DeepAI AI Chat
Log In Sign Up

Evolution of Neural Networks to Play the Game of Dots-and-Boxes

by   Lex Weaver, et al.
Charles Sturt University
Australian National University

Dots-and-Boxes is a child's game which remains analytically unsolved. We implement and evolve artificial neural networks to play this game, evaluating them against simple heuristic players. Our networks do not evaluate or predict the final outcome of the game, but rather recommend moves at each stage. Superior generalisation of play by co-evolved populations is found, and a comparison made with networks trained by back-propagation using simple heuristics as an oracle.


page 1

page 2

page 3

page 4


Teaching Deep Convolutional Neural Networks to Play Go

Mastering the game of Go has remained a long standing challenge to the f...

Dots Boxes is PSPACE-complete

Exactly 20 years ago at MFCS, Demaine posed the open problem whether the...

Sprague-Grundy values and complexity for LCTR

Given a Young diagram on n boxes as a non-increasing sequence of integer...

Impartial games with entailing moves

Combinatorial Game Theory has also been called `additive game theory', w...

Adaptive Forgetting Factor Fictitious Play

It is now well known that decentralised optimisation can be formulated a...

Improving Hearthstone AI by Combining MCTS and Supervised Learning Algorithms

We investigate the impact of supervised prediction models on the strengt...

Agent Madoff: A Heuristic-Based Negotiation Agent For The Diplomacy Strategy Game

In this paper, we present the strategy of Agent Madoff, which is a heuri...