Finite Regret and Cycles with Fixed Step-Size via Alternating Gradient Descent-Ascent

07/09/2019
by   James P. Bailey, et al.
0

Gradient descent is arguably one of the most popular online optimization methods with a wide array of applications. However, the standard implementation where agents simultaneously update their strategies yields several undesirable properties; strategies diverge away from equilibrium and regret grows over time. In this paper, we eliminate these negative properties by introducing a different implementation to obtain finite regret via arbitrary fixed step-size. We obtain this surprising property by having agents take turns when updating their strategies. In this setting, we show that an agent that uses gradient descent obtains bounded regret – regardless of how their opponent updates their strategies. Furthermore, we show that in adversarial settings that agents' strategies are bounded and cycle when both are using the alternating gradient descent algorithm.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
11/08/2015

Speed learning on the fly

The practical performance of online stochastic gradient descent algorith...
research
08/12/2013

Faster gradient descent and the efficient recovery of images

Much recent attention has been devoted to gradient descent algorithms wh...
research
11/29/2021

Optimal No-Regret Learning in General Games: Bounded Regret with Unbounded Step-Sizes via Clairvoyant MWU

In this paper we solve the problem of no-regret learning in general game...
research
07/01/2020

Decentralised Learning with Random Features and Distributed Gradient Descent

We investigate the generalisation performance of Distributed Gradient De...
research
11/02/2022

Gradient Descent and the Power Method: Exploiting their connection to find the leftmost eigen-pair and escape saddle points

This work shows that applying Gradient Descent (GD) with a fixed step si...
research
11/16/2021

Polymatrix Competitive Gradient Descent

Many economic games and machine learning approaches can be cast as compe...
research
09/11/2019

An Implicit Form of Krasulina's k-PCA Update without the Orthonormality Constraint

We shed new insights on the two commonly used updates for the online k-P...

Please sign up or login with your details

Forgot password? Click here to reset