Reinforcement Learning for Finance: Solve Problems in Finance with CNN and RNN Using the TensorFlow Library 1st Edition Samit Ahlawat pdf download
Reinforcement Learning for Finance: Solve Problems in Finance with CNN and RNN Using the TensorFlow Library 1st Edition Samit Ahlawat pdf download
https://ebookmeta.com/product/deep-reinforcement-learning-with-
python-with-pytorch-tensorflow-and-openai-gym-1st-edition-nimish-
sanghi-3/
https://ebookmeta.com/product/deep-reinforcement-learning-with-
python-with-pytorch-tensorflow-and-openai-gym-1st-edition-nimish-
sanghi/
https://ebookmeta.com/product/deep-reinforcement-learning-with-
python-with-pytorch-tensorflow-and-openai-gym-1st-edition-nimish-
sanghi-2/
https://ebookmeta.com/product/intersectionality-in-action-a-
guide-for-faculty-and-campus-leaders-for-creating-inclusive-
classrooms-and-institutions-1st-edition-peter-felten/
The Cambridge Handbook of Workplace Training and
Employee Development Kenneth G. Brown
https://ebookmeta.com/product/the-cambridge-handbook-of-
workplace-training-and-employee-development-kenneth-g-brown/
https://ebookmeta.com/product/all-the-beautiful-people-the-dread-
novels-1-1st-edition-jonathan-yanez/
https://ebookmeta.com/product/gini-inequality-index-methods-and-
applications-1st-edition-nitis-mukhopadhyay/
https://ebookmeta.com/product/berserker-bounty-hunter-1st-
edition-carlsson-roland/
https://ebookmeta.com/product/capitalism-and-the-emergence-of-
civic-equality-in-eighteenth-century-france-1st-edition-william-
h-sewell-jr/
Love of the Spirit The FGB 173 1st Edition Various
https://ebookmeta.com/product/love-of-the-spirit-the-fgb-173-1st-
edition-various/
Reinforcement
Learning for
Finance
Solve Problems in Finance with
CNN and RNN Using the
TensorFlow Library
—
Samit Ahlawat
Reinforcement
Learning for Finance
Solve Problems in Finance
with CNN and RNN Using
the TensorFlow Library
Samit Ahlawat
Reinforcement Learning for Finance: Solve Problems in Finance with CNN
and RNN Using the TensorFlow Library
Samit Ahlawat
Irvington, NJ, USA
Acknowledgments�������������������������������������������������������������������������������xi
Preface����������������������������������������������������������������������������������������������xiii
Introduction����������������������������������������������������������������������������������������xv
Chapter 1: Overview�����������������������������������������������������������������������������1
1.1 Methods for Training Neural Networks�����������������������������������������������������������2
1.2 Machine Learning in Finance��������������������������������������������������������������������������3
1.3 Structure of the Book��������������������������������������������������������������������������������������4
v
Table of Contents
vi
Table of Contents
vii
Table of Contents
Bibliography�������������������������������������������������������������������������������������403
Index�������������������������������������������������������������������������������������������������411
viii
About the Author
Samit Ahlawat is Senior Vice President in
Quantitative Research, Capital Modeling, at
JPMorgan Chase in New York, USA. In his
current role, he is responsible for building
trading strategies for asset management
and for building risk management models.
His research interests include artificial
intelligence, risk management, and
algorithmic trading strategies. He has given CQF Institute talks on artificial
intelligence, has authored several research papers in finance, and holds a
patent for facial recognition technology. In his spare time, he contributes
to open source code.
ix
Acknowledgments
I would like to express my heartfelt appreciation for my friends and
coworkers, in academia and the workplace, who encouraged me to write
this book.
xi
Preface
When I began using artificial intelligence tools in quantitative financial
research, I could not find a comprehensive introductory text focusing on
financial applications. Neural network libraries like TensorFlow, PyTorch,
and Caffe had made tremendous contributions in the rapid development,
testing, and deployment of deep neural networks, but I found most
applications restricted to computer science, computer vision, and robotics.
Having to use reinforcement learning algorithms in finance served as
another reminder of the paucity of texts in this field. Furthermore, I found
myself referring to scholarly articles and papers for mathematical proofs of
new reinforcement learning algorithms. This led me to write this book to
provide a one-stop resource for Python programmers to learn the theory
behind reinforcement learning, augmented with practical examples drawn
from the field of finance.
In practical applications, reinforcement learning draws upon deep
neural networks. To facilitate exposition of topics in reinforcement
learning and for continuity, this book also provides an introduction to
TensorFlow and covers neural network topics like convolutional neural
networks (CNNs) and recurrent neural networks (RNNs).
Finally, this book also introduces readers to writing modular, reusable,
and extensible reinforcement learning code. Having worked on developing
trading strategies using reinforcement learning and publishing papers,
I felt existing reinforcement learning libraries like TF-Agents are tightly
coupled with the underlying implementation framework and do not
xiii
Preface
xiv
Introduction
Reinforcement learning is a rapidly growing area of artificial intelligence
that involves an agent learning from past experience of rewards gained
by taking specific actions in certain states. The agent seeks to learn a
policy prescribing the optimum action in each state with the objective of
maximizing expected discounted future rewards. It is an unsupervised
learning technique where the agent learns the optimum policy by past
interactions with the environment. Supervised learning, by contrast, seeks
to learn the pattern of output corresponding to each state in training
data. It attempts to train the model parameters in order to get a close
correspondence between predicted and actual output for a given set of
inputs. This book outlines the theory behind reinforcement learning
and illustrates it with examples of implementations using TensorFlow.
The examples demonstrate the theory and implementation details of the
algorithms, supplemented with a discussion of corresponding APIs from
TensorFlow and examples drawn from quantitative finance. It guides
a reader familiar with Python programming from basic to advanced
understanding of reinforcement learning algorithms, coupled with a
comprehensive discussion on how to use state-of-the-art software libraries
to implement advanced algorithms in reinforcement learning.
Most applications of reinforcement learning have focused on robotics
or computer science tasks. By focusing on examples drawn from finance,
this book illustrates a spectrum of financial applications that can benefit
from reinforcement learning.
xv
CHAPTER 1
Overview
Deep neural networks have transformed virtually every scientific human
endeavor – from image recognition, medical imaging, robotics, and self-
driving cars to space exploration. The extent of transformation heralded
by neural networks is unrivaled in contemporary human history, judging
by the range of new products that leverage neural networks. Smartphones,
smartwatches, and digital assistants – to name a few – demonstrate the
promise of neural networks and signal their emergence as a mainstream
technology. The rapid development of artificial intelligence and machine
learning algorithms has coincided with increasing computational power,
enabling them to run rapidly. Keeping pace with new developments in
this field, various open source libraries implementing neural networks
have blossomed. Python has emerged as the lingua franca of the artificial
intelligence programming community. This book aims to equip Python-
proficient programmers with a comprehensive knowledge on how to use
the TensorFlow library for coding deep neural networks and reinforcement
learning algorithms effectively. It achieves this by providing detailed
mathematical proofs of key theorems, supplemented by implementation of
those algorithms to solve real-life problems.
Finance has been an early adopter of artificial intelligence algorithms
with the application of neural networks in designing trading strategies
as early as the 1980s. For example, White (1988) applied a simple
neural network to find nonlinear patterns in IBM stock price. However,
recent cutting-edge research on reinforcement learning has focused
2
Another Random Scribd Document
with Unrelated Content
constitutional powers, be aided in any of their relations; whether
laws are provided in all cases where they are wanting; whether
those provided are exactly what they should be; whether any abuses
take place in their administration, or in that of the public revenues;
whether the organization of the public agents or of the public force
is perfect in all its parts; in fine, whether anything can be done to
advance the general good, are questions within the limits of your
functions which will necessarily occupy your attention. In these and
other matters which you in your wisdom may propose for the good
of our country, you may count with assurance on my hearty co-
operation and faithful execution.