eCite Digital Repository

Generalisation over details: the unsuitability of supervised backpropagation networks for Tetris

Citation

Lewis, IJ and Beswick, SL, Generalisation over details: the unsuitability of supervised backpropagation networks for Tetris, Advances in Artificial Neural Systems, 2015 Article 157983. ISSN 1687-7594 (2015) [Refereed Article]


Preview
PDF
2Mb
  

Copyright Statement

Copyright 2015 I. J. Lewis and S. L. Beswick. This article is licensed under the terms of the Creative Commons Attribution License (CC BY 3.0 AU)

DOI: doi:10.1155/2015/157983

Abstract

We demonstrate the unsuitability of Artificial Neural Networks (ANNs) to the game of Tetris and show that their great strength, namely, their ability of generalization, is the ultimate cause. This work describes a variety of attempts at applying the Supervised Learning approach to Tetris and demonstrates that these approaches (resoundedly) fail to reach the level of performance of handcrafted Tetris solving algorithms. We examine the reasons behind this failure and also demonstrate some interesting auxiliary results. We show that training a separate network for each Tetris piece tends to outperform the training of a single network for all pieces; training with randomly generated rows tends to increase the performance of the networks; networks trained on smaller board widths and then extended to play on bigger boards failed to show any evidence of learning, and we demonstrate that ANNs trained via Supervised Learning are ultimately ill-suited to Tetris.

Item Details

Item Type:Refereed Article
Keywords:neural networks, Tetris, machine learning
Research Division:Information and Computing Sciences
Research Group:Artificial Intelligence and Image Processing
Research Field:Neural, Evolutionary and Fuzzy Computation
Objective Division:Information and Communication Services
Objective Group:Computer Software and Services
Objective Field:Computer Gaming Software
UTAS Author:Lewis, IJ (Dr Ian Lewis)
UTAS Author:Beswick, SL (Mr Sebastian Beswick)
ID Code:100063
Year Published:2015
Deposited By:Information and Communication Technology
Deposited On:2015-04-29
Last Modified:2017-11-01
Downloads:616 View Download Statistics

Repository Staff Only: item control page