University of Tasmania
Browse
157983.pdf (2.18 MB)

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

Download (2.18 MB)
journal contribution
posted on 2023-05-18, 09:28 authored by Ian LewisIan Lewis, Beswick, SL
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.

History

Publication title

Advances in Artificial Neural Systems

Volume

2015

Article number

157983

Number

157983

Pagination

1-8

ISSN

1687-7594

Department/School

School of Information and Communication Technology

Publisher

Hindawi Publishing Corporation

Place of publication

United States

Rights 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)

Repository Status

  • Open

Socio-economic Objectives

Animation, video games and computer generated imagery services

Usage metrics

    University Of Tasmania

    Categories

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC