John H. Andreae - An AGI Brain for a Robot

John H. Andreae

with Robot Cats by Gillian M Andreae

Academic Press is an imprint of Elsevier

125 London Wall, London EC2Y 5AS, United Kingdom

525 B Street, Suite 1650, San Diego, CA 92101, United States

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

2021 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publishers permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN 978-0-323-85254-8

For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Nikki Levy

Acquisitions Editor: Natalie Farra

Editorial Project Manager: Sam Young

Production Project Manager: Punithavathy Govindaradjane

Cover Designer: Richard Roberts

Robot cat illustrations: Gillian Andreae

Typeset by SPi Global, India

1. Figure 1 One Step: Events, Event-types, and Modules
2. Figure 2 Short Term Memory
3. Figure 3 Neuron for Association 1GA
4. Figure 4 Part of a Network
5. Figure 5 Basic Cycle
6. Figure 6 Plan Cycle
7. Figure 7 The World
8. Figure 8 The Reflex Motion Algorithm
9. Figure 9 Hierarchies of Action-Predicting Groups
10. Figure 10 Event-Types and Delays of Speech Action Groups
11. Figure 11 Choosing Groups
12. Figure 12 Four steps from the Interaction
13. Figure 13 What robot A says in the first 200 steps
14. Figure 14 The Successes of PP (robot B) in the first 200 steps
15. Figure 15 Steps when Cake or Roll is Squashed
16. Figure 16 A Short Plan on Step 146

The writing of this book was triggered by a remark of Dan Dennett. I was much encouraged by his approval 2 years later. Generous support for publication of the book was also given by Michael Arbib and Ian Witten. David Hill, Andy Barto, my son Peter, and an anonymous reviewer sharpened my ideas. My daughter Gillian did a lot more than provide the cats. She and Suzanne Brown showed me how to cater for the general reader. My wife, Molly, spent many hours correcting my English, pointing out flaws in my logic, and supporting me, as she has done for the past 67 years.

My interest in Artificial Intelligence (AI) goes back a long way, but my research career didnt start with AI. After graduating in Electrical Engineering in 1948 at Imperial College, I joined John Lambs research group studying fast chemical equilibria in liquids using ultrasonic waves, and, with a PhD, continued the research in ICIs Akers Research Laboratories, Old Welwyn, North of London. My publications from that era have been cited more than my AI research, but ICI suddenly closed the laboratories in 1961. Eric Ash, a fellow student from undergraduate days, found a job for me with Standard Telecommunication Laboratories in Harlow, Essex. From a list of topics offered by Len Lewin, I was unable to resist the appeal of The Electronic Simulation of Cerebral Functions.

My first learning machine was published in 1963 and called STeLLA after the name of the laboratories. Peter Joyce built STeLLA, using post office relays because transistors werent yet generally available and vacuum tubes were too large. STeLLA moved around the laboratory floor, but the relays were too unreliable and so we had to resort to computer simulation. Brian Gaines deepened my understanding of control systems theory and stochastic computing.

With Mollys family being in Dunedin and my family having moved from India to Matamata, it was an easy decision to move to the University of Canterbury in Christchurch, New Zealand, when the opportunity arose in 1966.

Work on STeLLA continued with an attempt by Peter Cashin to make it powerful enough to handle language, but it was the mathematical ideas of John Cleary that inspired my first version of PURR-PUSS, now called PP, and my first book. Igor Aleksander enabled publication of my second book. Bruce MacDonald, Shaun Ryan, and Kon Kuiper contributed to the development of PP as recorded here and in my second book.

I am grateful to Natalie Farra of Elsevier for guiding this book through the stages of publication at a difficult time.

Many thanks to everyone who helped me with this book and in my research

Note. My daughter Gillian Andreaes robot cat illustrations first appeared, with her permission, in my Man-Machine Studies reports, 1972-1991, ISSN 0110 1188, then in my two books Thinking with the Teachable Machine and Associative Learning for a Robot Intelligence, and more recently on her commercial items.