Last round convergence and no-dynamic regret in asymmetric repeated games
Last round convergence and no-dynamic regret in asymmetric repeated games
This paper considers repeated games in which one player has a different objective than others. In particular, we investigate repeated two-player zero-sum games where the column player not only aims to minimize her regret but also stabilize the actions. Suppose that while repeatedly playing this game, the row player chooses her strategy at each round by using a no-regret algorithm to minimize her regret. We develop a no-dynamic regret algorithm for the column player to exhibit last round convergence to a minimax equilibrium. We show that our algorithm is efficient against a large set of popular no-regret algorithms the row player can use, including the multiplicative weights update algorithm, general follow-the-regularized-leader and any no-regret algorithms satisfy a property so called “stability”.
553-577
Dinh, Le Cong
e89b4443-9eff-4790-b101-9eabe5ef947c
Nguyen, Tri-Dung
a6aa7081-6bf7-488a-b72f-510328958a8e
Zemkoho, Alain B.
30c79e30-9879-48bd-8d0b-e2fbbc01269e
Tran-Thanh, Long
633282bf-f7ff-4137-ada6-6d4f19262676
1 April 2021
Dinh, Le Cong
e89b4443-9eff-4790-b101-9eabe5ef947c
Nguyen, Tri-Dung
a6aa7081-6bf7-488a-b72f-510328958a8e
Zemkoho, Alain B.
30c79e30-9879-48bd-8d0b-e2fbbc01269e
Tran-Thanh, Long
633282bf-f7ff-4137-ada6-6d4f19262676
Dinh, Le Cong, Nguyen, Tri-Dung, Zemkoho, Alain B. and Tran-Thanh, Long
(2021)
Last round convergence and no-dynamic regret in asymmetric repeated games.
Feldman, Vitaly, Ligett, Katrina and Sabato, Sivan
(eds.)
In Proceedings of the 32nd International Conference on Algorithmic Learning Theory.
vol. 132,
PMLR.
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
This paper considers repeated games in which one player has a different objective than others. In particular, we investigate repeated two-player zero-sum games where the column player not only aims to minimize her regret but also stabilize the actions. Suppose that while repeatedly playing this game, the row player chooses her strategy at each round by using a no-regret algorithm to minimize her regret. We develop a no-dynamic regret algorithm for the column player to exhibit last round convergence to a minimax equilibrium. We show that our algorithm is efficient against a large set of popular no-regret algorithms the row player can use, including the multiplicative weights update algorithm, general follow-the-regularized-leader and any no-regret algorithms satisfy a property so called “stability”.
Text
dinh21a
- Accepted Manuscript
Restricted to Repository staff only
Request a copy
More information
Accepted/In Press date: 21 December 2020
e-pub ahead of print date: 1 April 2021
Published date: 1 April 2021
Identifiers
Local EPrints ID: 448061
URI: http://eprints.soton.ac.uk/id/eprint/448061
PURE UUID: 46435bc8-66f5-43df-a6b5-0802bbf3fda3
Catalogue record
Date deposited: 01 Apr 2021 15:41
Last modified: 17 Mar 2024 03:37
Export record
Contributors
Author:
Le Cong Dinh
Author:
Tri-Dung Nguyen
Author:
Long Tran-Thanh
Editor:
Vitaly Feldman
Editor:
Katrina Ligett
Editor:
Sivan Sabato
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics