Neural Networks and Deep Learning A Textbook 2nd

Edition Charu C. Aggarwal download

https://ebookmeta.com/product/neural-networks-and-deep-learning-
a-textbook-2nd-edition-charu-c-aggarwal/

Download more ebook from https://ebookmeta.com

We believe these products will be a great fit for you. Click
the link to download now, or visit ebookmeta.com
to discover even more!

Neural Networks and Deep Learning: A Textbook, 2nd

Edition Charu C. Aggarwal

https://ebookmeta.com/product/neural-networks-and-deep-learning-
a-textbook-2nd-edition-charu-c-aggarwal-2/

Machine Learning for Text 2nd Edition Charu C. Aggarwal

https://ebookmeta.com/product/machine-learning-for-text-2nd-
edition-charu-c-aggarwal/

Machine Learning for Text, 2nd Edition Charu Aggarwal

https://ebookmeta.com/product/machine-learning-for-text-2nd-
edition-charu-aggarwal/

Alternative Dispute Resolution and Peace building in

Africa 1st Edition Ernest Uwazie

https://ebookmeta.com/product/alternative-dispute-resolution-and-
peace-building-in-africa-1st-edition-ernest-uwazie/
Journal of Literature Culture Media Studies N.D.R.
Chandra (Editor-In-Chief)

https://ebookmeta.com/product/journal-of-literature-culture-
media-studies-n-d-r-chandra-editor-in-chief/

Girl At War Sara Novi■

https://ebookmeta.com/product/girl-at-war-sara-novic/

Enhancing Student Support in Higher Education: A

Subject-Focused Approach 1st Edition Nick Pilcher

https://ebookmeta.com/product/enhancing-student-support-in-
higher-education-a-subject-focused-approach-1st-edition-nick-
pilcher/

Collective Bargaining in Professional Sports Player

Salaries Free Agency Team Ownership League
Organizational Structures and the Power of
Commissioners 1st Edition Scott Bukstein
https://ebookmeta.com/product/collective-bargaining-in-
professional-sports-player-salaries-free-agency-team-ownership-
league-organizational-structures-and-the-power-of-
commissioners-1st-edition-scott-bukstein/

National Geographic Readers Helen Keller Level 2

Kitson Jazynka

https://ebookmeta.com/product/national-geographic-readers-helen-
keller-level-2-kitson-jazynka/
One Last Smile 1st Edition Gs Rhodes

https://ebookmeta.com/product/one-last-smile-1st-edition-gs-
rhodes/
Charu C. Aggarwal

Neural Networks and Deep Learning

A Textbook
2nd ed. 2023
Charu C. Aggarwal
IBM T. J. Watson Research Center, International Business Machines,
Yorktown Heights, NY, USA

ISBN 978-3-031-29641-3 e-ISBN 978-3-031-29642-0

https://doi.org/10.1007/978-3-031-29642-0

© Springer Nature Switzerland AG 2018, 2023

This work is subject to copyright. All rights are reserved by the

Publisher, whether the whole or part of the material is concerned,
specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other
physical way, and transmission or information storage and retrieval,
electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks,

service marks, etc. in this publication does not imply, even in the
absence of a specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general
use.

The publisher, the authors, and the editors are safe to assume that the
advice and information in this book are believed to be true and accurate
at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the
material contained herein or for any errors or omissions that may have
been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer

Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham,
Switzerland
To my wife Lata, my daughter Sayani,
and my late parents Dr. Prem Sarup and Mrs. Pushplata Aggarwal.
Preface
“Any A.I. smart enough to pass a Turing test is smart enough to
know to fail it.”–***Ian McDonald

Neural networks were developed to simulate the human nervous

system for machine learning tasks by treating the computational units
in a learning model in a manner similar to human neurons. The grand
vision of neural networks is to create artificial intelligence by building
machines whose architecture simulates the computations in the human
nervous system. Although the biological model of neural networks is an
exciting one and evokes comparisons with science fiction, neural
networks have a much simpler and mundane mathematical basis than a
complex biological system. The neural network abstraction can be
viewed as a modular approach of enabling learning algorithms that are
based on continuous optimization on a computational graph of
mathematical dependencies between the input and output. These ideas
are strikingly similar to classical optimization methods in control
theory, which historically preceded the development of neural network
algorithms.
Neural networks were developed soon after the advent of
computers in the fifties and sixties. Rosenblatt’s perceptron algorithm
was seen as a fundamental cornerstone of neural networks, which
caused an initial period of euphoria — it was soon followed by
disappointment as the initial successes were somewhat limited.
Eventually, at the turn of the century, greater data availability and
increasing computational power lead to increased successes of neural
networks, and this area was reborn under the new label of “deep
learning.” Although we are still far from the day that artificial
intelligence (AI) is close to human performance, there are specific
domains like image recognition, self-driving cars, and game playing,
where AI has matched or exceeded human performance. It is also hard
to predict what AI might be able to do in the future. For example, few
computer vision experts would have thought two decades ago that any
automated system could ever perform an intuitive task like categorizing
an image more accurately than a human. The large amounts of data
available in recent years together with increased computational power
have enabled experimentation with more sophisticated and deep
neural architectures than was previously possible. The resulting
success has changed the broader perception of the potential of deep
learning. This book discusses neural networks from this modern
perspective. The chapters of the book are organized as follows:
1. The basics of neural networks: Chapters 1, 2, and 3 discuss the
basics of neural network design and the backpropagation
algorithm. Many traditional machine learning models can be
understood as special cases of neural learning. Understanding the
relationship between traditional machine learning and neural
networks is the first step to understanding the latter. The
simulation of various machine learning models with neural
networks is provided in Chapter 3. This will give the analyst a feel
of how neural networks push the envelope of traditional machine
learning algorithms.

2. Fundamentals of neural networks: Although Chapters 1, 2, and 3

provide an overview of the training methods for neural networks, a
more detailed understanding of the training challenges is provided
in Chapters 4 and 5. Chapters 6 and 7 present radial-basis function
(RBF) networks and restricted Boltzmann machines.

3. Advanced topics in neural networks: A lot of the recent success of

deep learning is a result of the specialized architectures for various
domains, such as recurrent neural networks and convolutional
neural networks. Chapters 8 and 9 discuss recurrent and
convolutional neural networks. Graph neural networks are
discussed in Chapter 10. Several advanced topics like deep
reinforcement learning, attention mechanisms, neural Turing
mechanisms, and generative adversarial networks are discussed in
Chapters 11 and 12.

We have included some “forgotten” architectures like RBF networks

and Kohonen self-organizing maps because of their potential in many
applications. An application-centric view is highlighted throughout the
book in order to give the reader a feel for the technology.
What Is New in The Second Edition The second edition has focused
on improving the presentations in the first edition and also on adding
new material. Significant changes have been made to almost all the
chapters to improve presentation and add new material where needed.
In some cases, the material in different chapters has been reorganized
and reordered in order to improve exposition. The discussion of
training challenges with depth has been separated from the
backpropagation chapter, so that both topics could be discussed in
greater detail. Second-order methods have been explained with greater
clarity and examples. Chapter 10 on graph neural networks is entirely
new. The discussion on GRUs in Chapter 8 has been greatly enhanced.
New convolutional architectures using attention, such as Squeeze-and-
Excitation Networks, are discussed in Chapter 9. The chapter on
reinforcement learning has been significantly reorganized in order to
provide a clearer exposition. The Monte Carlo sampling approach for
reinforcement learning is discussed in greater detail in its own
dedicated section. Chapter 12 contains expanded discussions on
attention mechanisms for graphs and computer vision, transformers,
pre-trained language models, and adversarial learning.

Notations Throughout this book, a vector or a multidimensional data

point is annotated with a bar, such as or . A vector or
multidimensional point may be denoted by either small letters or
capital letters, as long as it has a bar. Vector dot products are denoted
by centered dots, such as . A matrix is denoted in capital letters
without a bar, such as R. Throughout the book, the n × d matrix
corresponding to the training data set is denoted by D, with n
documents and d dimensions. The individual data points in D are
therefore d-dimensional row vectors. On the other hand, vectors with
one component for each data point are usually n-dimensional column
vectors. An example is the n-dimensional column vector of class
variables. An observed value yi is distinguished from a predicted value
by a circumflex at the top of the variable.
Charu C. Aggarwal
Yorktown Heights, NY, USA
Author Biography
Charu C. Aggarwal is a Distinguished Research Staff Member (DRSM)
at the IBM T. J. Watson Research Center in Yorktown Heights, New York.
He completed his undergraduate degree in Computer Science from the
Indian Institute of Technology at Kanpur in 1993 and his Ph.D. from the
Massachusetts Institute of Technology in 1996.

He has worked extensively in the field of data mining. He has

published more than 400 papers in refereed conferences and journals
and authored over 80 patents. He is the author or editor of 20 books,
including textbooks on data mining, recommender systems, and outlier
analysis. Because of the commercial value of his patents, he has thrice
been designated a Master Inventor at IBM. He is a recipient of an IBM
Corporate Award (2003) for his work on bio-terrorist threat detection
in data streams, a recipient of the IBM Outstanding Innovation Award
(2008) for his scientific contributions to privacy technology, and a
recipient of two IBM Outstanding Technical Achievement Awards
(2009, 2015) for his work on data streams/high-dimensional data. He
received the EDBT 2014 Test of Time Award for his work on
condensation-based privacy-preserving data mining. He is a recipient of
the IEEE ICDM Research Contributions Award (2015) and ACM SIGKDD
Innovation Award, which are the two most prestigious awards for
influential research contributions in the field of data mining. He is also
a recipient of the W. Wallace McDowell Award, which is the highest
award given solely by the IEEE Computer Society across the field of
Computer Science.
 He has served as the general co-chair of the IEEE Big Data
Conference (2014) and as the program co-chair of the ACM CIKM
Conference (2015), the IEEE ICDM Conference (2015), and the ACM
KDD Conference (2016). He served as an associate editor of the IEEE
Transactions on Knowledge and Data Engineering from 2004 to 2008.
He is an associate editor of the IEEE Transactions on Big Data, an action
editor of the Data Mining and Knowledge Discovery Journal, and an
associate editor of the Knowledge and Information Systems Journal. He
has served or currently serves as the editor-in-chief of the ACM
Transactions on Knowledge Discovery from Data as well as the ACM
SIGKDD Explorations. He is also an editor-in-chief of ACM Books. He
serves on the advisory board of the Lecture Notes on Social Networks, a
publication by Springer. He has served as the vice-president of
the SIAM Activity Group on Data Mining and is a member of the
SIAM industry committee. He received the ACM SIGKDD Service Award
for the aforem3entioned contributions to running conferences and
journals— this honor is the most prestigious award for services to the
field of data mining. He is a fellow of the SIAM, ACM, and the IEEE, for
“contributions to knowledge discovery and data mining algorithms.”
Acknowledgments
Acknowledgements for the First Edition I would like to thank my
family for their love and support during the busy time spent in writing
this book. I would also like to thank my manager Nagui Halim for his
support during the writing of this book.
Several figures in this book have been provided by the courtesy of
various individuals and institutions. The Smithsonian Institution made
the image of the Mark I perceptron (cf. Figure 1.​5) available at no cost.
Saket Sathe provided the outputs in Chapter 8 for the tiny Shakespeare
data set, based on code available/described in [? ? ]. Andrew Zisserman
provided Figures 9.​13 and 9.​17 in the section on convolutional
visualizations. Another visualization of the feature maps in the
convolution network (cf. Figure 9.​16) was provided by Matthew Zeiler.
NVIDIA provided Figure 11.​10 on the convolutional neural network for
self-driving cars in Chapter 11, and Sergey Levine provided the image
on self-learning robots (cf. Figure 11.​9) in the same chapter. Alec
Radford provided Figure 12.​11, which appears in Chapter 12. Alex
Krizhevsky provided Figure 9.​9(b) containing AlexNet.
This book has benefitted from significant feedback and several
collaborations that I have had with numerous colleagues over the years.
I would like to thank Quoc Le, Saket Sathe, Karthik Subbian, Jiliang
Tang, and Suhang Wang for their feedback on various portions of this
book. Shuai Zheng provided feedback on the section on regularized
autoencoders in Chapter 5. I received feedback on the sections on
autoencoders from Lei Cai and Hao Yuan. Feedback on the chapter on
convolutional neural networks was provided by Hongyang Gao,
Shuiwang Ji, and Zhengyang Wang. Shuiwang Ji, Lei Cai, Zhengyang
Wang and Hao Yuan also reviewed the Chapters 4 and 8, and suggested
several edits. They also suggested the ideas of using Figures 9.​6 and 9.​7
for elucidating the convolution/deconvolution operations.
For their collaborations, I would like to thank Tarek F. Abdelzaher,
Jinghui Chen, Jing Gao, Quanquan Gu, Manish Gupta, Jiawei Han,
Alexander Hinneburg, Thomas Huang, Nan Li, Huan Liu, Ruoming Jin,
Daniel Keim, Arijit Khan, Latifur Khan, Mohammad M. Masud, Jian Pei,
Magda Procopiuc, Guojun Qi, Chandan Reddy, Saket Sathe, Jaideep
Srivastava, Karthik Subbian, Yizhou Sun, Jiliang Tang, Min-Hsuan Tsai,
Haixun Wang, Jianyong Wang, Min Wang, Suhang Wang, Joel Wolf,
Xifeng Yan, Mohammed Zaki, ChengXiang Zhai, and Peixiang Zhao. I
would also like to thank my advisor James B. Orlin for his guidance
during my early years as a researcher.
I would like to thank Lata Aggarwal for helping me with some of the
figures created using PowerPoint graphics in this book. My daughter,
Sayani, was helpful in incorporating special effects (e.g., image color,
contrast, and blurring) in several JPEG images used at various places in
this book.

Acknowledgements for the Second Edition The chapter on graph

neural networks benefited from collaborations with Jiliang Tang,
Lingfei Wu, and Suhang Wang. Shuiwang Ji, Meng Liu, and Hongyang
Gao provided detailed comments on the chapter on graph neural
networks as well as other parts of the book. Zhengyang Wang and
Shuiwang Ji provided feedback on the new section on transformer
networks. Yao Ma and Jiliang Tang provided the permission to use
Figure 10.​3. Yao Ma also provided detailed comments on the chapter on
graph neural networks. Sharmishtha Dutta, a PhD student at RPI,
proofread Chapter 10, and also pointed out the sections that needed
further clarity. My manager, Horst Samulowitz, provided support
during the writing of the second edition of this book.
Contents
1 An Introduction to Neural Networks
1.​1 Introduction
1.​2 Single Computational Layer:​The Perceptron
1.​2.​1 Use of Bias
1.​2.​2 What Objective Function Is the Perceptron
Optimizing?​
1.​3 The Base Components of Neural Architectures
1.​3.​1 Choice of Activation Function
1.​3.​2 Softmax Activation Function
1.​3.​3 Common Loss Functions
1.​4 Multilayer Neural Networks
1.​4.​1 The Multilayer Network as a Computational Graph
1.​5 The Importance of Nonlinearity
1.​5.​1 Nonlinear Activations in Action
1.​6 Advanced Architectures and Structured Data
1.​7 Two Notable Benchmarks
1.​7.​1 The MNIST Database of Handwritten Digits
1.​7.​2 The ImageNet Database
1.​8 Summary
1.​9 Bibliographic Notes and Software Resources
1.​10 Exercises
2 The Backpropagation Algorithm
2.​1 Introduction
2.​2 The Computational Graph Abstraction
2.​2.​1 Computational Graphs Create Complex Functions
2.​3 Backpropagation in Computational Graphs
2.​3.​1 Computing Node-to-Node Derivatives with the Chain
Rule
2.​3.​2 Dynamic Programming for Computing Node-to-
NodeDerivatives
2.​3.​3 Converting Node-to-Node Derivatives into Loss-to-
Weight Derivatives
2.​4 Backpropagation in Neural Networks
2.​4.​1 Some Useful Derivatives of Activation Functions
2.​4.​2 Examples of Updates for Various Activations
2.​5 The Vector-Centric View of Backpropagation
2.​5.​1 Derivatives with Respect to Vectors
2.​5.​2 Vector-Centric Chain Rule
2.​5.​3 A Decoupled View of Vector-Centric Backpropagation
2.​5.​4 Vector-Centric Backpropagation with Non-
LayeredArchitect​ures
2.​6 The Not-So-Unimportant Details
2.​6.​1 Mini-Batch Stochastic Gradient Descent
2.​6.​2 Learning Rate Decay
2.​6.​3 Checking the Correctness of Gradient Computation
2.​6.​4 Regularization
2.​6.​5 Loss Functions on Hidden Nodes
2.​6.​6 Backpropagation Tricks for Handling Shared Weights
2.​7 Tuning and Preprocessing
2.​7.​1 Tuning Hyperparameters
2.​7.​2 Feature Preprocessing
2.​7.​3 Initialization
 2.​8 Backpropagation Is Interpretable
2.​9 Summary
2.​10 Bibliographic Notes and Software Resources
2.​11 Exercises
3 Machine Learning with Shallow Neural Networks
3.​1 Introduction
3.​2 Neural Architectures for Binary Classification Models
3.​2.​1 Revisiting the Perceptron
3.​2.​2 Least-Squares Regression
3.​2.​2.​1 Widrow-Hoff Learning
3.​2.​2.​2 Closed Form Solutions
3.​2.​3 Support Vector Machines
3.​2.​4 Logistic Regression
3.​2.​5 Comparison of Different Models
3.​3 Neural Architectures for Multiclass Models
3.​3.​1 Multiclass Perceptron
3.​3.​2 Weston-Watkins SVM
3.​3.​3 Multinomial Logistic Regression (Softmax Classifier)
3.​4 Unsupervised Learning with Autoencoders
3.​4.​1 Linear Autoencoder with a Single Hidden Layer
3.​4.​1.​1 Connections with Singular Value Decomposition
3.​4.​1.​2 Sharing Weights in the Encoder and Decoder
3.​4.​2 Nonlinear Activation Functions and Depth
3.​4.​3 Application to Visualization
3.​4.​4 Application to Outlier Detection
3.​4.​5 Application to Multimodal Embeddings
Exploring the Variety of Random
Documents with Different Content
 THE MOSQUITO.

T HERE is nothing in the appearance of the mosquito to excite

alarm even in the most timid breasts, no sign of his almost
diabolical nature, or of his power of making himself obnoxious. And
yet he is endowed with a subtlety, a malice, and a fiendish thirst for
blood unparalleled save in the leech. The mosquito is found in
almost every climate and country, sounding his trumpet as
vehemently by the shores of the Arctic Sea as beside a sluggish
stream on the Equator, the British Islands being almost alone in their
happy immunity from its presence; and among all the varied
blessings for which a Briton has cause to be thankful there is
scarcely one so peculiar and so marked as the absence of this
creature. It is probably seen at its worst in the north of Russia,
Norway and Sweden, and in some of the Northern States of
America. In these countries it is hardly safe to leave a horse out at
night, for although we may safely discredit the legends that horses
have been carried off bodily by mosquitoes, these animals have
undoubtedly been killed by the poisonous bites of their innumerable
foes. It is the methods of the mosquito rather than the injury it
inflicts that drive men to madness. It is not that they are greatly
grudged the drop or two of blood they extract, and the pain and
inflammation of the wound, though often considerable, are not very
much more so than those of our own midnight assailants, the bug
and the flea. If they would but come and have their meal in peace
and quiet, man might bear it. It is their shrill trumpeting, their
approaches and departures, and the long and agonising suspense
that precedes the moment when, their investigation complete, they
fix on what appears to them the most penetrable point, settle, and
begin their meal, that cows the spirit of the bravest man. Heroes
who would face the spring of an infuriated tiger, and lead a column
to the cannon’s mouth, will quail and cover their head with the sheet
when they hear the shrill challenge of the mosquito.
Man has endeavoured by many means to defend himself from this
persecutor. He has rubbed himself with medicaments, and has hung
up boughs of shrubs to which it is supposed that the mosquito has
an objection. He has invented pastilles, whose smoke, it was hoped,
would lull his foe into a lethargy; but at all these and similar
measures the mosquito laughs. The only resource affording even a
partial protection is the mosquito curtain. In theory this device is
excellent. Man enclosed within a curtain of gauze ought to be
unassailable. Unfortunately the practice does not follow the theory.
However secure the curtains, however great the pains bestowed in
seeing that no mosquito was present when the man was tucked up
inside them, we doubt whether history records a single example of
complete success having attended the arrangement. Do what man
will, the mosquito will be there. Its favourite plan is to be
beforehand with a man, and to hide somewhere until man has
entered his muslin tent. Every effort will, it knows, be made to
dislodge it; the curtains will be shaken, towels will be flapped here
and there, every nook and corner will, as it seems, be examined, but
the mosquito will manage in one way or other to evade the search.
But even in the exceptional cases where it is routed out, the
mosquito knows that it is but for a time. If there is a hole in the
curtains, be it only the size of a knitting-needle, it will find it and get
through; and in the event of the curtains being absolutely new, it is
sure to find some point at which the tucking up has been imperfectly
done. But most of all it relies upon entering with the would-be
sleeper. The latter is well aware of this. He listens first for the sound
of wings, but at this moment the mosquito is discreetly silent. Then
he untucks a small portion of the curtain, his attendant flaps a towel
wildly, and under cover of this he plunges hastily through the orifice,
which is at once closed behind him. Then, in spite of a thousand
similar experiences, the man flatters himself that this time he has
evaded the mosquito, and lies down to rest. Stronger and stronger
grows the hope as the minutes pass on, and at last it almost blooms
into certainty as he finally turns over and composes himself for
sleep. Drowsiness steals over him, when, just as consciousness is
leaving him, the mosquito sounds a triumphant bugle-blast close to
his ear. Then the ordinary man sits up in bed as if he were shot, and
swears. This is, unfortunately, all but universal. The best and most
patient of men have found it absolutely impossible to avoid using
bad language at this crisis. There is a shout for the attendant, a light
is brought and placed on a table near the curtain. Then the battle
begins in grim earnest, the man against the mosquito; the one silent
and watchful, his arms outside the sheet ready for instant action, the
other, agile, ubiquitous, intent on exasperating and not on attacking
its victim, now resting for a time in a corner, then making a rapid
dash at the nose or ear, then disappearing again, and lying silent for
some minutes. Occasionally, very occasionally, the man is victor, and
with a rapid clutch will grasp and annihilate the mosquito as it
passes by his face. In the vast majority of cases the man’s
watchfulness is in vain. Hours pass, and Nature asserts herself. The
mosquito has had amusement enough, and now, meaning business,
remains quiet until its victim dozes off. Not until he is sound asleep
will it this time move. Then it settles lightly upon him, inserts its
delicate proboscis in one of the pores of his skin, pours in a tiny drop
of venom to dilute the blood, and then having drunk till its body has
swelled to many times its original size, heavily flies away, and
fastens itself to the curtain, where it falls an easy victim to the
vengeance of the sleeper in the morning. Such is the conflict when
one mosquito has found an entrance. When, as is more usual, half a
dozen have entered, it is, as may be imagined, still more dire and
disastrous; and the sleeper in the morning wakes with perhaps an
eye closed, and his face swollen and disfigured by bumps almost
beyond knowledge.
The existence of the mosquito can be accounted for only upon the
ground that he was sent as a special trial to man’s temper, but in
that case Nature evidently miscalculated the amount of self-control
that man possesses. A trial can hardly be considered as a trial when
the result is certain, and the breakdown of man’s temper under the
attacks of the mosquito is universal and complete. It would have
been enough had the mosquito been endowed with activity, craft,
and voracity. The trial would have been in that case ample, but
exceptional men might have passed through it unscathed. It was the
addition of the trumpet that settled the matter. No such exasperating
sound is to be heard on earth. Good resolutions crumble to nought
before it. The most patient and the most stoical of mortals are as
much moved by it as their weaker brethren, and the native of the
Arctic Circle and he of the Equator alike in their respective languages
utter words of despair and profanity. We may hope, however, that
science has not yet spoken its last word, and that some future
Pasteur or Koch may discover a bacillus capable of creating a
contagious and fatal disease among mosquitoes, and that by this
means man may be relieved of a burden almost too heavy for him.
 THE COW.

A LTHOUGH the cow is always with us, we know but little about her
beyond her likes and dislikes in the matter of food. We have,
indeed, by dint of long perseverance, transformed the wild cow into
an eating machine—a vehicle for the conversion of feeding stuffs into
milk and meat. Her brain is to us a sealed book, which so far no sage
has made it his business to open. No one, however, can doubt that
the cow does a great deal of thinking. In this respect it is among
beasts as is the owl among birds. No one can watch a herd of cattle
ruminating tranquilly, without being impressed with the conviction
that they are thinking deeply. Whether they are meditating over the
legends that have been handed down to them of the time when they
wandered wild and free on mountain and moor, or are wondering
why man busies himself in supplying them with the food most to
their liking, while he requires no active service in return, as he does
from the horse, we know not.
The eye of the ox is soft and meditative; it has not inspired modern
poets, but the ancients recognised its beauty, and the Greeks could
find no more complimentary epithet for the Queen of the Gods than
to call her ox-eyed. Such an eye should certainly indicate a
philosophic mind, and it is in this direction that we must regard it as
probable that the cow’s ruminations are directed. We may credit her
with having arrived at a conclusion to her own satisfaction as to the
points that have engaged the attention of a Darwin or a Spencer, but
one can scarce conjecture that the cerebral organisation of the cow
was beforehand with man in the discovery of the steam-engine or the
electric telegraph. The Arabs and the Orientals, with their deep
knowledge of the occult, were evidently impressed with the idea that
the cow’s brain is so stored with knowledge that it would be a danger
to mankind were she able to put her thoughts into words. This is
shown by the fact that, while in their legends the gift of speech is
frequently bestowed on horses, storks, and birds of many kinds,
there is no instance of a cow being so favoured. It may be said that
the dog is similarly omitted; but the dog is an animal looked down
upon in the East. It is there never admitted to the intimacy of man,
and, having been habitually repressed, has not acquired the traits of
character that distinguish it in Western countries. But in whatever
light the matter is looked at, it cannot be doubted that it is
unfortunate for the world that so profound a thinker as the cow is
unable to communicate her conclusions to man.
 The cow, as distinct from the bull, is in its wild state a timid animal,
and it is somewhat singular that although she has lost much of that
timidity, she largely inspires the feeling among the female sex. Next
to the mouse, the ordinary woman fears the cow. The dog, a really
more alarming animal, she is not afraid of; the horse inspires her
with no terror; but the sight of two or three cows in a lane throws
her off her balance. On such an occasion a woman will perform feats
of activity quite beyond her at ordinary times: she will climb a five-
barred gate, or squeeze herself through a gap in a hedge, regardless
of rents or scratches, with as much speed and alacrity as she would
manifest in leaping on a chair in the presence of that ferocious
animal the mouse. We believe that this unreasoning terror has its
origin in the pernicious nursery legend of the cow with the crumpled
horn. It is true that that animal is related to have suffered the
maiden all forlorn to milk her, but she afterwards tossed the dog; and
it is the pictorial representations of her while performing this feat that
have impressed the juvenile mind. The mere fact that there are few
precedents for a woman being tossed by a cow goes for nothing, nor
that the animal’s disposition is peaceable in the extreme; it can,
therefore, be hardly questioned that the timidity excited in the female
mind by the cow must be founded upon some lost legend of
antiquity. It may be that Eve had trouble in her first efforts to procure
lacteal fluid from the cow, or that the specimen chosen to perpetuate
the race in the Ark was rendered savage and dangerous from its long
imprisonment there; but no legend that would give favour to either
theory has come down to us.
In her wild state the cow is compelled to take considerable exercise
in order to obtain a sufficient amount of sustenance; the
domesticated animal, having no need to do so, has developed habits
of laziness. She has become constitutionally averse to exertion; but
Providence, by sending the fly, has done much to counteract the
effects of this tendency. It has been calculated by mechanical
engineers that the amount of energy required to switch away flies
with a cow’s tail is equivalent to that which would raise a weight of
seven pounds one foot. Intelligent observers estimate that upon a
hot day when the flies are troublesome, a cow will switch her tail
thirty times in the course of a minute, thus expending an amount of
energy per hour sufficient, if otherwise employed, to lift nearly six
tons’ weight one foot from the ground; so that, considering the
number of cows in Great Britain, it is clear that an amount of power
in comparison to which that of Niagara is as nothing is being wasted.
The thoughtful agriculturist will surely perceive that as an
expenditure of energy means loss of flesh and decreased production
of milk, it would be to his interest to envelop his cattle in mosquito
curtains during the summer months.
The cow is best seen in a state of repose. Either as lying down or
standing in the shade of a tree, dreamily chewing the cud, and
vaguely wondering whether beet or turnips will form the staple of her
supper, there are few animals more taking to the eye. She can walk,
too, without forfeiting our respect, but she is a lamentable spectacle
when she runs. The poetry of motion does not exist in the case of
the cow, and yet it is clear that she takes the greatest pains about
her running, and puts her whole heart into it; personally, then, she is
not to blame in that the result is, as an exhibition, a failure. The fault
lies in nature rather than in the individual. In the course of the
Darwinian process of transforming, let us say a mole into a cow, it
was clearly in the creature’s mind that the day would come when she
would be milked. Each of the countless generations required to bring
her to her present form kept this contingency steadily in view, and
practised kicking sideways. The result is, so far as the milkmaid is
concerned, a superb success, and the cow is able to kick sideways in
a manner that excites the envious admiration of the horse; but, as
was to be expected, with the acquisition of the sideway motion the
cow’s leg lost the power possessed so pre-eminently by the horse
and mule of delivering a good, fair, square kick backwards; and even
in running, what may be called the side action predominates over the
fore and aft. Doubtless the cow knew her own business, and
deliberately sacrificed gracefulness of action to the joy of being able
to kick over a milkmaid. The lover of grace may regret that it should
be so, but has no right to complain of the cow pleasing herself. The
original mole probably foresaw that her far-off descendant would be
a creature of few active enjoyments, and of a steady and tranquil
nature, and considered that she was perfectly justified in making
some sacrifice in order to enable the cow of the future to enjoy at
least one piece of lively fun.
 On the whole, however, the cow may fairly claim to be an
eminently worthy and respectable animal, and to be of great
importance to man. Some may feel inclined to say, of vital
importance; but this may be disputed. It is due in a great degree to
the attention that man has bestowed upon her that she has
developed her capacity for putting on flesh, and her abnormal
secretion of milk. Had man not found her ready to his hand, and
foreseen her capacity in this direction, he might have turned his
attention to the mastodon, which in that case would now be grazing
in vast numbers among the woods planted for his sustenance, and
would be affording mountains of flesh and tuns of milk, while
mastodon butter might have been able to hold its own against
margarine and other fatty compounds. The cow deserves great credit
for developing herself into her wild type from some wandering germ
or other, but for her progression to her present status she has to
thank the care and attention she has received from man.
 THE OCTOPUS AND CUTTLE FISH.

A LTHOUGH dignified by the name of a fish, the cuttle fish has

nothing in common with the finny inhabitants of the sea, save that
its existence is passed beneath the surface of the water. It stands
alone, apart from all living creatures, with scarcely a point of
resemblance to any of them, its nearest relations being, perhaps, the
sea anemones—those lovely inhabitants of pools among rocks.
Nature would seem to have created the octopus in an idle moment,
in order to show how she could diverge from her regular course, and
turn out a creature with a multiplicity of arms, without body or legs,
and with its head in the middle of its stomach. As usual, she
succeeded to perfection, but was so horrified with the monster she
had made that she threw it into the sea, and endowed it with a
diabolical disposition. The octopus resembles an ogre dwelling in its
cave, conscious that its distorted shape will not bear the light, and
stretching out its arms studded with suckers to grasp and draw down
to its mouth any living thing that passes within its reach. The cuttle
fish varies in size from the squid, beloved by gourmands who dwell
on the shores of the Mediterranean, to the monster octopus who
throws his arms round boats and drags them to the bottom. Some,
indeed, in the Indian seas, are reported to grow to a size that
renders them formidable even to ships, wrapping them in its embrace
and dragging the sailors from the deck or shrouds. Even allowing for
exaggeration, there can be little doubt that enormous specimens are
occasionally met with, and that these would be formidable to small
vessels. Bodies have been cast ashore whose arms have measured
thirty feet in length, and these could well pluck a sailor from the deck
of a ship. On our own shores they are, happily, never met with of
formidable size, but comparatively large ones are encountered not far
south; for it may be taken that the desperate struggle described by
Victor Hugo in “The Toilers of the Sea” was at least not considered by
him to be impossible, and that he had heard from fishermen of the
existence of creatures as large as the one he described. The octopus
appears almost insensible to pain, and the hacking off of one or more
of its tentacles does not seem to cause it any inconvenience. Its body
—or rather its stomach—is its only vital part, and even this must be
almost cut into pieces before it will relinquish the hold it has obtained
of a prey. The beak of a parrot is the last thing one would expect to
find in the centre of these waving tentacles, and Nature apparently
placed it there as the crowning effort in the work of construction of
this monster.
Among birds, beasts, and fishes we may seek in vain for a
prototype of the octopus. To find one we must go to man, and we
shall find that, in his way, the professional money-lender bears a
close resemblance to this creature. The waving arms, that by their
resemblance to great seaweeds lull a passing fish into a sense of
security, are represented in the case of the money-lender by
flattering and unctuous advertisements, which, catching the eye of
the unwary, persuade him that money is to be had for asking, upon
terms to suit all pockets; but, as in the case of the octopus, once the
suckers catch hold, there is no escape; nearer and nearer the victim
is drawn, in spite of his struggles, to the parrot mouth that will tear
him to pieces, and swallow up him and his belongings. The analogy is
in all ways extremely close, and yet the man who would shudder at
the thought of entering a cave in the depth of whose waters the
octopus is lurking, will enter the professional money-lender’s den with
an unmoved countenance and an even pulse. Happily, there is every
reason for supposing that the fish which form the staple of the diet of
the octopus suffer less in the process of destruction than does the
victim of the money-lender. Fish are certainly almost, if not entirely,
insensible to pain, and there is no reason to suppose that they are
gifted with strong powers of imagination; it may therefore be
believed that although a fish may struggle to escape from the grip of
the tentacle, it feels none of the horror that seizes a human victim
when once grasped by one of the larger species, and that its doom is
hidden from it until the savage beak seizes it, and at once puts an
end to its existence.
While man can to a certain extent enter into the feelings of a large
proportion of the animal creation, it is beyond his power to imagine
himself an octopus, or to get himself en rapport with its thoughts.
Has it any higher impulses? Is it naturally cruel, or does it view its
own methods and conduct from a strictly business point? Does it
persuade itself that it is an estimable character? Is it in its own
private circle affectionate and domesticated? Has it the power of
discussing passing events with its congeners, and exchanging views
as to the flavour of the various fish that form its diet, or as to
advantageous spots for ambush? We can answer none of these
questions. It certainly has but a small chance of leading a higher life.
The subterranean world it sees around it is full of strife and
destruction. “The large fish eat the smaller fish, and so on ad
infinitum.” It only plays the same game as those around it, but by
different methods, and there is no reason, because those methods
are repugnant to us, that the octopus should be of the same opinion.
Man is singularly intolerant in such matters. He himself kills the
creatures he requires for food either by knocking them on the head,
by cutting their throats, or by shooting them. Fish he captures either
with nets or with a hook which sticks into their mouth or throat. And
yet he criticises severely the methods of the animal creation. He
dislikes the spider because like a fisherman it catches its prey in nets.
He shudders at the cat because it plays with its victim just as the
angler does. He is shocked because the octopus lies in wait for its
prey and lassoes it as it passes. There is, in fact, no pleasing man,
and he is shocked at all methods of killing, even at that most closely
resembling those which he himself employs in slaying the creatures
on which he feeds. We fear that there is a great deal of humbug
about human susceptibilities.
Some of the cuttle fish are large manufacturers of ink. These,
instead of anchoring themselves to the bottom, float near the
surface, and their chance of obtaining food would be small were it
not for their power of ejecting ink, and thus clouding the water and
veiling themselves from sight—a habit which also affords them a
method of escape when themselves attacked by the shark or other
formidable enemy. This method is not unknown to man, and several
well-known instances might be adduced of public men who, after
having by loose assertions brought a formidable opponent down
upon them, escape under a cloud of misleading words, phrases, and
explanations that explain nothing, and retractions that leave the
matter as it was before. Seeing that the peculiar variety of ink
secreted by the cuttle fish is of a very valuable kind, it is somewhat
remarkable that no enterprising manufacturer has as yet taken the
matter in hand and established an aqueous farm for the breeding and
rearing of cuttle fish. Indian ink and sepia are both so valuable that
such an enterprise ought to pay handsome profits, and if the oyster
can be cultivated, why not the cuttle fish? It would, of course, be
necessary that the retaining walls of the gigantic aquarium indicated
should be impervious to the passing of cuttle fish even in their
earliest stage. Otherwise the proprietors would be liable very speedily
to be indicted as a nuisance by the lodging-house keepers and
owners of bathing machines of the nearest sea-side watering places.
But this could doubtless be effected, and then no argument could be
adduced that the cuttle fish should necessarily be a nuisance to their
neighbours that would not equally apply to the wild beasts at a
menagerie. In the latter case one occasionally breaks out and causes
consternation, and, possibly, damage, and even if an octopus should
do the same there could be no very valid ground for complaint. As
the squid when cooked furnishes a somewhat gelatinous food not
altogether dissimilar to calf’s head, it is probable that the flesh of the
larger varieties might be utilised for the manufacture of mock turtle,
and another source of revenue would, therefore, be open to their
breeders. It is clear from these remarks that the cuttle fish has not
hitherto received the careful consideration that it deserves, and the
dislike we feel for its form and habits has blinded us to the benefits
that might with culture and domestication be derived from it.
 THE BACILLUS.

H AD the learned Linnæus been informed that there existed a

creature of which he had taken no account, which exercised a
much larger influence upon the fortunes and happiness of man than
any of those which he so laboriously arranged and classified, he
would have smiled the smile of incredulity. But just as it is but within
the present century that mankind has awoke to the enormous power
and usefulness of steam and electricity, so it is only within the last
ten or fifteen years that he has attained to the knowledge of the
existence of the demon bacillus, who has sprung at a bound into the
position of man’s deadliest enemy. Secretiveness must be assigned
the first place among the characteristics of the bacillus. Since man
first appeared upon earth this scourge must have carried on its
deadly work, and heaped up a hecatomb of victims in comparison to
which those who have perished by war or by famine are but an
insignificant handful; and yet man has pursued his way in the
blindest ignorance of the very existence of his indefatigable enemy.
Even yet comparatively few people are aware of the personal
peculiarities of the bacillus, or could describe with any approach to
accuracy the difference between the allied tribes, each of which
represents some form or other of disease or death, and the scientific
men who are so actively busying themselves in counteracting its
work are very chary of describing its personal peculiarities. When
these are more generally understood it will probably lead to a
revolution in art. The artist of other days who wished to convey to
the beholder that the personage depicted was in imminent peril of
his life could find no better means of doing so than by placing
behind him a shadowy figure with a death’s head and skeleton arms
holding a dart. This childish representation can no longer be
tolerated, and the artist of the future will have only to depict
hovering over the principal figure a bacillus, and the beholder will at
once understand not only that death is impending, but will be able
to distinguish from the characteristics of the bacillus whether it will
take the form of consumption, typhoid, small-pox, or other disease.
This will be of vast utility in the painting of historical personages, as
no questions can arise centuries later as to the cause of their death,
the disease of which they died being clearly indicated by the
accompanying bacillus, which, of course, will in future be appended
to every posthumous portrait.
It is mortifying to human vanity to reflect that for some sixty
centuries, at the shortest computation, man has been taking all sorts
of pains to protect himself against minor dangers, in absolute
ignorance of the bacillus fiend in his midst. Against the wild beast
and the snake he has waged open warfare. He has covered himself
with armour to protect himself from the weapons of human foes. He
has furnished his ships with lifeboats, he has placed trap-doors in
the roofs of his houses to afford an escape in case of fire, and has
invented the safety lamp as a protection for those who work in
mines. He has muzzled the dog in order to escape the fabulously
remote risk of hydrophobia, and he has laid down strict regulations
to diminish the chances of his being blown up by explosives. He has
fenced himself in by sanitary regulations to preserve himself against
the evil effect of foul smells, and has flattered himself that by these
and many other precautions he has done what he could to ensure
for himself prolonged life. And yet all this time the bacillus has been
carrying on his work unsuspected, laughing, in whatever passes as
his sleeve, as he yearly sweeps away his tens of millions of victims.
It has, in fact, been a new and terrible illustration of the saying, “Out
of sight, out of mind.” Proud man, who slays the whale for its oil,
and the elephant for its ivory, has been slain by his invisible foe, the
bacillus; and, like a soldier brought down by a long range bullet, has
not even had the satisfaction of knowing who was his slayer.
 The
microsc
ope has
long
since
discove
red to
him the
existen
ce of
innume
rable
Small-pox Bacillus Cholera Bacillus
creatur
(Natural Size). (Natural Size).
es,
invisible
to the naked eye; he has learnt that
the water he drank teemed with
animated atoms; that many of the
rocks were composed solely of their
minute skeletons; that a layer of them
reposed on the depth of ocean; that
countless numbers of them were
borne with the floating dust in the air.
Some of these discoveries caused him
wonder and admiration, others a
certain sense of uneasiness and
disgust; but when he discovered that Typhoid Bacillus
neither he nor his ancestors had (Natural Size).
suffered any material inconvenience
from imbibing these countless hosts in
their drinks, or inhaling them in the atmosphere, he ceased to
trouble himself about them, and went on his way regardless of their
existence. The case has been wholly changed by the discovery of
the bacillus, and man stands aghast alike at the terribly destructive
and deadly nature of his foe, and at his own impotency to guard
himself against its attacks. His feelings resemble those of the solitary