Convex Analysis at Infinity: An Introduction to Astral Space

• URL: http://arxiv.org/abs/2205.03260v3
• Date: Tue, 16 Jul 2024 14:26:24 GMT
• Title: Convex Analysis at Infinity: An Introduction to Astral Space
• Abstract summary: Not all convex functions on $mathbbRn$ have finite minimizers; some can only be minimized by a sequence as it heads to infinity. We study astral space, a compact extension of $mathbbRn$ to which such points at infinity have been added.
• Score: 25.760842971037576
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Not all convex functions on $\mathbb{R}^n$ have finite minimizers; some can only be minimized by a sequence as it heads to infinity. In this work, we aim to develop a theory for understanding such minimizers at infinity. We study astral space, a compact extension of $\mathbb{R}^n$ to which such points at infinity have been added. Astral space is constructed to be as small as possible while still ensuring that all linear functions can be continuously extended to the new space. Although astral space includes all of $\mathbb{R}^n$, it is not a vector space, nor even a metric space. However, it is sufficiently well-structured to allow useful and meaningful extensions of concepts of convexity, conjugacy, and subdifferentials. We develop these concepts and analyze various properties of convex functions on astral space, including the detailed structure of their minimizers, exact characterizations of continuity, and convergence of descent algorithms.

Related papers

• $\mathbb{SL}(n)$ Representation Learning: An Intrinsic Mixed-Curvature Space with Higher Curvature Capacities and Deeper Order-Aware Composition [47.80325814873361]
  Mixed-curvature representation learning seeks to capture rich geometric structures that cannot be adequately modeled by a single curvature regime.<n>We introduce the $mathbbSL(n)$ space, a representation geometry defined by the simple $det(A)variant=1$ constraint and a left in Schatten-$p$ structure.<n>Despite this minimal construction, $mathbbSL(n)$ exhibits pointwise negative, zero, and flag curvature around a common flagpole.
  arXiv  Detail & Related papers  (2026-09-14T05:52:14Z)
• Characterising the sets of quantum states with non-negative Wigner function [0.15393457051344295]
  We consider the convex sets $mathcal D_+(mathcal H)$ of Wigner-positive states (WPS) over $mathcal H$ with non-negative Wigner function.<n>We construct minimal sets of states that generate $mathcal D_+(mathcal H)$ through convex combinations.
  arXiv  Detail & Related papers  (2025-12-16T19:00:02Z)
• Flat Channels to Infinity in Neural Loss Landscapes [46.76940650038536]
  Loss landscapes of neural networks contain minima and saddle points that may be connected in flat regions or appear in isolation.<n>We identify and characterize a special structure in the loss landscape: channels along which the loss decreases extremely slowly.<n>The emergence of gated linear units at the end of the channels highlights a surprising aspect of the computational capabilities of fully connected layers.
  arXiv  Detail & Related papers  (2025-06-17T20:04:15Z)
• Two-Timescale Gradient Descent Ascent Algorithms for Nonconvex Minimax Optimization [77.3396841985172]
  We provide a unified analysis of two-timescale gradient ascent (TTGDA) for solving structured non minimax optimization problems.<n>Our contribution is to design TTGDA algorithms are effective beyond the setting.
  arXiv  Detail & Related papers  (2024-08-21T20:14:54Z)
• Geometry of degenerate quantum states, configurations of $m$-planes and invariants on complex Grassmannians [55.2480439325792]
  We show how to reduce the geometry of degenerate states to the non-abelian connection $A$. We find independent invariants associated with each triple of subspaces. Some of them generalize the Berry-Pancharatnam phase, and some do not have analogues for 1-dimensional subspaces.
  arXiv  Detail & Related papers  (2024-04-04T06:39:28Z)
• Nearly Optimal Regret for Decentralized Online Convex Optimization [53.433398074919]
  Decentralized online convex optimization (D-OCO) aims to minimize a sequence of global loss functions using only local computations and communications. We develop novel D-OCO algorithms that can respectively reduce the regret bounds for convex and strongly convex functions. Our algorithms are nearly optimal in terms of $T$, $n$, and $rho$.
  arXiv  Detail & Related papers  (2024-02-14T13:44:16Z)
• A Minimal Control Family of Dynamical Syetem for Universal Approximation [6.164223149261533]
  We show that a control family can generate flow maps that can approximate diffeomorphisms of $mathbbRd$ on any compact domain. Our result reveals an underlying connection between the approximation power of neural networks and control systems.
  arXiv  Detail & Related papers  (2023-12-20T10:36:55Z)
• Completely entangled subspaces of entanglement depth $k$ [0.0]
  We introduce a class of entangled subspaces: completely entangled subspaces of entanglement depth $k$ ($k$-CESs) We discuss the relation between these subspaces and unextendible product bases (UPBs)
  arXiv  Detail & Related papers  (2023-12-13T19:25:01Z)
• Global universal approximation of functional input maps on weighted spaces [3.8059763597999012]
  We introduce so-called functional input neural networks defined on a possibly infinite dimensional weighted space with values also in a possibly infinite dimensional output space. We prove a global universal approximation result on weighted spaces for continuous functions going beyond the usual approximation on compact sets. We emphasize that the reproducing Hilbert kernel space of the signature kernels are Cameron-Martin spaces of certain Gaussian processes.
  arXiv  Detail & Related papers  (2023-06-05T23:06:32Z)
• Activation Functions Not To Active: A Plausible Theory on Interpreting Neural Networks [0.0]
  Researchers believe that neural networks model a high-dimensional space but cannot give a clear definition of this space. We develop a plausible theory on interpreting neural networks in terms of the role of activation functions in neural networks. We show that the activation function acts as a magnifying function that maps the low-dimensional linear space into an infinite-dimensional space.
  arXiv  Detail & Related papers  (2023-05-01T05:23:58Z)
• Quantum and classical low-degree learning via a dimension-free Remez inequality [52.12931955662553]
  We show a new way to relate functions on the hypergrid to their harmonic extensions over the polytorus. We show the supremum of a function $f$ over products of the cyclic group $exp(2pi i k/K)_k=1K$. We extend to new spaces a recent line of work citeEI22, CHP, VZ22 that gave similarly efficient methods for learning low-degrees on hypercubes and observables on qubits.
  arXiv  Detail & Related papers  (2023-01-04T04:15:40Z)
• Bound states of weakly deformed soft waveguides [0.0]
  We consider the two-dimensional Schr"odinger operator with an attractive potential which is a multiple of the characteristic function of an unbounded strip-shaped region. In the critical case $int_mathbbR f,mathsfd x = 0$, we derive a sufficient condition for the existence of a unique bound state for all sufficiently small $varepsilon > 0$.
  arXiv  Detail & Related papers  (2022-11-03T16:56:39Z)
• A Variational Quantum Algorithm For Approximating Convex Roofs [0.0]
  entanglement measures are first defined for pure states of a bipartite Hilbert space, and then extended to mixed states via the convex roof extension. We produce a sequence of extensions that we call $f$-$d$ extensions, for $d in mathbbN$. We introduce a quantum variational algorithm which aims to approximate the $f$-$d$ extensions of entanglement measures defined on pure states.
  arXiv  Detail & Related papers  (2022-03-04T02:30:35Z)
• Proximal Gradient Descent-Ascent: Variable Convergence under K{\L} Geometry [49.65455534654459]
  The finite descent-ascent parameters (GDA) has been widely applied to solve minimax optimization problems. This paper fills such a gap by studying the convergence of the KL-Lized geometry.
  arXiv  Detail & Related papers  (2021-02-09T05:35:53Z)
• Stochastic Linear Bandits with Protected Subspace [51.43660657268171]
  We study a variant of the linear bandit problem wherein we optimize a linear objective function but rewards are accrued only to an unknown subspace. In particular, at each round, the learner must choose whether to query the objective or the protected subspace alongside choosing an action. Our algorithm, derived from the OFUL principle, uses some of the queries to get an estimate of the protected space.
  arXiv  Detail & Related papers  (2020-11-02T14:59:39Z)
• A deep network construction that adapts to intrinsic dimensionality beyond the domain [79.23797234241471]
  We study the approximation of two-layer compositions $f(x) = g(phi(x))$ via deep networks with ReLU activation. We focus on two intuitive and practically relevant choices for $phi$: the projection onto a low-dimensional embedded submanifold and a distance to a collection of low-dimensional sets.
  arXiv  Detail & Related papers  (2020-08-06T09:50:29Z)
• Metrizing Weak Convergence with Maximum Mean Discrepancies [88.54422104669078]
  This paper characterizes the maximum mean discrepancies (MMD) that metrize the weak convergence of probability measures for a wide class of kernels. We prove that, on a locally compact, non-compact, Hausdorff space, the MMD of a bounded continuous Borel measurable kernel k, metrizes the weak convergence of probability measures if and only if k is continuous.
  arXiv  Detail & Related papers  (2020-06-16T15:49:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.