Constant-Time Algorithms for Continuous Optimization Problems

In this chapter, we consider constant-time algorithms for continuous optimization problems. Specifically, we consider quadratic function minimization and tensor decomposition, both of which have numerous applications in machine learning and data mining. The key component in our analysis is graph limit theory, which was originally developed to study graphs analytically.

Action convergence of operators and graphs

We present a new approach to graph limit theory that unifies and generalizes the two most well-developed directions, namely dense graph limits (even the more general $L^p$ limits) and Benjamini–Schramm limits (even in the stronger local-global setting). We illustrate by examples that this new framework provides a rich limit theory with natural limit objects for graphs of intermediate density. Moreover, it provides a limit theory for bounded operators (called P-operators) of the form $L^\infty (\Omega )\to L^1(\Omega )$ for probability spaces $\Omega $ . We introduce a metric to compare P-operators (for example, finite matrices) even if they act on different spaces. We prove a compactness result, which implies that, in appropriate norms, limits of uniformly bounded P-operators can again be represented by P-operators. We show that limits of operators, representing graphs, are self-adjoint, positivity-preserving P-operators called graphops. Graphons, $L^p$ graphons, and graphings (known from graph limit theory) are special examples of graphops. We describe a new point of view on random matrix theory using our operator limit framework.

Limit of a nonpreferential attachment multitype network model

Here, we deal with a model of multitype network with nonpreferential attachment growth. The connection between two nodes depends asymmetrically on their types, reflecting the implication of time order in temporal networks. Based upon graph limit theory, we analytically determined the limit of the network model characterized by a kernel, in the sense that the number of copies of any fixed subgraph converges when network size tends to infinity. The results are confirmed by extensive simulations. Our work thus provides a theoretical framework for quantitatively understanding grown temporal complex networks as a whole.

Pointwise topological convergence and topological graph convergence of set-valued maps

Let X,Y be topological spaces and {Fn : n ? ?} be a sequence of set-valued maps from X to Y with the pointwise topological limit G and with the topological graph limit F. We give an answer to the question from ([19]): which conditions on X,Y and/or {F,G,Fn : n ? ?} are needed to F = G.