Search

In this article, we consider set optimization problems with variable ordering structures. Within the framework of the set less order relation with variable ordering structures, we investigate the existence, the upper convergence, and the lower convergence of solutions to such problems in the image spaces. For both the existence and the upper convergence of solutions, we employ new techniques to obtain various results without assuming the compactness of the constraint sets. Additionally, we utilize the domination property concerning the variable ordering cones to address the lower convergence of solutions. The obtained results are presented in several versions from different aspects for convenient comparison with existing results. Many examples are provided to illustrate the novelty of our results or to compare them with existing ones in the literature. [ABSTRACT FROM AUTHOR]

This paper focuses on the description and computation of the B-differential of the componentwise minimum of two affine vector functions. This issue arises in the reformulation of the linear complementarity problem with the Min C-function. The question has many equivalent formulations and we identify some of them in linear algebra, convex analysis and discrete geometry. These formulations are used to state some properties of the B-differential, like its symmetry, condition for its completeness, its connectivity, bounds on its cardinality, etc. The set to specify has a finite number of elements, which may grow exponentially with the range space dimension of the functions, so that its description is most often algorithmic. We first present an incremental-recursive approach avoiding to solve any optimization subproblem, unlike several previous approaches. It is based on the notion of matroid circuit and the related introduced concept of stem vector. Next, we propose modifications, adapted to the problem at stake, of an algorithm introduced by Rada and Černý (SIAM J Discret Math 32(1):455-473, 2018, https://doi.org/10.1137/15M1027930) to determine the cells of an arrangement in the space of hyperplanes having a point in common. Measured in CPU time on the considered test-problems, the mean acceleration ratios of the proposed algorithms, with respect to the one of Rada and Černý, are in the range 15..31, and this speed-up can exceed 100, depending on the problem, the approach and the chosen linear optimization and matroid solvers. [ABSTRACT FROM AUTHOR]

We consider an incomplete Cauchy problem governed by a quasi-convex potential and a quasi-nonexpansive nonpotential effect. We state and prove several results on the weak and strong convergence of solutions to this problem with the Lipschitz continuity condition. Discrete-time counterparts of these results are investigated as well. [ABSTRACT FROM AUTHOR]

In this paper, a novel algorithm is proposed as the conjugate gradient subspace pursuit (CGSP). The main idea of CGSP is given to combine the subspace pursuit (SP) algorithm with the conjugate gradient method. Then, with mild assumption, the recovery guarantee condition is established, which implied that the algorithm can converge to an exact sparse solution after a finite number of iterations. Finally, the numerical simulation of CGSP is given. The simulation results show that the sparse signal satisfied the specific conditions can be recovered by CGSP with high probability and the high sampling rate, and CGSP has the best performance. [ABSTRACT FROM AUTHOR]

For a linear equality constrained convex optimization problem, we initially propose a mixed primal-dual dynamical system with Hessian driven damping. This dynamical system comprises a second-order ordinary differential equation (ODE) with $ \alpha /t $ α / t damping and Hessian driven damping for the primal variable, and a first-order ODE for the dual variable. Utilizing the Lyapunov analysis approach, we prove that all the Lagrangian residual, the objective residual and the feasibility violation enjoy fast convergence under general scaling coefficients. We further analyse the convergence rate results under specific scaling coefficients. Our mixed dynamical system is extended to solve multi-block optimization problems, and we also consider the discrete case of the mixed dynamical system. This is the first study to explore primal-dual dynamical systems with Hessian driven damping for linear equality constrained problems. [ABSTRACT FROM AUTHOR]

This paper is concerned with the study of a family of fixed point iterations combining relaxation with different inertial (acceleration) principles. We provide a systematic, unified and insightful analysis of the hypotheses that ensure their weak, strong and linear convergence, either matching or improving previous results obtained by analysing particular cases separately. We also show that these methods are robust with respect to different kinds of perturbations–which may come from computational errors, intentional deviations, as well as regularisation or approximation schemes–under surprisingly weak assumptions. Although we mostly focus on theoretical aspects, numerical illustrations in image inpainting and electricity production markets reveal possible trends in the behaviour of these types of methods. [ABSTRACT FROM AUTHOR]

The minimization of a data-fidelity term and an additive regularization functional gives rise to a powerful framework for supervised learning. In this paper, we present a unifying regularization functional that depends on an operator L and on a generic Radon-domain norm. We establish the existence of a minimizer and give the parametric form of the solution(s) under very mild assumptions. When the norm is Hilbertian, the proposed formulation yields a solution that involves radial-basis functions and is compatible with the classical methods of machine learning. By contrast, for the total-variation norm, the solution takes the form of a two-layer neural network with an activation function that is determined by the regularization operator. In particular, we retrieve the popular ReLU networks by letting the operator be the Laplacian. We also characterize the solution for the intermediate regularization norms ‖ · ‖ = ‖ · ‖ L p with p ∈ (1 , 2 ] . Our framework offers guarantees of universal approximation for a broad family of regularization operators or, equivalently, for a wide variety of shallow neural networks, including the cases (such as ReLU) where the activation function is increasing polynomially. It also explains the favorable role of bias and skip connections in neural architectures. [ABSTRACT FROM AUTHOR]

The first two authors of this paper asserted in Lemma 4 of 'New Farkas-type constraint qualifications in convex infinite programming' (DOI: ) that a given reverse convex inequality is consequence of a given convex system satisfying the Farkas–Minkowski constraint qualification if and only if certain set depending on the data contains a particular point of the vertical axis. This paper identifies a hidden assumption in this reverse Farkas lemma which always holds in its applications to nontrivial optimization problems. Moreover, it shows that the statement remains valid when the Farkas–Minkowski constraint qualification fails by replacing the mentioned set by its closure. This hidden assumption is also characterized in terms of the data. Finally, the paper provides some applications to convex infinite systems and to convex infinite optimization problems. [ABSTRACT FROM AUTHOR]

We study the convergence of a random iterative sequence of a family of operators on infinite-dimensional Hilbert spaces, inspired by the stochastic gradient descent (SGD) algorithm in the case of the noiseless regression. We identify conditions that are strictly broader than previously known for polynomial convergence rate in various norms, and characterize the roles the randomness plays in determining the best multiplicative constants. Additionally, we prove almost sure convergence of the sequence. [ABSTRACT FROM AUTHOR]

It is shown that a lsc function is convex if and only if the minimal set of any linear perturbation of the function is convex. That fact also yields that the convexity of a function is equivalent to the quasiconvexity of its all linear perturbations. [ABSTRACT FROM AUTHOR]

The paper deals with fractional optimization problems where the objective function (ratio of two functions) is defined on a Cartesian product of two real normed spaces X and Y. Within this framework, we are interested to determine the so-called partial minimizers, i.e. points in $ X \times Y $ X × Y with the property that any of its variables minimizes the objective function, restricted to this variable, with respect to the other one. While any global minimizer is obviously a partial minimizer, the reverse implication holds true only under additional assumptions (e.g. separability properties of the involved functions). By exploiting the particularities of the objective function, we deliver a Dinkelbach type algorithm for computing partial minimizers of fractional optimization problems. Further assumptions on the involved spaces and functions, such as Lipschitz-type continuity, partial Fréchet differentiability, and coercivity, enable us to establish the convergence of our algorithm to a partial minimizer. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of identification of a Robin coefficient (also known as impedance coefficient) in a parabolic PDE from terminal observations of the temperature distributions. The problem is ill-posed in the sense that small perturbation in the observation may lead to a large deviation in the solution. Thus, in order to obtain stable approximations, we employ the Tikhonov-regularization. We propose a weak source condition motivated by the work of Engl and Zou (2000) and obtain a convergence rate of O (δ 1 2) , where δ is the noise level of the observed data. The obtained rate is better than some of the previous known rates. Moreover, the advantage of the proposed source condition is that the above mentioned convergence rate is obtained without the need for characterizing the range space of modelling operator and its Fréchet derivative, which is in contrast to the general convergence theory of Tikhonov-regularization for non linear operators. [ABSTRACT FROM AUTHOR]

In the paper, we consider both finite-dimensional and infinite-dimensional optimization problems with inclusion-type and equality-type constraints. We obtain sufficient conditions for the stability in the weak topology of a solution to this problem with respect to small perturbations of the problem parameters. In the finite-dimensional case, conditions for the stability in the strong topology of the solution are obtained for the problem with equality-type constraints. These conditions are based on a certain implicit function theorem. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the asymptotic behavior of inertial dynamics with dry friction within the context of a Hilbert framework for convex differentiable optimization. Our study focuses on a doubly nonlinear first-order evolution inclusion that encompasses two potentials. In our analysis, we specifically focus on two main components: the differentiable function f that needs to be minimized, which influences the system's state through its gradient, and the nonsmooth dry friction potential denoted as φ = r ‖ · ‖ . It's important to note that the dry friction term acts on a linear combination of the velocity vector and the gradient of f. Consequently, any stationary point in our system corresponds to a critical point of f, unlike the case where only the velocity vector is involved in the dry friction term, resulting in an approximate critical point of f. To emphasize the crucial role of ∇ f (x) , we also explore the dual formulation of this dynamic, which possesses a Riemannian gradient structure. To address these dynamics, we employ the recently developed generic acceleration approach by Attouch, Bot, and Nguyen. This approach involves the time scaling of a continuous first-order differential equation, followed by the application of the method of averaging. By applying this methodology, we derive fast convergence results for second-order time-evolution systems with dry friction, asymptotically vanishing viscous damping, and implicit Hessian-driven damping. [ABSTRACT FROM AUTHOR]

This paper is intended to provide an updated survey of recent optimality theory for infinite-dimensional convex programming. It aims at establishing theoretical support for algorithmic developments. Two alternative strategies inspire the approaches presented in the paper. The first one consists of replacing the family of constraints by a single one, appealing to the supremum function, and is based on various characterizations of the subdifferential of the pointwise supremum of convex functions. The second one uses appropriate characterizations of affine consequent inequalities of the constraint system exploiting ad hoc constraint qualifications. [ABSTRACT FROM AUTHOR]

Consider three closed linear subspaces C 1 , C 2 , and C 3 of a Hilbert space H and the orthogonal projections P 1 , P 2 and P 3 onto them. Halperin showed that a point in C 1 ∩ C 2 ∩ C 3 can be found by iteratively projecting any point x 0 ∈ H onto all the sets in a periodic fashion. The limit point is then the projection of x 0 onto C 1 ∩ C 2 ∩ C 3 . Nevertheless, a non-periodic projection order may lead to a non-convergent projection series, as shown by Kopecká, Müller, and Paszkiewicz. This raises the question how many projection orders in { 1 , 2 , 3 } N are "well behaved" in the sense that they lead to a convergent projection series. Melo, da Cruz Neto, and de Brito provided a necessary and sufficient condition under which the projection series converges and showed that the "well behaved" projection orders form a large subset in the sense of having full product measure. We show that also from a topological viewpoint the set of "well behaved" projection orders is a large subset: it contains a dense G δ subset with respect to the product topology. Furthermore, we analyze why the proof of the measure theoretic case cannot be directly adapted to the topological setting. [ABSTRACT FROM AUTHOR]

In this paper, we consider a class of second-order ordinary differential equations with Hessian-driven damping and Tikhonov regularization, which arises from the minimization of a smooth convex function in Hilbert spaces. Inspired by Attouch et al. (J Differ Equ 261:5734–5783, 2016), we establish that the function value along the solution trajectory converges to the optimal value, and prove that the convergence rate can be as fast as o (1 / t 2) . By constructing proper energy function, we prove that the trajectory strongly converges to a minimizer of the objective function of minimum norm. Moreover, we propose a gradient-based optimization algorithm based on numerical discretization, and demonstrate its effectiveness in numerical experiments. [ABSTRACT FROM AUTHOR]

In this paper, we prove BV regularity on the transport density in the mass transport problem to the boundary in two dimensions under certain conditions on the domain, the boundary cost and the mass distribution. Moreover, we show by a counter-example that the smoothness of the mass distribution, the boundary and the boundary cost does not imply that the transport density is W 1 , p , for some p > 1 . [ABSTRACT FROM AUTHOR]

In a Hilbert setting we aim to study a second order in time differential equation, combining viscous and Hessian-driven damping, containing a time scaling parameter function and a Tikhonov regularization term. The dynamical system is related to the problem of minimization of a nonsmooth convex function. In the formulation of the problem as well as in our analysis we use the Moreau envelope of the objective function and its gradient and heavily rely on their properties. We show that there is a setting where the newly introduced system preserves and even improves the well-known fast convergence properties of the function and Moreau envelope along the trajectories and also of the gradient of Moreau envelope due to the presence of time scaling. Moreover, in a different setting we prove strong convergence of the trajectories to the element of minimal norm from the set of all minimizers of the objective. The manuscript concludes with various numerical results. [ABSTRACT FROM AUTHOR]

A Fan-Theobald-von Neumann system [7] is a triple (V , W , λ) , where V and W are real inner product spaces and λ : V → W is a norm-preserving map satisfying a Fan-Theobald-von Neumann type inequality together with a condition for equality. Examples include Euclidean Jordan algebras, systems induced by certain hyperbolic polynomials, and normal decomposition systems (Eaton triples). The present article is a continuation of [9] where the concepts of commutativity, automorphisms, majorization, and reduction were introduced and elaborated. Here, we describe some transfer principles and present Fenchel conjugate and subdifferential formulas. [ABSTRACT FROM AUTHOR]

The optimal control problems for the wave equation are considered on networks. The turnpike property is shown for the state equation, the adjoint state equation as well as the optimal cost. The shape and topology optimization is performed for the network with the shape functional given by the optimality system of the control problem. The set of admissible shapes for the network is compact in finite dimensions, thus the use of turnpike property is straightforward. The topology optimization is analysed for an example of nucleation of a small cycle at the internal node of network. The topological derivative of the cost is introduced and evaluated in the framework of domain decomposition technique. Numerical examples are provided. [ABSTRACT FROM AUTHOR]

We provide a new, short proof of the density in energy of Lipschitz functions into the metric Sobolev space defined by using plans with barycenter (and thus, a fortiori, into the Newtonian–Sobolev space). Our result covers first-order Sobolev spaces of exponent p ∈ (1 , ∞) , defined over a complete separable metric space endowed with a boundedly-finite Borel measure. Our proof is based on a completely smooth analysis: first we reduce the problem to the Banach space setting, where we consider smooth functions instead of Lipschitz ones, then we rely on classical tools in convex analysis and on the superposition principle for normal 1-currents. Along the way, we obtain a new proof of the density in energy of smooth cylindrical functions in Sobolev spaces defined over a separable Banach space endowed with a finite Borel measure. [ABSTRACT FROM AUTHOR]

We present an analysis of generalized Nash equilibrium problems in infinite-dimensional spaces with possibly non-convex objective functions of the players. Such settings arise, for instance, in games that involve nonlinear partial differential equation constraints. Due to non-convexity, we work with equilibrium concepts that build on first order optimality conditions, especially Quasi-Nash Equilibria (QNE), i.e. first-order optimality conditions for (Generalized) Nash Equilibria, and Variational Equilibria (VE), i.e. first-order optimality conditions for Normalized Nash Equilibria. We prove existence of these types of equilibria and study characterizations of them via regularized (and localized) Nikaido-Isoda merit functions. We also develop continuity and (continuous) differentiability results for these merit functions under quite weak assumptions, using a generalization of Danskin's theorem. They provide a theoretical foundation for, e.g. using globalized descent methods for computing QNE or VE. [ABSTRACT FROM AUTHOR]

The aim of this paper is to introduce a nonlinear scalarization function in set optimization based on the concept of null set which was introduced by Wu (J Math Anal Appl 472(2):1741–1761, 2019). We introduce a notion of pseudo algebraic interior of a set and define a weak set order relation using the concept of null set. We investigate several properties of this nonlinear scalarization function. Further, we characterize the set order relations and investigate optimality conditions for solution sets in set optimization based on the concept of null set. Finally, a numerical example is provided to compute a weak minimal solution using this nonlinear scalarization function. [ABSTRACT FROM AUTHOR]

In a real Hilbert space setting, we reconsider the classical Arrow–Hurwicz differential system in view of solving linearly constrained convex minimization problems. We investigate the asymptotic properties of the differential system and provide conditions for which its solutions converge towards a saddle point of the Lagrangian associated with the convex minimization problem. Our convergence analysis mainly relies on a 'Lagrangian identity' which naturally extends on the well-known descent property of the classical continuous steepest descent method. In addition, we present asymptotic estimates on the decay of the solutions and the primal-dual gap function measured in terms of the Lagrangian. These estimates are further refined to the ones of the classical damped harmonic oscillator provided that second-order information on the objective function of the convex minimization problem is available. Finally, we show that our results directly translate to the case of solving structured convex minimization problems. Numerical experiments further illustrate our theoretical findings. [ABSTRACT FROM AUTHOR]

This work is devoted to show the efficiency of a new numerical approach in solving geometrical shape optimization problems constrained to partial differential equations, on a family of convex domains. More precisely, we are interested to an improved numerical optimization process based on the new shape derivative formula, using the Minkowski deformation of convex domains, recently established in Boulkhemair and Chakib (J. Convex Anal. 21( n ∘ 1 ), 67–87 2014), Boulkhemair (SIAM J. Control Optim. 55( n ∘ 1 ), 156–171 2017). This last formula allows to express the shape derivative by means of the support function, in contrast to the classical one expressed in term of vector fields Henrot and Pierre 2005, Delfour and Zolésio 2011, Sokolowski and Zolesio 1992. This avoids some of the disadvantages related to the classical shape derivative approach, when one use the finite elements discretization for approximating the auxiliary boundary value problems in shape optimization processes Allaire 2007. So, we investigate here the performance of the proposed shape optimization approach through the numerical resolution of some shape optimization problems constrained to boundary value problems governed by Laplace or Stokes operator. Notably, we carry out a comparative numerical study between its resulting numerical optimization process and the classical one. Finally, we give some numerical results showing the efficiency of the proposed approach and its ability in producing good quality solutions and in providing better accuracy for the optimal solution in less CPU time compared to the classical approach. [ABSTRACT FROM AUTHOR]

In this note, we provide a characterization for the set of extreme points of the Lipschitz unit ball in a specific vectorial setting. While the analysis of the case of real-valued functions is covered extensively in the literature, no information about the vectorial case has been provided up to date. Here, we aim at partially filling this gap by considering functions mapping from a finite metric space to a strictly convex Banach space that satisfy the Lipschitz condition. As a consequence, we present a representer theorem for such functions. In this setting, the number of extreme points needed to express any point inside the ball is independent of the dimension, improving the classical result from Carathéodory. [ABSTRACT FROM AUTHOR]

In a Hilbertian framework, for the minimization of a general convex differentiable function f, we introduce new inertial dynamics and algorithms that generate trajectories and iterates that converge fastly towards the minimizer of f with minimum norm. Our study is based on the non-autonomous version of the Polyak heavy ball method, which, at time t, is associated with the strongly convex function obtained by adding to f a Tikhonov regularization term with vanishing coefficient ε (t) . In this dynamic, the damping coefficient is proportional to the square root of the Tikhonov regularization parameter ε (t) . By adjusting the speed of convergence of ε (t) towards zero, we will obtain both rapid convergence towards the infimal value of f, and the strong convergence of the trajectories towards the element of minimum norm of the set of minimizers of f. In particular, we obtain an improved version of the dynamic of Su-Boyd-Candès for the accelerated gradient method of Nesterov. This study naturally leads to corresponding first-order algorithms obtained by temporal discretization. In the case of a proper lower semicontinuous and convex function f, we study the proximal algorithms in detail, and show that they benefit from similar properties. [ABSTRACT FROM AUTHOR]

This paper concerns second-order sufficient conditions of optimality for a class of infinite dimensional optimal control problems under control constraints and end-point constraints. The distinctive feature of our formulation is the use of variational analysis techniques. Our approach is based on a suitable decomposition of controls, using second-order tangents, and it is developed in the case of functional constraints. An adaptation of the tools in the case of pointwise constraints on the controls is proposed too. We provide applications to concrete optimal control problems involving PDEs. [ABSTRACT FROM AUTHOR]

In a Hilbert setting, our study focuses on the dynamical system introduced by Su-Boyd-Candès as a low resolution ODE of Nesterov’s accelerated gradient method (NAG). This inertial system, denoted by (A V D) α , is driven by the gradient of the function f to be minimized, and is damped with an asymptotic vanishing coefficient of the form α / t , with α ≥ 3 . Taking α large enough plays a crucial role in the asymptotic convergence properties of the trajectories. For a general convex function f , taking α > 3 guarantees the asymptotic convergence rate of the values o (1 / t 2) , as well as the convergence of the trajectories towards optimal solutions. For strongly convex f , the asymptotic rate of convergence is of order 1 / t 2 α 3 , which increases with α . To analyze the effect of the parameter α in the convergence properties of (A V D) α , we show that a judicious time scaling of (A V D) α produces trajectories close to those of the continuous steepest descent method associated with f when α is sufficiently large. This limiting process involves a singular perturbation property, as we move from a second-order evolution equation to a first-order one. This transition enables us to understand the change in the rate of convergence from 1 / t to 1 / t 2 between the steepest descent method and (NAG). Based on a complexity analysis over a finite time interval, new results are obtained regarding the optimal tuning of the parameter α and the involved constants C α in the estimations. Numerical experiments have been conducted to illustrate and confirm the theoretical results. [ABSTRACT FROM AUTHOR]

We develop a new perturbation method in Orlicz sequence spaces ℓ M with Orlicz function M satisfying Δ 2 condition at zero. This result allows one to support from below any bounded below lower semicontinuous function with bounded support, with a perturbation of the defining function σ M . We give few examples how the method can be used for determining the type of the smoothness of certain Orlicz spaces. [ABSTRACT FROM AUTHOR]

In this paper, a general formula concerning the multi-composition rule of convex subdifferential calculus is provided in the setting of Banach spaces under an appropriate regularity condition. As an application, this calculus rule is applied to obtain necessary and sufficient Karush–Kuhn–Tucker type optimality conditions for constrained convex minmax location problems with perturbed minimal time functions and set-up costs. [ABSTRACT FROM AUTHOR]

We prove that solutions to several shape optimization problems in the plane, with a convexity constraint on the admissible domains, are polygons. The main terms of the shape functionals we consider are either the Dirichlet energy E f ⁢ (Ω) of the Laplacian in the domain Ω or the first eigenvalue λ 1 ⁢ (Ω) of the Dirichlet-Laplacian. Usually, one considers minimization of such functionals (often with measure constraint), as for example for the famous Saint-Venant and Faber-Krahn inequalities. By adding the convexity constraint (and possibly other natural constraints), we instead consider the rather unusual and difficult question of maximizing these functionals. This paper follows a series of papers by the authors, where the leading idea is that a certain concavity property of the shape functional that is minimized leads optimal shapes to locally saturate their convexity constraint, which geometrically means that they are polygonal. In these previous papers, the leading term in the shape functional was usually the opposite of the perimeter, for which the aforementioned concavity property was rather easy to obtain through computations of its second order shape derivative. By carrying classical shape calculus, a similar concavity property can be observed for the opposite of E f ⁢ (Ω) or λ 1 ⁢ (Ω) when shapes are smooth and convex. The main novelty in the present paper is the proof of a weak convexity property of E f ⁢ (Ω) and λ 1 ⁢ (Ω) among planar convex shapes, namely rather nonsmooth shapes. This involves new computations and estimates of the second order shape derivatives of E f ⁢ (Ω) and λ 1 ⁢ (Ω) interesting for themselves. [ABSTRACT FROM AUTHOR]

The class of convex–concave bilinear saddle point problems encompasses many important convex optimization models arising in a wide array of applications. The most of existing primal–dual dynamical systems for saddle point problems are based on first order ordinary differential equations (ODEs), which only own the O (1 / t) convergence rate in the convex case, and fast convergence rate analysis always requires some additional assumption such as strong convexity. In this paper, based on second order ODEs, we consider a general inertial primal–dual dynamical system, with damping, scaling and extrapolation coefficients, for a convex–concave bilinear saddle point problem. By the Lyapunov analysis approach, under appropriate assumptions, we investigate the convergence rates of the primal–dual gap and velocities, and the boundedness of the trajectories for the proposed dynamical system. With special parameters, our results can recover the Polyak's heavy ball acceleration scheme and Nesterov's acceleration scheme. We also provide numerical examples to support our theoretical claims. [ABSTRACT FROM AUTHOR]

A risk analyst assesses potential financial losses based on multiple sources of information. Often, the assessment does not only depend on the specification of the loss random variable, but also on various economic scenarios. Motivated by this observation, we design a unified axiomatic framework for risk evaluation principles which quantify jointly a loss random variable and a set of plausible probabilities. We call such an evaluation principle a generalised risk measure. We present a series of relevant theoretical results. The worst-case, coherent and robust generalised risk measures are characterised via different sets of intuitive axioms. We establish the equivalence between a few natural forms of law-invariance in our framework, and the technical subtlety therein reveals a sharp contrast between our framework and the traditional one. Moreover, coherence and strong law-invariance are derived from a combination of other conditions, which provides additional support for coherent risk measures such as expected shortfall over value-at-risk, a relevant issue for risk management practice. [ABSTRACT FROM AUTHOR]

In the framework of a real Hilbert space, we address the problem of finding the zeros of the sum of a maximally monotone operator A and a cocoercive operator B. We study the asymptotic behaviour of the trajectories generated by a second order equation with vanishing damping, attached to this problem, and governed by a time-dependent forward–backward-type operator. This is a splitting system, as it only requires forward evaluations of B and backward evaluations of A. A proper tuning of the system parameters ensures the weak convergence of the trajectories to the set of zeros of A + B , as well as fast convergence of the velocities towards zero. A particular case of our system allows to derive fast convergence rates for the problem of minimizing the sum of a proper, convex and lower semicontinuous function and a smooth and convex function with Lipschitz continuous gradient. We illustrate the theoretical outcomes by numerical experiments. [ABSTRACT FROM AUTHOR]

This paper studies the problem of learning an unknown function f from given data about f. The learning problem is to give an approximation f ^ to f that predicts the values of f away from the data. There are numerous settings for this learning problem depending on (i) what additional information we have about f (known as a model class assumption), (ii) how we measure the accuracy of how well f ^ predicts f, (iii) what is known about the data and data sites, (iv) whether the data observations are polluted by noise. A mathematical description of the optimal performance possible (the smallest possible error of recovery) is known in the presence of a model class assumption. Under standard model class assumptions, it is shown in this paper that a near optimal f ^ can be found by solving a certain finite-dimensional over-parameterized optimization problem with a penalty term. Here, near optimal means that the error is bounded by a fixed constant times the optimal error. This explains the advantage of over-parameterization which is commonly used in modern machine learning. The main results of this paper prove that over-parameterized learning with an appropriate loss function gives a near optimal approximation f ^ of the function f from which the data is collected. Quantitative bounds are given for how much over-parameterization needs to be employed and how the penalization needs to be scaled in order to guarantee a near optimal recovery of f. An extension of these results to the case where the data is polluted by additive deterministic noise is also given. [ABSTRACT FROM AUTHOR]

In this paper, we provide an explicit expression for the proximity operator of a perspective of any proper lower semicontinuous convex function defined on a Hilbert space. Our computation enhances and generalizes known formulae for the case when the Fenchel conjugate of the convex function has an open domain or when it is radial. We show numerically that our approach is more efficient than previous approaches in the literature and we provide several examples of nonradial functions for which the domain of its conjugate is not open and we compute the proximity operators of their perspectives. [ABSTRACT FROM AUTHOR]