Read Anytime Anywhere Easy Ebook Downloads at ebookluna.

com

Deep Learning on Edge Computing Devices: Design

Challenges of Algorithm and Architecture 1st
Edition Xichuan Zhou - eBook PDF

https://ebookluna.com/download/deep-learning-on-edge-
computing-devices-design-challenges-of-algorithm-and-
architecture-ebook-pdf/

OR CLICK HERE

DOWLOAD EBOOK

Visit and Get More Ebook Downloads Instantly at https://ebookluna.com

 Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Artificial Intelligence and Machine Learning for EDGE

Computing 1st Edition Rajiv Pandey - eBook PDF

https://ebookluna.com/download/artificial-intelligence-and-machine-
learning-for-edge-computing-ebook-pdf/

ebookluna.com

(eBook PDF) Deep Learning for Medical Image Analysis by S.

Kevin Zhou

https://ebookluna.com/product/ebook-pdf-deep-learning-for-medical-
image-analysis-by-s-kevin-zhou/

ebookluna.com

(eBook PDF) Algorithm Design by Jon Kleinberg

https://ebookluna.com/product/ebook-pdf-algorithm-design-by-jon-
kleinberg/

ebookluna.com

Fractional-order Design: Devices, Circuits, And Systems

1st Edition - eBook PDF

https://ebookluna.com/download/fractional-order-design-devices-
circuits-and-systems-ebook-pdf/

ebookluna.com
Medical image recognition, segmentation and parsing :
machine learning and multiple object approaches 1st
Edition Zhou - eBook PDF
https://ebookluna.com/download/medical-image-recognition-segmentation-
and-parsing-machine-learning-and-multiple-object-approaches-ebook-pdf/

ebookluna.com

(Original PDF) Algorithm Design and Applications by

Michael T. Goodrich

https://ebookluna.com/product/original-pdf-algorithm-design-and-
applications-by-michael-t-goodrich/

ebookluna.com

Artificial Intelligence and Deep Learning in Pathology 1st

Edition Stanley Cohen Md (Editor) - eBook PDF

https://ebookluna.com/download/artificial-intelligence-and-deep-
learning-in-pathology-ebook-pdf/

ebookluna.com

Deep Learning in Bioinformatics: Techniques and

Applications in Practice - eBook PDF

https://ebookluna.com/download/deep-learning-in-bioinformatics-
techniques-and-applications-in-practice-ebook-pdf/

ebookluna.com

Circular Design for Zero Emission Architecture and

Building Practice 1st Edition - eBook PDF

https://ebookluna.com/download/circular-design-for-zero-emission-
architecture-and-building-practice-ebook-pdf/

ebookluna.com
DEEP LEARNING
ON EDGE
COMPUTING
DEVICES
This page intentionally left blank
DEEP LEARNING
ON EDGE
COMPUTING
DEVICES
Design Challenges of Algorithm
and Architecture

XICHUAN ZHOU
HAIJUN LIU
CONG SHI
JI LIU
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
Copyright © 2022 Tsinghua University Press. Published by 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 Publisher’s 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: 978-0-323-85783-3

For information on all Elsevier publications

visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mara Conner

Acquisitions Editor: Glyn Jones
Editorial Project Manager: Naomi Robertson
Production Project Manager: Selvaraj Raviraj
Designer: Christian J. Bilbow
Typeset by VTeX
Contents

Preface vii
Acknowledgements ix

PART 1 Introduction

1. Introduction 3
1.1. Background 3
1.2. Applications and trends 5
1.3. Concepts and taxonomy 8
1.4. Challenges and objectives 13
1.5. Outline of the book 14
References 16

2. The basics of deep learning 19

2.1. Feedforward neural networks 19
2.2. Deep neural networks 22
2.3. Learning objectives and training process 29
2.4. Computational complexity 33
References 34

PART 2 Model and algorithm

3. Model design and compression 39

3.1. Background and challenges 39
3.2. Design of lightweight neural networks 40
3.3. Model compression 47
References 56

4. Mix-precision model encoding and quantization 59

4.1. Background and challenges 59
4.2. Rate-distortion theory and sparse encoding 61
4.3. Bitwise bottleneck quantization methods 65
4.4. Application to efficient image classification 67
References 73

5. Model encoding of binary neural networks 75

5.1. Background and challenges 75

v
vi Contents

5.2. The basic of binary neural network 77

5.3. The cellular binary neural network with lateral connections 79
5.4. Application to efficient image classification 84
References 92

PART 3 Architecture optimization

6. Binary neural network computing architecture 97

6.1. Background and challenges 97
6.2. Ensemble binary neural computing model 98
6.3. Architecture design and optimization 102
6.4. Application of binary computing architecture 105
References 108

7. Algorithm and hardware codesign of sparse binary network

on-chip 111
7.1. Background and challenges 111
7.2. Algorithm design and optimization 115
7.3. Near-memory computing architecture 120
7.4. Applications of deep adaptive network on chip 124
References 135

8. Hardware architecture optimization for object tracking 139

8.1. Background and challenges 139
8.2. Algorithm 140
8.3. Hardware implementation and optimization 143
8.4. Application experiments 147
References 152

9. SensCamera: A learning-based smart camera prototype 155

9.1. Challenges beyond pattern recognition 155
9.2. Compressive convolutional network model 159
9.3. Hardware implementation and optimization 164
9.4. Applications of SensCamera 166
References 175

Index 179
Preface

We first started working in the field of edge computing-based machine

learning in 2010. With project funding, we tried to accelerate support
vector machine algorithms on integrated circuit chips to support embed-
ded applications such as fingerprint recognition. In recent years, with the
development of deep learning and integrated circuit technology, artificial
intelligence applications based on edge computing devices, such as intel-
ligent terminals, autonomous driving, and AIOT, are emerging one after
another. However, the realization of an embedded artificial intelligence ap-
plication involves multidisciplinary knowledge of mathematics, computing
science, computer architecture, and circuit and system design. Therefore
we arrived at the idea of writing a monograph focusing on the research
progress of relevant technologies, so as to facilitate the understanding and
learning of graduate students and engineers in related fields.
Deep learning application development based on embedded devices is
facing the theoretical bottleneck of high complexity of deep neural network
algorithms. Realizing the lightweight of various fast developing deep learn-
ing models is one of the keys to realize AIOT pervasive artificial intelligence
in the future. In recent years, we have been focusing on the development
of automated deep learning tools for embedded devices. This book covers
some of the cutting-edge technologies, currently developing in embedded
deep learning, and introduces some core algorithms, including lightweight
neural network design, model compression, model quantization, etc., aim-
ing to provide reference for the readers to design embedded deep learning
algorithm.
Deep learning application development based on embedded devices is
facing the technical challenge of limited development of integrated circuit
technology in the post-Moore era. To address this challenge, in this book,
we propose and elaborate a new paradigm of algorithm-hardware codesign
to realize the optimization of energy efficiency and performance of neural
network computing in embedded devices. The DANoC sparse coding neu-
ral network chip developed by us is taken as an example to introduce the
new technology of memory computing, hoping to give inspiration to em-
bedded design experts. We believe that, in the post-Moore era, the system
collaborative design method across multiple levels of algorithms, software,
and hardware will gradually become the mainstream of embedded intelli-
vii
viii Preface

gent design to meet the design requirements of high real-time performance

and low power consumption under the condition of limited hardware re-
sources.
Due to time constraints and the authors’ limited knowledge, there may
be some omissions in the content, and we apologize to the readers for this.

Xichuan Zhou
Acknowledgements

First of all, we would like to thank all the students who participated in the
relevant work for their contributions to this book, including Shuai Zhang,
Kui Liu, Rui Ding, Shengli Li, Songhong Liang, Yuran Hu, etc.
We would like to take the opportunity to thank our families, friends,
and colleagues for their support in the course of writing this monograph.
We would also like to thank our organization, School of Microelectron-
ics and Communication Engineering, Chongqing University, for providing
supportive conditions to do research on intelligence edge computing.
The main content of this book is compiled from a series of research,
partly supported by the National Natural Science Foundation of China
(Nos. 61971072 and 62001063).
We are most grateful to the editorial staff and artists at Elsevier and
Tsinghua University Press for giving us all the support and assistance needed
in the course of writing this book.

ix
Visit https://testbankfan.com
now to explore a rich
collection of testbank or
solution manual and enjoy
exciting offers!
This page intentionally left blank
PART 1

Introduction

1
This page intentionally left blank
CHAPTER 1

Introduction
1.1 Background
At present, the human society is rapidly entering the era of Internet of
Everything. The application of the Internet of Things based on the smart
embedded device is exploding. The report “The mobile economy 2020”
released by Global System for Mobile Communications Assembly (GSMA)
has shown that the total number of connected devices in the global Inter-
net of Things reached 12 billion in 2019 [1]. It is estimated that by 2025
the total scale of the connected devices in the global Internet of Things
will reach 24.6 billion. Applications such as smart terminals, smart voice
assistants, and smart driving will dramatically improve the organizational
efficiency of the human society and change people’s lives. With the rapid
development of artificial intelligence technology toward pervasive intelli-
gence, the smart terminal devices will further deeply penetrate the human
society.
Looking back at the development process of artificial intelligence, at
a key time point in 1936, British mathematician Alan Turing proposed
an ideal computer model, the general Turing machine, which provided
a theoretical basis for the ENIAC (Electronic Numerical Integrator And
Computer) born ten years later. During the same period, inspired by
the behavior of the human brain, American scientist John von Neumann
wrote the monograph “The Computer and the Brain” [2] and proposed
an improved stored program computer for ENIAC, i.e., Von Neumann
Architecture, which became a prototype for computers and even artificial
intelligence systems.
The earliest description of artificial intelligence can be traced back to
the Turing test [3] in 1950. Turing pointed out that “if a machine talks
with a person through a specific device without communication with the
outside, and the person cannot reliably tell that the talk object is a machine
or a person, this machine has humanoid intelligence”. The word “artificial
intelligence” actually appeared at the Dartmouth symposium held by John
McCarthy in 1956 [4]. The “father of artificial intelligence” defined it as
“the science and engineering of manufacturing smart machines”. The pro-
posal of artificial intelligence has opened up a new field. Since then, the
Deep Learning on Edge Computing Devices Copyright © 2022 Tsinghua University Press.
https://doi.org/10.1016/B978-0-32-385783-3.00008-9 Published by Elsevier Inc. All rights reserved. 3
4 Deep Learning on Edge Computing Devices

Figure 1.1 Relationship diagram of deep learning related research fields.

academia has also successively presented research results of artificial intel-

ligence. After several historical cycles of development, at present, artificial
intelligence has entered a new era of machine learning.
As shown in Fig. 1.1, machine learning is a subfield of theoretical re-
search on artificial intelligence, which has developed rapidly in recent years.
Arthur Samuel proposed the concept of machine learning in 1959 and con-
ceived the establishment of a theoretical method “to allow the computer
to learn and work autonomously without relying on certain coded instruc-
tions” [5]. A representative method in the field of machine learning is the
support vector machine [6] proposed by Russian statistician Vladimir Vap-
nik in 1995. As a data-driven method, the statistics-based SVM has perfect
theoretical support and excellent model generalization ability, and is widely
used in scenarios such as face recognition.
Artificial neural network (ANN) is one of the methods to realize
machine learning. The artificial neural network uses the structural and
functional features of the biological neural network to build mathematical
models for estimating or approximating functions. ANNs are computing
systems inspired by the biological neural networks that constitute animal
brains. An ANN is based on a collection of connected units or nodes
called artificial neurons, which loosely model the neurons in a biologi-
cal brain. The concept of the artificial neural network can be traced back
to the neuron model (MP model) [7] proposed by Warren McCulloch
and Walter Pitts in 1943. In this model the input multidimensional data
are multiplied by the corresponding weight parameters and accumulated,
Random documents with unrelated
content Scribd suggests to you:
 Hence, the policy of Campaign of that army superior in Numbers,
is:
Incessantly to proffer battles which:

(a) Accepted, constantly reduces the inferior army and increases

its disproportion in numbers, or,
(b) Evaded, compels the inferior army to abandon important
posts, for whose defence it cannot afford the resulting loss
of troops; thus permitting to the numerically superior army
a continually increasing advantage in Position.

PRINCIPLE
All else being equal the advantage of Numbers is decisive of victory
in battle and Campaign.

Things being unequal, the advantage in Numbers may be nullified by

adverse advantages in Organization, Topography, Mobility, Time
and Position.

Victory resulting from advantage in Numbers is achieved by

simultaneously attacking two or more Tactical Keys from a
Kindred Strategic Key and two or more Kindred Points of
Command.

TO LOCATE THE AREA OF CONCENTRATION

That relative advantage in Mobility expressed by the situation of

the Strategic Front upon the Strategetic Center is materially
manifested upon the Chess-board by Kindred Chess-pieces posted
upon that great central diagonal which extends towards the
Objective Plane. Such advantage determines those points which
should be occupied in the proper development of the front so
posted; and consequently designates the direction and location of
that battlefield upon which the kindred army may concentrate in
overwhelming force, despite all possible resistance by the enemy.

MOST FAVORABLE BATTLEFIELD

That relative advantage in Organization expressed by superior

potential totality, is materially manifested upon the Chess-board by
the geometric and sub-geometric symbols of those Chess-pieces
possessed of the superior potential complement. Such symbols taken
in combination, describe that field of battle most favorable for the
execution of those Major Tactical evolutions which appertain to the
Chess-pieces of superior organization.

POSTS OF MAXIMUM SECURITY

That relative defensive advantage in Topography expressed by

inaccessibility to hostile attack is materially manifested upon the
Chess-board by Corps of Position, posted upon points of different
color to that occupied by the adverse Bishop; and this advantage
designates those posts situated on a projected field of battle which
may be occupied with the maximum of security.
That relative offensive advantage in Topography expressed by
accessibility to kindred attack is materially manifested upon the
Chess-board by Corps of Position posted upon points of the same
color as that occupied by the kindred Bishop; and this advantage
designates those posts situated on a projected field of battle which
may be attacked with the maximum facility.

CHARACTER OF THE MOST FAVORABLE BATTLE

That relative advantage in Position with the Column of Attack,

expressed:

1. By superior location, direction and development of the

Kindred Strategic Front of Operations; and
 2. By the occupation of Points of Departure, of Manoeuvre, of
Command and of the Strategic Key of a True Strategic
Horizon, indicates that a Strategic Grand Battle in the first
instance; and in the second case that a Tactical Grand
Battle is most favorable in the existing situation.

That relative advantage in Position with the Column of Support,

expressed by superior facilities for occupying with the Kindred
Promotable Factors their corresponding Points of Junction in the
Kindred Logistic Horizon, is materially manifested upon the Chess-
board by the larger number of Pawn Altitudes which either are open,
or may be opened, despite all possible resistance by the enemy; and
such advantage designates those adverse Points of Impenetrability
and Points of Resistance to the march of the Kindred Promotable
Factors, which it is necessary to nullify.
That relative advantage in Position with the Column of Manoeuvre,
expressed by the security of the Kindred and the exposure of the
adverse Strategetic Rear to attack by the Kindred Column of
Support, is materially manifested upon the Chess-board by the
occupation by a Kindred Promotable Factor of the Point of Proximity;
and such advantage indicates that the advance with all possible
celerity of such Promotable Factor and Point of Proximity toward the
corresponding Point of Junction is a dominating influence in the
existing situation.

PROJECTED GRAND BATTLE

From the advantage in Position appertaining to the three Grand

Columns is deduced the character of the Grand Battle properly in
sequence.
Advantage in Position with the Column of Attack indicates the
opportunity, all else being equal, to engage in a victorious Strategic
Grand Battle against the hostile Formation in Mass, or in a Tactical
Grand Battle against the hostile Formation by Wings.
 Advantage in Position with the Column of Support indicates the
opportunity to engage effectively in a series of minor battles, as
though having the advantage in Numbers.
Advantage in Position with the Column of Manoeuvre indicates the
opportunity to engage in a victorious Logistic Grand Battle against
the adverse Formation by Grand Columns.

LEAST FAVORABLE ADVERSE CONDITION

That relative advantage in Time expressed by restrictions of the

adversary’s choice of movements at his turn to play, is materially
manifested upon the Chess-board by Feints operated by Kindred
Chess-pieces against adverse vital points; and such advantage of the
Initiative dictates the next move of the opposing army.
The advantage of the Initiative determines which of the adverse
corps d’armee may and may not move.
The material expression of this advantage always is a Feint by a
Kindred Corps against a vital point either occupied or unoccupied,
which necessitates that upon his next move, the enemy either
evacuate, support, cover or sustain the post so menaced.
Such feint, therefore, restricts the move of the enemy to those of
his corps as are able to obviate the threatened loss and
proportionately reduces the immediate activity of his army.

RELATIVE ADVANTAGES IN LOCATION

“It is only the force brought into action that avails in battles
and campaigns—the rest does not count.”—Napoleon.

The distance which separates opposing Corps d’armee always

modifies the values of the Prime Strategetic Means.
 Hence in the making of Grand Reconnaissance, it is next in
sequence to determine whether the Chess-pieces are:

I. In Contact.
II. In Presence.
III. At Distance.

Corps d’armee are in Contact with each other whenever their

logistic radii intersect; or, their radii offensive and the corresponding
adverse radii defensive are opposed to each other.
Corps d’armee are in Presence whenever the posts which they
occupy are contained within the same Strategic front, the same
Strategetic Horizon, or are in communication with their
corresponding posts of mobilization, development, or manoeuvre.
Corps d’armee are at Distance when the posts which they occupy
are not in communication with Kindred Corps d’armee posted upon
the strategic front adopted, or with posts of mobilization or
development contained within the corresponding Primary Base of
Operations, or, within the True Strategetic Horizon.

REQUISITES FOR SUCCESSFUL CAMPAIGNING

Every Campaign, whether upon the surface of the Earth or upon

the Chess-board is decided and usually is terminated by a Grand
Battle.
Those movements of opposing Grand Columns, whereby such
decisive conflict is brought about under circumstances which ensure
victory, by reason of superior advantages in Strategetic Means, are
termed Grand Manoeuvres; and a proper series of Grand
Manoeuvres, combined with their corresponding feints, strategems,
ambuscades and minor battles, the whole terminated by a resulting
Grand Battle, is termed a Grand Operation.
 Those processes of Grand Manoeuvre, which produce an
opportunity to victoriously engage in battle, are the most subtle and
difficult known to the Strategetic Art.
Successful application of these processes in practice depends
wholly upon proper use of the MEANS at hand and the doing of the
utmost that can be done in the TIME available.
Nothing can be more repugnant to high art in Strategetics than
those crudities termed in the specious mouthings of pretentious
mediocrity “waiting moves,” “delayed strokes,” “defensive-
offensives,” “masterly inactivities,” and the like.
“Time past is gone and cannot be regained; time future is not and
may never be; time present is” and with it Opportunity, which an
instant later may be gone.
The gain of but “a foot of ground and a minute of time” would
have saved the French army at Rosbach and have cost Frederic the
Great one of his most lustrous victories and perhaps his army and
his crown.

PRINCIPLE
In Strategetics there is but a single method whereby Opportunity
may be availed of, and that is by so augmenting kindred
advantages and so depreciating adverse advantages as to
acquire for the kindred army that particular advantage of
Strategetic means which in the given situation is the proper
basis of the Strategetic movement next in sequence.

At Distance.

The chief requisite for success when acting against an adverse army
at Distance, is the advantage in MOBILITY.
 The primary process is that of a Grand Manoeuvre against an
adverse army acting in the formation by Grand Columns, and the
object of such Grand Manoeuvre always is, by superior celerity of
movement, to occupy:

1. The Strategic Center by the Kindred Column of Attack, thus

intersecting the Route of Communication between the
adverse main body and its Base of Operation; or to occupy:
2. The Logistic Center with the Kindred Columns of Support and
of Manoeuvre, thus intersecting the Route of
Communication between the adverse main body and its
Chief Supporting Column and clearing the way for the
advance of the Kindred Column of Support against the flank
and rear of the adverse Main Body.

Obviously, the united Kindred Columns of Attack and of Support

always will constitute an overwhelming superiority in Numbers as
compared with the adverse main body.

In Presence.

The chief requisite for success when acting against an adverse

Grand Column in Presence, is the advantage in POSITION.

The primary process is that of a Grand Manoeuvre against an

adverse army acting in the Formation by Wings, and the object of
such Grand Manoeuvre always is, by availing to the uttermost of its
situation upon the Tactical Center, i.e., upon the area midway
between the adverse Wings thus isolated from each other; to act in
overwhelming Numbers, first against one and then against the other
hostile bodies.

In Contact.
The chief requisite for success when acting offensively against an
adverse Grand Column, or Wing, or Corps d’armee, in Contact,
is the advantage in NUMBERS.

The primary process is that of a Grand Battle in which the kindred

army has an overwhelming superiority in Numbers in contact, and at
least the equality in all other Prime Strategetic Means.
In this circumstance, the object of such Grand Battle always is:

1. To attack the hostile Formation in Mass frontally at the center,

and upon both wings obliquely; all three attacks being
made simultaneously and the evolutions so executed that
the hostile army never is able to penetrate between either
kindred wing and the kindred center, nor to outflank that
kindred wing which may be in the air.
2. In case the kindred army has the equality or inferiority in all
other Prime Strategetic Means, then the object of a Grand
Battle on the Offensive is to attack the hostile Formation in
Mass obliquely with the whole kindred army, and preferably
upon that wing which covers the route of communication of
the adverse army with its Base of Operations, but always
upon that wing which contains the Tactical Key of the actual
Battlefield.

Obviously, the concentration of the entire kindred army against a

single adverse wing always will constitute an overwhelming
superiority in Numbers.
In making such attack obliquely against a single adverse wing, the
center and remaining wing of the kindred army must not engage
until the kindred Van and Corps of Position of the attacking wing first
have formed the center of three sides of an octagon; of which the
Kindred Corps of Evolution will form the farthest side and the
Kindred Center and left wing Corps d’armee will form the nearest
and latest constructed side.
The chief requisite for success when acting defensively against a
Grand Column, or Wing, or Corps d’armee is the advantage in
TOPOGRAPHY.

The primary process is that of a Grand Battle in which the kindred

army, decidedly inferior in Numbers in the aggregate, has the
advantage in Topography and equality in all other Prime Strategetic
Means.
In this case the object is to support both flanks of the inferior
army upon impassable natural barriers, strengthening both wings at
the expense of the center, both in quantity and in quality of troops.
If the Tactical Defensive be selected, the center should retire
before the oncoming of the hostile army in order to enclose it
between the Kindred Wings, which will then overwhelm it by
superior Numbers, while the natural barriers on the flanks being
impassable will prevent the remaining hostile corps from
participating in the battle otherwise than as spectators.
Should the Tactical Offensive be selected, that kindred wing best
adapted for attack should engage supported by all kindred Corps of
Evolution, while advancing the Kindred Center in reserve and holding
the remaining wing refused and in observation.
All else being equal, relative advantage in either branch of Prime
Strategetic Means is sufficient to ensure victory in battle, and the
proper use of such advantage for securing victory is outlined thus:

PRINCIPLE
Utilize advantage in Prime Strategetic Means to obtain the
superiority in Numbers at the Point of Contact in an Offensive
Battle; and to nullify the adverse superiority in Numbers at the
point of contact in a Defensive Battle.
 Between War and Chess there is a seeming incongruity, which is
the basis of that doubt of the utility of Chess-play, so commonly held
by laymen, and which fallacy few, even among proficients, are
competent to combat.
This doubt most frequently is voiced by the query:
 If Chess and War are analagous, why was not
Napoleon a Master Chess-player and Morphy a great
military Commander?
This query readily is answered in the words of Frederic the Great,
viz.:
“To be possessed of talent is not sufficient.
Opportunity to display such talent and to its full extent
is necessary. All depends on the time in which we live.”
The Strategetic talent possessed in common by Morphy and
Napoleon, in both was brought to perfection by long and expert
training.
But circumstances placed the twelve year old Napoleon in the
midst of soldiers and in an era of war, while circumstances placed
the twelve year old Morphy in the midst of Chess-players and in an
era of Peace.
Napoleon was educated a General; Morphy was educated a lawyer.
To develop his self-evident and superlative Strategetic talent,
Napoleon’s education was of the best; to develop his self-evident
and superlative Strategetic talent, Morphy’s education was of the
worst.
Napoleon succeeded as a General; Morphy failed as a lawyer.
The innate capability of Napoleon for Strategetics was developed
in the direction of Warfare; the innate capability of Morphy for
Strategetics was developed in the direction of Chess-play.
In War, Napoleon is superlative; in Chess, Morphy is superlative.
Educated in the law, Napoleon might have proved like Morphy a
non-entity; educated in Chess, Napoleon might have proved like
Morphy a phenomenon.
Educated in War, Morphy might have rivalled Napoleon.
 For the Chess-play of Morphy displays that perfect comprehension
of Strategetics, to which none but the great Captains in warfare have
attained.
Perfection in Strategetics consists in exactly interpreting in battle
and campaign, the System of Warfare invented by Epaminondas.
Those able to do this in War have achieved greatness, and the
great at Chess-play are those who best have imitated that exactness
with which Morphy employed this system on the Chess-board.
To those who imagine that Strategetic talent, as exemplified in
Warfare, is different from Strategetic talent as exemplified in Chess-
play, the following may afford matter for reflection.

“Frederic the Great was one of the finest Chess-players that

Germany ever produced.”—Wilhelm Steinitz.
PRIME STRATEGETIC PROPOSITION
SECTION ONE
PRIME STRATEGETIC PROPOSITION
SECTION ONE

(First Phase.)
In the consideration of every Strategetic Situation possible in
Warfare, or in Chess-play, the initial process always is a Grand
Reconnaissance.
Grand Reconnaissance is that exact scrutiny of existing conditions,
whereby is determined the relative advantages and disadvantages
possessed by the opposing armies in:

1. Time.
2. Numbers.
3. Position.
4. Organization.
5. Mobility.
6. Topography.

The First Phase in the demonstration of every Prime Strategetic

Proposition consists:

1. In determining by comparison of the relative advantages and

disadvantages in Time, which of the opposing armies has
the ability to MOVE, while the other must remain stationary.
2. In deducing the MOTIF of such movement.
3. In designating the DIRECTION of such movement.
 The making of Grand Reconnaissance is a special privilege which
exclusively appertains to the advantage in Time. It always should be
made by the Commander-in-chief of that army which is able to put
itself in motion, while the opposing army must remain stationary,
and it never should be confounded with the advance of the Cavalry
Corps, nor confused with the work of scouts and spies; all of which
are matters entirely separate and distinct from Grand
Reconnaissance.
In the Grand Reconnaissance of any given Strategetic Situation
the element of Numbers primarily is to be considered, for the reason
that the basic fact of the Science of Strategetics is:
“THE GREATER FORCE ALWAYS OVERCOMES THE
LESSER.”—Napoleon.
Hence, unless more immediately vital considerations prevent,
superiority in Numbers, of itself, is decisive of victory; and thus it
readily is to be deduced that all else being equal, the advantage in
Time plus the advantage in Numbers constitutes the easiest and
simplest winning combination known to Strategetic Art.
But it so happens that the advantage in Time may be combined
not only with the greater force, but also with an equal, or even with
the lesser force, and from this it is self-evident that Strategetic
Situations are divided into three classes, viz.:

I. Numerical superiority, plus right to move.

II. Numerical equality, plus right to move.
III. Numerical inferiority, plus right to move.

There are two primary methods for availing of superiority in

Numbers to destroy the opposing lesser force, viz.:

1. By the Process of Attrition, i.e., by maintaining an incessant

tactical offensive and thus wearing down the opposing
army by exchanging pieces at every opportunity.
 2. By Acting in Detachments, i.e., by means of the extra corps,
simultaneously to attack more points of vital importance
than the hostile army is able simultaneously to defend.

From the foregoing it is obvious that conversely there are two

principal considerations, which all else being equal, must dominate
the procedure of the Numerically inferior force, viz.:

I. To avoid further diminution of its aggregate.

II. To avoid creating indefensible vital points.

The second consideration in the making of a Grand

Reconnaissance by the commander-in-chief of an army having the
advantage in Time, is the element of Position; for the reason that by
unscientific posting of Corps d’armee, relative advantages in Time,
or in Numbers or in both, may be rendered nugatory, on account of
inability of the kindred Columns of Attack, of Support and of
Manoeuvre to perform their functions.
In case the Corps are scientifically posted and are in positions to
avail of advantage in Time and Numbers, those adverse vital points
whose occupation may be effected by superior force, always will be
the objectives of the movements of the latter.
Hence, the following:

PRINCIPLE
As the advantage in Time gives the right to MOVE and the
advantage in Numbers indicates the MOTIF of movement; so
does the advantage in Position, as expressed by the Strategic
Syllogism, specify the DIRECTION of that movement which
normally appertains to the army having the advantage in Time.
 The proper direction of that movement which normally appertains
to the advantage in Time always is indicated by the plus signs in the
Strategic Syllogism, viz.:

+A. Signifies that the Normal direction of movement for the

army having the advantage in Time is along the
Strategetic Center towards the Objective Plane.
+S. Signifies that the Normal direction of movement for the
army having the advantage in Time is along one or
more pawn altitudes towards the Kindred Logistic
Horizon.
+M. Signifies that the Normal direction of movement for the
army having the advantage in Time is along the
shortest open pawn altitude towards the Kindred Point
of Junction.
+A+S. Signifies that the Normal direction of movement for the
army having the advantage in Time is double, i.e.,
+A. towards the Objective Plane.
+S. along one or more open Pawn altitudes toward
the Kindred Logistic Horizon.
+A+M. Signifies that the Normal direction of movement for the
army having the advantage in Time, is double, i.e.,
+A. towards the Objective Plane.
+M. Along the shortest open pawn altitude toward
the Kindred Point of Junction.
+S+M. Signifies that the Normal direction of movement for the
army having the advantage in Time is double, i.e.,
+S. Along one or more open Pawn altitudes toward
the Kindred Logistic Horizon.
+M. Along the shortest open Pawn Altitude toward
the Kindred Point of Junction.
+A+S+M. Signifies that the Normal direction of movement is
triple, i.e.,
 +A. Toward the Objective Plane.
+S. Along one or more open Pawn altitudes, toward
the Kindred Logistic Horizon.
+M. Along the shortest open Pawn altitude, toward
the Kindred Point of Junction.

The First Phase in the demonstration of every Strategetic

Proposition is determined by the following:

THEOREM
Given the Normal ability to move, to determine the Normal motif and
direction of movement.

1. Designate that army having the advantage in Time and express

such advantage by the symbol +T, express the corresponding
disadvantage in Time which appertains to the opposing army, by the
symbol -T, and such symbols will constitute the First Term of the
First Phase of the demonstration of any Prime Strategetic
Proposition.
2. Express that superiority, equality, or inferiority in Numbers,
which appertains to each of the opposing armies by the symbols +N,
=N,-N, respectively; and such symbols will constitute the Second
Term of the First Phase of the demonstration of any Prime
Strategetic Proposition.
3. Express the objectives designated by the plus terms of the
Strategic Syllogism, viz.:

(a) Objective of +A = Objective Plane, i.e., O. P.

(b) ” +S = Logistic Horizon, i.e., L. H.
(c) ” +M = Point of Junction, i.e., P. J.
and the symbols denoting such objectives will constitute the Third
Term in the First Phase of the demonstration of any Prime
Strategetic Proposition.
4. Combine those three terms which appertain to the advantage in
Time, then combine those three terms which appertain to the
disadvantage in Time, and the resulting equation when expanded
will depict:

(a) The normal ability to move.

(b) The normal motif of movement.
(c) The normal directions of movement which appertain to each
of the opposing armies.

EXAMPLE

White. (+T+N) + (+A+S+M)

Black. (-T-N) + (-A-S-M)

EXPANDED

First Term. +T = Normal ability to move.

Second Term. +N = Normal motif of movement,
(a) Detachments, (b) Exchanges.
Third Term +O. P. = Normal objective of +A.
+L. H. = ” ” ” +S.
+P. J. = ” ” ” +M.

Hence, in the foregoing example the normal direction of

movement for White may be either toward the Objective Plane with
the Column of Attack, or toward the Logistic Horizon, or the Point of
Junction with the Column of Support, or toward both objectives, with
both columns simultaneously.
 Meanwhile, the Black army having the disadvantage in Time is
unable to move, and consequently is stationary.
Furthermore, White having the superiority in Numbers may move
with an equal force against either objective designated by the Third
Term of the equation, and with his excess force against one or more
adverse vital points, simultaneously, against which latter movement,
Black obviously has no adequate defence.

TACTICO-LOGISTIC INEQUALITY
The Tactico-Logistic Inequality is the algebraic expression of the
relative advantages and disadvantages in Time and in Numbers
appertaining to opposing Strategetic Entireties.
Such advantages and disadvantages are denoted by the terms,
viz.:

+T. Signifies the absolute advantage in Time, i.e., the ability of an

army, a grand column, a wing or a corps d’armee to move,
while the opposing force must remain stationary.
-T. Signifies the absolute disadvantage in Time, i.e., the
obligation of an army, a grand column, a wing, or a corps
d’armee to remain stationary, while the opposing force is in
motion.
+N. Signifies the absolute advantage in Numbers, i.e., the larger
number of corps d’armee.
-N. Signifies the absolute disadvantage in Numbers, i.e., the
smaller number of corps d’armee.
=N. Signifies the equality in Numbers, i.e., the same number of
corps d’armee.

There are six forms of the Tactico-Logistic Inequality, viz.:

1. +T+N
 -T-N
2. +T=N
-T=N
3. +T-N
-T+N
4. -T+N
+T-N
5. -T=N
+T=N
6. -T-N
+T+N

INITIAL STRATEGETIC EQUATION

The Initial Strategetic Equation is made up of those terms which

compose the Strategic Syllogism and the Tactico-Logistic Inequality,
viz.:

(+A+S+M) + (+T+N) - (-A-S-M) + (-T-N) =

the Normal Motif and Direction of Effort.

RULE
1. Set down in parenthesis those terms of the Strategic Syllogism
which appertain to White.
Set down in parenthesis those terms of the Tactico-Logistic
Inequality which appertain to White.
Connect the two kindred terms thus constructed, by the sign of
addition, to show that each is to augment the other, and superscore
all by the same vincula to show that all are to be taken together to
form one side of the resulting equation.
2. Repeat this process for the Black terms to construct the second
side of the Initial Strategetic Equation, and separate the White from
the Black terms by a minus sign.

STRATEGETIC VALUES
The Strategetic Values of the terms contained in the Strategic
Syllogism and in the Tactico-Logistic Inequality are shown by the
appended tables, viz.:

TABLE OF STRATEGIC VALUES.

Term. Post. Direction. Motif.

1. +A Grand Vertex Tactical Key of To give checkmate
Objective Plane
2. +M Point Proximity Point of Junction To queen a Pawn
en command
3. +A Major Vertex 1. Grand Vertex To gain winning
2. Point Aligned Position with
3. Point en Column of Attack
Potence
4. +M Point Proximity Point en To gain winning
en Menace Command Position with
Column of Support
5. +M Point Proximity Point en Menace To gain winning
en Presence Position with
Column of Support
6. +A Minor Vertex 1. Major Vertex To gain Superior
2. Point Aligned Position with
Column of Attack
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookluna.com