Concept

Contraction mapping

Related publications (18)

Stable Nonconvex-Nonconcave Training via Linear Interpolation

Volkan Cevher, Thomas Michaelsen Pethick, Wanyun Xie

This paper presents a theoretical analysis of linear interpolation as a principled method for stabilizing (large-scale) neural network training. We argue that instabilities in the optimization process are often caused by the nonmonotonicity of the loss lan ...

2023

Recursive evaluation and iterative contraction of N-body equivariant features

Michele Ceriotti, Sergey Pozdnyakov, Jigyasa Nigam

Mapping an atomistic configuration to a symmetrized N-point correlation of a field associated with the atomic positions (e.g., an atomic density) has emerged as an elegant and effective solution to represent structures as the input of machine-learning algo ...

2020

Efficient Proximal Mapping of the 1-path-norm of Shallow Networks

Volkan Cevher, Paul Thierry Yves Rolland, Fabian Ricardo Latorre Gomez, Shaul Nadav Hallak

We demonstrate two new important properties of the 1-path-norm of shallow neural networks. First, despite its non-smoothness and non-convexity it allows a closed form proximal operator which can be efficiently computed, allowing the use of stochastic proxi ...

2020

Deep Spline Networks With Control Of Lipschitz Regularity

Michaël Unser, Shayan Aziznejad

The motivation for this work is to improve the performance of deep neural networks through the optimization of the individual activation functions. Since the latter results in an infinite-dimensional optimization problem, we resolve the ambiguity by search ...

IEEE2019

Avoidance of Convex and Concave Obstacles with Convergence ensured through Contraction

Aude Billard, Lukas Huber

This paper presents a closed-form approach to obstacle avoidance for multiple moving convex and star-shaped concave obstacles. The method takes inspiration in harmonic-potential fields. It inherits the convergence properties of harmonic potentials. We prov ...

2019

Distributed Inference over Multitask Graphs under Smoothness

Ali H. Sayed, Stefan Vlaski, Roula Nassif

This paper formulates a multitask optimization problem where agents in the network have individual objectives to meet, or individual parameter vectors to estimate, subject to a smoothness condition over the graph. The smoothness requirement softens the tra ...

IEEE2018

THE LMMP FOR LOG CANONICAL 3-FOLDS IN CHARACTERISTIC p > 5

Joseph Allen Waldron

We prove that one can run the log minimal model program for log canonical 3-fold pairs in characteristic p > 5. In particular, we prove the cone theorem, contraction theorem, the existence of flips and the existence of log minimal models for pairs with log ...

2018

Optimal reinsurance with multiple tranches

Semyon Malamud

Motivated by common practices in the reinsurance industry and in insurance markets such as Lloyd's, we study the general problem of optimal insurance contracts design in the presence of multiple insurance providers. We show that the optimal risk allocation ...

Elsevier Science Sa2016

Decomposition Strategies for Nonconvex Problems, a Parametric Approach

Jean-Hubert Hours

This thesis deals with the development of numerical methods for solving nonconvex optimisation problems by means of decomposition and continuation techniques. We first introduce a novel decomposition algorithm based on alternating gradient projections and ...

EPFL2016

Large subsets of discrete hypersurfaces in Z d contain arbitrarily many collinear points

Florian Karl Richter

In 1977 L.T. Ramsey showed that any sequence in Z 2 with bounded gaps contains arbitrarily many collinear points. Thereafter, in 1980, C. Pomerance provided a density version of this result, relaxing the condition on the sequence from having bounded gaps t ...

2016

Graph Chatbot

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.