interest. As a corollary to our main theorem, we derive new approximation results for functions on multisets and measures, and new separation results for graph neural networks. We also provide two negative results: (1) moments of piecewise-linear neural networks cannot be injective multiset functions; and (2) even when moment-based multiset functions are injective, they can never be bi-Lipschitz.

cs.LG↗

arXiv2023

Trilateration using Unlabeled Path or Loop Lengths

Let $\mathbf{p}$ be a configuration of $n$ points in $\mathbb{R}^d$ for some $n$ and some $d \ge 2$. Each pair of points defines an edge, which has a Euclideanlength in the configuration. A path is an ordered sequence of the points, and a loop is a path that begins and ends at the same point. A path or loop, as a sequence of edges, also has a Euclidean length, which is simply the sum of its Euclidean edge lengths. We are interested in reconstructing $\mathbf{p}$ given a set of edge, path and loop lengths. In particular, we consider the unlabeled setting where the lengths are given simply as a set of real numbers, and are not labeled with the combinatorial data describing which paths or loops gave rise to these lengths. In this paper, we study the question of when $\mathbf{p}$ will be uniquely determined (up to an unknowable Euclidean transform) from some given set of path or loop lengths through an exhaustive trilateration process. Such a process has already been used for the simpler problem of reconstruction using unlabeled edge lengths. This paper also provides a complete proof that this process must work in that edge-setting when given a sufficiently rich set of edge measurements and assuming that $\mathbf{p}$ is generic.

math.MG↗

arXiv2023

Proving the Existence of a GOR Without Probability

In this note, we provide a new proof that a $D$-connected graph $G$ on $n$ vertices has a general position orthogonal representation in $\RR^{n-D}$. Our argument, while based on many of the concepts from the original proof due to Lovász, Saks and Schrijver, does not use the probabilistic method.

math.CO↗

arXiv2023

General position stresses

Let $G$ be a graph with $n$ vertices, and $d$ be a target dimension. In this paper we study the set of rank $n-d-1$ matrices that are equilibrium stress matrices for at least one (unspecified) $d$-dimensional framework of $G$ in general position. In particular, we show that this set is algebraically irreducible. Likewise, we show that the set of frameworks with such equilibrium stress matrices is irreducible. As an application, this leads to a new and direct proof that every generically globally rigid graph has a generic framework that is universally rigid.

math.MG↗

arXiv2022

$K_{5,5}$ is fully reconstructible in $\mathbb{C}^3$

A graph $G$ is fully reconstructible in $\mathbb{C}^d$ if the graph is determined from its $d$-dimensional measurement variety. The full reconstructibility problem has been solved for $d=1$ and $d=2$. For $d=3$, some necessary and some sufficient conditions are known and $K_{5,5}$ falls squarely within the gap in the theory. In this paper, we show that $K_{5,5}$ is fully reconstructible in $\mathbb{C}^3$.

math.MG↗

arXiv2022

Computing maximum likelihood thresholds using graph rigidity

math.CO↗

arXiv2022

Universal Rigidity of Ladders on the line

In "Universal rigidity on the line, point orde" it is shown, answering a question of Jordán and Nguyen, that universal rigidity of a generic bar-joint framework in R^1 depends on more than the ordering of the vertices. The graph G that was used in that paper is a ladder with three rungs. Here we provide a general answer when that ladder with three rungs in the line is universally rigid and when it is not.

math.MG↗

arXiv2022

Globally rigid graphs are fully reconstructible

A $d$-dimensional framework is a pair $(G,p)$, where $G=(V,E)$ is a graph and $p$ is a map from $V$ to $\mathbb{R}^d$. The length of an edge $uv\in E$ in $(G,p)$ is the distance between $p(u)$ and $p(v)$. The framework is said to be globally rigid in $\mathbb{R}^d$ if the graph $G$ and its edge lengths uniquely determine $(G,p)$, up to congruence. A graph $G$ is called globally rigid in $\mathbb{R}^d$ if every $d$-dimensional generic framework $(G,p)$ is globally rigid. In this paper, we consider the problem of reconstructing a graph from the set of edge lengths arising from a generic framework. Roughly speaking, a graph $G$ is strongly reconstructible in $\mathbb{C}^d$ if the set of (unlabeled) edge lengths of any generic framework $(G,p)$ in $d$-space, along with the number of vertices of $G$, uniquely determine both $G$ and the association between the edges of $G$ and the set of edge lengths. It is known that if $G$ is globally rigid in $\mathbb{R}^d$ on at least $d+2$ vertices, then it is strongly reconstructible in $\mathbb{C}^d$. We strengthen this result and show that under the same conditions, $G$ is in fact fully reconstructible in $\mathbb{C}^d$, which means that the set of edge lengths alone is sufficient to uniquely reconstruct $G$, without any constraint on the number of vertices (although still under the assumption that the edge lengths come from a generic realization). As a key step in our proof, we also prove that if $G$ is globally rigid in $\mathbb{R}^d$ on at least $d+2$ vertices, then the $d$-dimensional generic rigidity matroid of $G$ is connected. Finally, we provide new families of fully reconstructible graphs and use them to answer some questions regarding unlabeled reconstructibility posed in recent papers.

math.MG↗

arXiv2022

Lattices Without a Big Constant and With Noise

We show how Frieze's analysis of subset sum solving using lattices can be done with out any large constants and without flipping. We apply the variant without the large constant to inputs with noise.

cs.DS↗

arXiv2021

Unique Geometry and Texture from Corresponding Image Patches

We present a sufficient condition for recovering unique texture and viewpoints from unknown orthographic projections of a flat texture process. We show that four observations are sufficient in general, and we characterize the ambiguous cases. The results are applicable to shape from texture and texture-based structure from motion.

cs.CV↗

arXiv2021

Transverse Rigidity is Prestress Stability

Recently, V. Alexandrov proposed an intriguing sufficient condition for rigidity, which we will call "transverse rigidity". We show that transverse rigidity is actually equivalent to the known sufficient condition for rigidity called "prestress stability". Indeed this leads to a novel interpretation of the prestress condition.

math.MG↗