arXiv subjects

Yuuki Takai

Publications and source records attributed to Yuuki Takai.

9 recordsLinked to original sources

arXiv2024

Decomposition of Equivariant Maps via Invariant Maps: Application to Universal Approximation under Symmetry

In this paper, we develop a theory about the relationship between invariant and equivariant maps with regard to a group $G$. We then leverage this theory in the context of deep neural networks with group symmetries in order to obtain novel insight into their mechanisms. More precisely, we establish a one-to-one relationship between equivariant maps and certain invariant maps. This allows us to reduce arguments for equivariant maps to those for invariant maps and vice versa. As an application, we propose a construction of universal equivariant architectures built from universal invariant networks. We, in turn, explain how the universal architectures arising from our construction differ from standard equivariant architectures known to be universal. Furthermore, we explore the complexity, in terms of the number of free parameters, of our models, and discuss the relation between invariant and equivariant networks' complexity. Finally, we also give an approximation rate for G-equivariant deep neural networks with ReLU activation functions for finite group G.

cs.LG↗

arXiv2023

An equivalent condition for abelian varieties over finite fields to have QM

In this paper, we give an equivalent condition for an abelian variety over a finite field to have multiplication by a quaternion algebra over a number field. We prove the result by combining Tate's classification of the endomorphism algebras of abelian varieties over finite fields with Yu's criterion of the existence of homomorphisms between semi-simple algebras.

math.NT↗

arXiv2022

Cover time of graphs with bounded genus

The cover time of a finite connected graph is the expected number of steps needed for a simple random walk on the graph to visit all vertices of the graph. It is known that the cover time of any finite connected $n$-vertex graph is at least $(1 + o(1)) n \log n$ and at most $(1 + o(1)) \frac{4}{27} n^3$. By Jonasson and Schramm, the cover time of any bounded-degree finite connected $n$-vertex planar graph is at least $c n(\log n)^2$ and at most $6n^2$, where $c$ is a positive constant depending only on the maximal degree of the graph. In particular, the lower bound is established via the use of circle packing of planar graphs on the Riemann sphere. In this paper, we show that the cover time of any finite $n$-vertex graph $G$ with maximum degree $Δ$ on the compact Riemann surface $S$ of given genus $g$ is at least $c n(\log n)^2/ Δ(g + 1)$ and at most $(6 + o(1))n^2$, where $c$ is an absolute constant, if $n$ is sufficiently large and three sufficient conditions for $S$ and a circle packing of $G$ filling $S$.

cs.DM↗

arXiv2021

On the Number of Linear Functions Composing Deep Neural Network: Towards a Refined Definition of Neural Networks Complexity

The classical approach to measure the expressive power of deep neural networks with piecewise linear activations is based on counting their maximum number of linear regions. This complexity measure is quite relevant to understand general properties of the expressivity of neural networks such as the benefit of depth over width. Nevertheless, it appears limited when it comes to comparing the expressivity of different network architectures. This lack becomes particularly prominent when considering permutation-invariant networks, due to the symmetrical redundancy among the linear regions. To tackle this, we propose a refined definition of piecewise linear function complexity: instead of counting the number of linear regions directly, we first introduce an equivalence relation among the linear functions composing a piecewise linear function and then count those linear functions relative to that equivalence relation. Our new complexity measure can clearly distinguish between the two aforementioned models, is consistent with the classical measure, and increases exponentially with depth.

cs.LG↗

arXiv2021

Coarse Ricci curvature of hypergraphs and its generalization

In the present paper, we introduce a concept of Ricci curvature on hypergraphs for a nonlinear Laplacian. We prove that our definition of the Ricci curvature is a generalization of Lin-Lu-Yau coarse Ricci curvature for graphs to hypergraphs. We also show a lower bound of nonzero eigenvalues of Laplacian, gradient estimate of heat flow, and diameter bound of Bonnet-Myers type for our curvature notion. This research leads to understanding how nonlinearity of Laplacian causes complexity of curvatures.

math.MG↗

arXiv2020

Hypergraph Clustering Based on PageRank

A hypergraph is a useful combinatorial object to model ternary or higher-order relations among entities. Clustering hypergraphs is a fundamental task in network analysis. In this study, we develop two clustering algorithms based on personalized PageRank on hypergraphs. The first one is local in the sense that its goal is to find a tightly connected vertex set with a bounded volume including a specified vertex. The second one is global in the sense that its goal is to find a tightly connected vertex set. For both algorithms, we discuss theoretical guarantees on the conductance of the output vertex set. Also, we experimentally demonstrate that our clustering algorithms outperform existing methods in terms of both the solution quality and running time. To the best of our knowledge, ours are the first practical algorithms for hypergraphs with theoretical guarantees on the conductance of the output set.

cs.DS↗

arXiv2019

Universal approximations of permutation invariant/equivariant functions by deep neural networks

In this paper, we develop a theory about the relationship between $G$-invariant/equivariant functions and deep neural networks for finite group $G$. Especially, for a given $G$-invariant/equivariant function, we construct its universal approximator by deep neural network whose layers equip $G$-actions and each affine transformations are $G$-equivariant/invariant. Due to representation theory, we can show that this approximator has exponentially fewer free parameters than usual models.

cs.LG↗

arXiv2019

Finding Cheeger Cuts in Hypergraphs via Heat Equation

Cheeger's inequality states that a tightly connected subset can be extracted from a graph $G$ using an eigenvector of the normalized Laplacian associated with $G$. More specifically, we can compute a subset with conductance $O(\sqrt{ϕ_G})$, where $ϕ_G$ is the minimum conductance of a set in $G$. It has recently been shown that Cheeger's inequality can be extended to hypergraphs. However, as the normalized Laplacian of a hypergraph is no longer a matrix, we can only approximate to its eigenvectors; this causes a loss in the conductance of the obtained subset. To address this problem, we here consider the heat equation on hypergraphs, which is a differential equation exploiting the normalized Laplacian. We show that the heat equation has a unique solution and that we can extract a subset with conductance $\sqrt{ϕ_G}$ from the solution. An analogous result also holds for directed graphs.

cs.DS↗

arXiv2010

An effective isomorphy criterion for mod $\ell$ Galois representations

In this paper, we consider mod $\ell$ Galois representations of $\mathbb{Q}$. In particular, we obtain an effective criterion to distinguish two semisimple 2-dimensional, odd mod $\ell$ Galois representations up to isomorphism. Serre's conjecture (Khare-Wintenberger's theorem), Sturm's theorem, and its modification by Kohnen are used in our proof.

math.NT↗