Analytics, Data Science and A I:

Systems for Decision Support

Eleventh Edition

Chapter 6
Deep Learning and Cognitive
Computing

Slide in this Presentation Contain Hyperlinks.

JAWS users should be able to get a list of links by
using INSERT+F7

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Introduction
• Artificial intelligence (AI) is making a re-entrance into the
world of commuting and in our lives, this time far stronger
and much more promising than before.
• This unprecedented re-emergence and the new level of
expectations can largely be attributed to deep learning and
cognitive computing.
• These two latest buzzwords define the leading edge of AI
and machine learning today. Evolving out of the traditional
artificial neural networks (ANN), deep learning is changing
the very foundation of how machine learning works.

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
 Introduction
• Thanks to large collections of data and improved
computational resources, deep learning is making a profound
impact on how computers can discover complex patterns
using the self-extracted features from the data (as opposed to
a data scientist providing the feature vector to the learning
algorithm).
• Cognitive computing—first popularized by IBM Watson and its
success against the best human players in the game show
Jeopardy!—makes it possible to deal with a new class of
problems, the type of problems that are thought to be solvable
only by human ingenuity and creativity, ones that are
characterized by ambiguity and uncertainty. This chapter
covers the concepts, methods, and application of these two
cutting-edge AI technology trends.
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Learning Objectives (1 of 2)
6.1 Learn what deep learning is and how it is changing the
world of computing
6.2 Know the placement of deep learning within the broad
family of A I learning methods
6.3 Understand how traditional “shallow” artificial neural
networks (AN N) work
6.4 Become familiar with the development and learning
processes of AN N
6.5 Develop an understanding of the methods to shed light
into the AN N black box

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Learning Objectives (2 of 2)
6.6 Know the underlying concept and methods for deep
neural networks
6.7 Become familiar with different types of deep
learning methods
6.8 Understand how convolutional neural networks (CN N),
recurrent neural networks (RN N), and long short-
memory networks (LST M) work
6.9 Become familiar with the computer frameworks for
implementing deep learning
6.10 Know the foundational details about cognitive
Computing and I B M Watson

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Opening Vignette (1 of 4)
Fighting Fraud with Deep Learning and Artificial
Intelligence

• Business problem
– Danske Bank
– Predictive analytics in banking
 Fraud detection
• The solution
– Deep learning
• The results

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Opening Vignette (2 of 4)
Fighting Fraud with Deep Learning and Artificial
Intelligence

• Accuracy
– RO C curve
• D L vs traditional M L
techniques

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Opening Vignette (3 of 4)
Fighting Fraud with Deep Learning and Artificial
Intelligence
• A Generalized Framework for A I and Deep Learning–
Based Analytics Solutions

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Opening Vignette (4 of 4)
Discussion Questions:
1. What is fraud in banking?
2. What are the types of fraud that banking firms are facing
today?
3. What do you think are the implications of fraud on banks
and on their customers?
4. Compare the old and new methods for identifying and
mitigating fraud.
5. Why do you think deep learning methods provided better
prediction accuracy?
6. Discuss the trade-off between false positive and false
negative (type 1 and type 2 errors) within the context of
predicting fraudulent activities.
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Introduction to Deep Learning (1 of 3)
• Imaginative things in the SciFi movies are turning into
realities-tanks to A I and Machine Learning
– Siri, Google assistant, Alexa, Google home, …
• Deep learning is the newest member of the A I/Machine
Learning family
– Learn better than ever before
• The reason for Deep Learning superiority
– Automatic feature extraction and representation

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Introduction to Deep Learning (2 of 3)
• The placement of Deep Learning within the overarching
A I-based learning methods

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Introduction to Deep Learning (3 of 3)
• Differences between Classic Machine-Learning Methods
and Representation Learning/Deep Learning

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.1 (1 of 2)
Finding the Next Football Star with Artificial
Intelligence

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.1 (2 of 2)
Finding the Next Football Star with Artificial
Intelligence
Discussion Questions:
1. What does SciSports do? Look at its Web site for more
information.
2. How can advanced analytics help football teams?
3. What is the role of deep learning in solutions provided by
SciSports?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Basics of “Shallow” Learning (1 of 4)
• Artificial Neural Networks – abstractions of human brain
and its complex biological network of neurons
• Neurons = Processing Elements (PE s)
• Single-input and single-output neuron/P E

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Basics of “Shallow” Learning (2 of 4)
• Common transfer
(activation)
functions

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Basics of “Shallow” Learning (3 of 4)
• Typical multiple-input neuron with R individual inputs

n  w1,1 p1  w1,2 p2  w1,3 p3  ...  w1, R pR  b

n  Wp  b
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Basics of “Shallow” Learning (4 of 4)
• Typical Neural Network with three layers and eight
neurons

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.2
Gaming Companies Use Data Analytics to Score
Points with Players
Watch: Art of Analytics – The Sword

Discussion Questions:
1. What are the main challenges for gaming companies?
2. How can analytics help gaming companies stay
competitive?
3. What types of data can gaming companies obtain and
use for analytics?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Technology Insight 6.1 (1 of 3)
Elements of an Artificial Neural Network
• Processing element (P E)
• Network structure
– Hidden layer(s)
• Input
• Output
• Connection weights
• Summation function
• Transfer function

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Technology Insight 6.1 (2 of 3)
Elements of an Artificial Neural Network
• Neural Network with One Hidden Layer

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Technology Insight 6.1 (3 of 3)
Elements of an Artificial Neural Network
Summation Functions

Transfer Function

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.3
Artificial Intelligence Helps Protect Animals
from Extinction
Watch: WildTrack

Discussion Questions
1. What is WildTrack and what does it do?
2. How can advanced analytics help WildTrack?
3. What are the roles that deep learning plays in this
application case?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Process of Developing Neural-Network Based
Systems
• A process with constant
feedbacks for changes and
improvements!

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Learning Process in AN N
1. Compute temporary
outputs.
2. Compare outputs with
desired targets.
3. Adjust the weights and
repeat the process.

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Backpropagation for AN N Training
(1 of 2)
1. Initialize weights with random values
2. Read in the input vector and the desired output
3. Compute the actual output via the calculations
4. Compute the error.
5. Change the weights by working backward

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Backpropagation for AN N Training
(2 of 2)
• Illustration of the Overfitting in AN N

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Illuminating the Black Box of AN N
• AN N are typically known as black boxes
• Sensitivity analysis can shed light to the black-box

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.4 (1 of 2)
Sensitivity Analysis Reveals Injury Severity
Factors in Traffic Accidents
Discussion Questions:
1. How does sensitivity analysis shed light on the black box
(i.e., neural networks)?
2. Why would someone choose to use a black-box tool
such as neural networks over theoretically sound, mostly
transparent statistical tools like logistic regression?
3. In this case, how did neural networks and sensitivity
analysis help identify injury-severity factors in traffic
accidents?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.4 (2 of 2)
Sensitivity Analysis Reveals Injury Severity
Factors in Traffic Accidents
• Graphical representation of the sensitivity analysis results
for the eight binary AN N model configurations

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Deep Neural Networks (1 of 3)
• Deep: more hidden layers
• In addition to CP U, it also uses G P U
– With programming languages like CUD A by NVI DI A
• Needs large datasets
• Deep learning uses tensors as inputs
– Tensor: N-dimensional arrays
– Image representation with 3-D tensors
• There are different types and capabilities of Deep Neural
Networks for different tasks/purposes

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Deep Neural Networks (2 of 3)
Feedforward Multilayer Perceptron (M L P)-Type Deep
Networks
• Most common type of deep networks
• Vector Representation of the First Three Layers in a
Typical M L P Network.

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Deep Neural Networks (3 of 3)
• Impact of Random
Weights in Deep M L P
• The Effect of Pre-
training Network
Parameters on
Improving Results of a
Classification-Type
Deep Neural Network.
• More hidden layers
versus more neurons?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.5 (1 of 2)
Georgia DO T Variable Speed Limit Analytics
Help Solve Traffic Congestions
Discussion Questions:
1. What was the nature of the problems that G DO T was
trying to solve with data science?
2. What type of data do you think was used for the
analytics?
3. What were the data science metrics developed in this
pilot project? Can you think of other metrics that can be
used in this context?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.5 (2 of 2)
Georgia DO T Variable Speed Limit Analytics
Help Solve Traffic Congestions

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Convolutional “Deep” Neural
Networks
• Most popular M L P-base D L method
• Used for image/video processing, text recognition
• Has at least one convolution weight function
– Convolutional layer
• Convolutional layer  Polling (sub-sampling)
– Consolidating the large tensors into one with a smaller
size-and reducing the number of model parameters
while keeping only the important features
– There can be different types of polling layers

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Convolution Function
• Typical Convolutional Network Unit

• Convolution of a 2 x 2 Kernel by a 3 x 6 Input Matrix

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Image Processing Using CN N (1 of 3)
• ImageNet (http://www.image-net.org)
• Architecture of AlexNet, a CN N for Image Classification

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Image Processing Using CN N (2 of 3)
• Conceptual Representation of the Inception Feature in
GoogLeNet

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Image Processing Using CN N (3 of 3)
• Examples of Using the Google Lens
Figure 6.28 Two Examples of Using the Google Lens, a Service Based
on Convolutional Deep Networks for Image Recognition.

Source: ©2018 Google LL C, used with permission. Google and the Google logo are
registered trademarks of Google LL C.
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.6
From Image Recognition to Face Recognition
Discussion Questions:
1. What are the technical challenges in face recognition?
2. Beyond security and surveillance purposes, where else
do you think face recognition can be used?
3. What are the foreseeable social and cultural problems
with developing and using face recognition technology?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Text Processing Using CN N (1 of 2)
• Google word2vec project
– Word embeddings
• Typical Vector Representation of Word Embeddings in a
Two-Dimensional Space

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Text Processing Using CN N (2 of 2)
• CN N Architecture for Relation Extraction Task in Text
Mining

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Recurrent Neural Networks (RN N) &
Long Short-Term Memory (LST M)
(1 of 3)
• RN N designed to process sequential inputs
• Typical recurrent unit

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Recurrent Neural Networks (RN N) &
Long Short-Term Memory (LST M)
(2 of 3)
• LST M is a variant of RN N
– In a dynamic network, the weights are called the long-
term memory while the feedbacks role is the short-
term memory

Typical Long
Short-Term
Memory (L S T M)
Network
Architecture

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.7 (1 of 4)
Deliver Innovation by Understanding Customer
Sentiments

Discussion Questions:
1. Why do you think sentiment analysis is gaining overwhelming popularity?
2. How does sentiment analysis work? What does it produce?
3. In addition to the specific examples in this case, can you think of other
businesses and industries that can benefit from sentiment analysis? What is
common among the companies that can benefit greatly from sentiment
analysis?
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Recurrent Neural Networks (RN N) &
Long Short-Term Memory (LST M)
(3 of 3)
• LST M Network Applications

Example Indicating
the Close-to-
Human
Performance of the
Google Neural
Machine Translator
(G N M T)

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Computer Frameworks for
Implementation of Deep Learning
• Torch (http://www.torch.ch)
– M L with G P U
• Caffe (caffe.berkeleyvision.org)
– Facebook’s improved version (www.caffe2.ai)
• TensorFlow (www.tensorflow.org)
– Google - Tensor Processing Units (TPU s)
• Theano (deeplearning.net/software/theano)
– Deep Learning Group at the University of Montreal
• Keras (keras.io)
– Application Programming Interface
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Cognitive Computing (1 of 3)
• Systems that use mathematical models to emulate (or
partially simulate) the human cognition process to find
solutions to complex problems and situations where the
potential answers can be imprecise
• I B M Watson on Jeopardy!
• How does cognitive computing work?
– Adaptive • Data mining,
– • Pattern recognition,
Interactive
• Deep learning, and
– Iterative and stateful
• NL P
– Contextual – Mimic the way the
human brain works

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Conceptual Framework for Cognitive
Computing and Its Promises

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Cognitive Computing (2 of 3)
• How does cognitive computing differ from A I?
Table 6.3 Cognitive Computing versus Artificial Intelligence (A I).
Characteristic Cognitive Computing Artificial Intelligence (A I)

Technologies used • Machine learning • Machine learning

• Natural language processing • Natural language processing
• Neural networks • Neural networks
• Deep learning • Deep learning
• Text mining
• Sentiment analysis

Capabilities offered Simulate human thought processes Find hidden patterns in a variety of
to assist humans in finding solutions data sources to identify problems
to complex problems and provide potential Solutions
Purpose Augment human capability Automate complex processes by
acting like a human in certain
Situations
Industries Customer service, marketing, Manufacturing, finance,
healthcare, entertainment, service healthcare, banking, securities,
Sector retail, government

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Cognitive Computing (3 of 3)
• Cognitive computing use cases
– Development of smart and adaptive search engines
– Effective use of natural language processing
– Speech recognition
– Language translation
– Context-based sentiment analysis
– Face recognition and facial emotion detection
– Risk assessment and mitigation
– Fraud detection and mitigation
– Behavioral assessment and recommendations, …
• Cognitive analytics?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Cognitive Search
• Can handle a variety of data types
• Can contextualize the search space
• Employ advanced A I technologies.
• Enable developers to build enterprise-specific search
applications

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.7 (2 of 4)
I B M Watson Competes against the Best at
Jeopardy!
Discussion Questions:
1. In your opinion, what are the most unique features about
Watson?
2. In what other challenging games would you like to see
Watson compete against humans? Why?
3. What are the similarities and differences between
Watson’s and humans’ intelligence?

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.7 (3 of 4)
I B M Watson Competes against the Best at
Jeopardy!
A High-Level Depiction of Watson’s DeepQ A Architecture

Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved
Application Case 6.7 (4 of 4)
I B M Watson Competes against the Best at
Jeopardy!
• How does Watson do it?
– Massive parallelism
– Many experts
– Pervasive confidence estimation
– Integration of shallow and deep knowledge
• Future of Watson and Cognitive Computing
– Healthcare and medicine
– Security and Government
– Finance and Retail
– Education and Scientific Research, …
Copyright © 2020, 2015, 2011 Pearson Education, Inc. All Rights Reserved