The Digital Journey of Banking and Insurance, Volume II: Digitalization and Machine Learning (Digital Journey of Banking and Insurance, 2)

Acknowledgments
Introduction to Volume II—Digitalization and Machine Learning
Contents
Notes on Contributors
List of Figures
List of Tables
List of Equations
Use Cases
Use Case: Optimization of Regression Tests—Reduction of the Test Portfolio Through Representative Identification
	1 Introduction
		1.1 Initial Situation—By Example
		1.2 General Idea
		1.3 Structure of the Article
	2 Phases in the Approach
		2.1 Step I: Select Portfolio
		2.2 Step II: Configure and Train Classification Algorithm
		2.3 Step III: Determine Representatives and Select Reduced Test Portfolio
		2.4 Step IV: Apply Business Methods to the Reduced Test Portfolio
		2.5 Optimize Classification Algorithm
	3 Model Development and Application
		3.1 Model
			3.1.1 Cluster Algorithms
			3.1.2 Hard and Soft Clustering
		3.2 Experiment Results—German Credit
		3.3 Backtesting and Continuous Improvement
	4 Challenges in the Practical Application
	5 Summary
	Literature
Use Case—Nostro Accounts Match
	1 Introduction
		1.1 General Idea
		1.2 Structure of the Chapter
	2 Business Requirements
		2.1 Correspondent Bank System
	3 Match Analysis
		3.1 Data Collection/Generation
		3.2 Brute Force Search
		3.3 Data Analysis/String Similarity
		3.4 Model Training/Features Selection
		3.5 Example of a Match Result
	4 Challenges in the Practical Application—NLP
	5 Summary
	Literature
Use Case—Fraud Detection Using Machine Learning Techniques
	1 Introduction
	2 A Primer on Insurance Fraud
		2.1 Current State of Fraud Detection
	3 Use Case: Fraud Detection with Machine Learning on Car Insurance Claims
		3.1 Data Collection and Preparation
		3.2 Model Selection
			3.2.1 Supervised Model
			3.2.2 Unsupervised Model
			3.2.3 Ensemble Model
		3.3 Model Calibration
	4 Process Design
	5 Implementing an AI-Based Fraud Detection System
	6 Summary
	Literature
Use Case: NFR—HR Risk
	1 Introduction
	2 Resignation Risk: A Fraction of Non-financial Risk
		2.1 Taxonomy of the Risk Category HR Risk
		2.2 Position of the Resignation Risk in the HR Taxonomy
	3 Impact Graph and Indicators of Resignation Risk
		3.1 Impact Graph
		3.2 Indicators
		3.3 Relationship Between Impact Graphs and Indicators
	4 Measure Resignation Risk
		4.1 Scoring Resignation Indicators
		4.2 Merging Score Values
	5 Use of Machine Learning
		5.1 Training Machine Learning Algorithms: General Considerations and Preliminaries
		5.2 Drafting an Example Model: Resignation Risk
		5.3 Strategies to Find the Best Model and Enable Systematic Improvements
	6 Summary
	Literature
Sentiment Analysis for Reputational Risk Management
	1 Introduction
		1.1 Reputational Risk
		1.2 Sentiment Analysis
		1.3 General Approach
		1.4 Management of Reputational Risk
	2 Relevance Analysis
		2.1 Preprocessing
			2.1.1 Training the Word Embedding Model
			2.1.2 Querying the Vocabulary
			2.1.3 Interim Summary
		2.2 Topic Modeling
			2.2.1 Latent Dirichlet Allocation
			2.2.2 Relevance Probability
		2.3 Interim Summary and Outlook
	3 Sentiment Analysis
		3.1 Dictionary-Based Approach
			3.1.1 Preprocessing and Sentiment Score Calculation
			3.1.2 Challenges
			3.1.3 Augmented Dictionary Approach
		3.2 The NLP-Based Approach and Outlook
	Literature
Use Case: NFR—Using GraphDB for Impact Graphs
	1 Introduction
		1.1 Connected World
		1.2 Non-financial Risk Management
		1.3 Structure of the Chapter
	2 Example ESG Risk
		2.1 Business Setting
		2.2 Impact Graph
	3 Graph Functions
		3.1 Storage and Query in Graph Databases
		3.2 Descriptive Analysis Functions in the Impact Graph Context
			3.2.1 Predecessor and Successor Analysis
			3.2.2 Risk Driver (Sources)
			3.2.3 Impacted Nodes (Sinks)
			3.2.4 Graph Aggregation
	4 Traditional Oprisk Approach
		4.1 Abstract Example
	5 NFR Approach—Estimation of the Threat Situation
		5.1 General Idea
		5.2 Example HR Risk
			5.2.1 Change Activity Status
		5.3 Reporting
	6 Summary
	Literature
High-Performance Applications
Distributed Calculation Credit Portfolio Models
	1 Introduction
		1.1 Initial Situation
		1.2 Structure of the Chapter
	2 Credit Portfolio Models
		2.1 Overview of the Common Types of Credit Portfolio Risk Model
		2.2 Short Introduction to CreditMetrics
		2.3 Parallel Computing—Business-Driven Angle
			2.3.1 Natural Splitting and Block Building
			2.3.2 Hybrid Approach to Splitting and Block Building
		2.4 Comparison of the Splitting Variants
	3 Distributed Calculation
		3.1 Technical Environment
		3.2 Configuration in Sparklyr
		3.3 Data Repartition
		3.4 Experimental Result
	4 Summary
	Literature
BSDS: Balance Sheet Dynamics Simulator
	1 Introduction
		1.1 Financial Navigator and Other Applications
		1.2 Structure of the Chapter
	2 Model
		2.1 Recap: Elements of an Agent-Based Model
		2.2 Agent Types
		2.3 Environmental Definitions
		2.4 Sequence of Agent Actions and Messages
		2.5 Measures of the Matching Efficiency
		2.6 Limiting Conditions and Model Parameters
	3 Technical Implementation
		3.1 Technical Scenario
		3.2 Simudyne
		3.3 Hadoop Cloudera (Distributed Calculation)
	4 Ergodicity and Preliminary Results
		4.1 Initial States
		4.2 Simulation Runs
	5 Summary and Outlook
		5.1 Summary
		5.2 Reasonable Model Extensions
	Literature
Dynamic Dashboards
	1 Introduction
		1.1 Initial Situation
		1.2 Definition—Dynamic Dashboarding
		1.3 Structure of the Article
	2 Reporting Demands in a Dynamic World
		2.1 Why We Must Be Dynamic
		2.2 Risk Reporting
		2.3 Exploring Sensitivity and Combined Parameter Shifts
		2.4 Inverse Stress Test
	3 Use Cases for Dynamic Dashboards
		3.1 Use Case—Client Interface: Portfolio Manager Dashboard
			3.1.1 Purpose of the Portfolio Manager Dashboard
			3.1.2 Goal and Aim of the Asset Manager Client Dashboard
			3.1.3 Example Implementation in R Shiny Implementation in R Shiny
			3.1.4 Example Implementation in MicroStrategy
		3.2 Use Case—Impairment Projection
			3.2.1 Purpose of the Impairment Projection Dashboard
			3.2.2 Example Implementation in R Shiny
		3.3 Use Case—Intraday Liquidity Stress Test
		3.4 Use Case—Integrated Stress Test
	4 Tools for Dynamic Dashboards
		4.1 Microsoft—Power BI
		4.2 SAP—SAC
		4.3 MicroStrategy
		4.4 R Shiny
		4.5 Python—Plotly and Dash
	5 Summary
		5.1 Demands and Tools
		5.2 Outlook
	Literature
High-Performance Applications
	1 Introduction
	2 Hash Functions
	3 Password Cracking
		3.1 Attack Mechanisms
			3.1.1 Brute Force Attack
			3.1.2 Mask Attacks
		3.2 Benefit of Slow Hash Functions
	4 Summary
Quantum Computing
Post-quantum Secure Cryptographic Algorithms
	1 Introduction
		1.1 Quantum Technologies
		1.2 Benefits
		1.3 Threats
	2 Solutions
		2.1 Quantum Key Distribution
	3 Post-Quantum Cryptography
		3.1 Code-Based
		3.2 Hash-Based
		3.3 Lattice-Based
		3.4 Multivariate
		3.5 Isogeny-Based
	4 Summary
	Literature
Quantum Technologies
	1 Introduction
	2 The Basis of Quantum Technologies
	3 (Evolved) Quantum Technologies
		3.1 Quantum Communication
		3.2 Quantum Imaging
		3.3 Quantum Sensing and Metrology
		3.4 Quantum Simulation
		3.5 Quantum Computing
	4 Quantum Annealing
	5 Gate-Based Quantum Computing
		5.1 Foundations—Gate-Based Quantum Computing
		5.2 Algorithms—For Gate-Based Quantum Computing
		5.3 Entanglement Quantum Circuit
		5.4 Deutsch’s Quantum Algorithm
		5.5 Grover’s Search Algorithm
		5.6 Shor’s Factoring Algorithm
		5.7 The HHL Algorithm
	6 Topological Quantum Computing
	7 Summary
	Literature
Categorical Quantum Theory
	1 Introduction
	2 Basic Definitions
		2.1 Definition Category
		2.2 Definition Functor
		2.3 Definition Natural Transformation
		2.4 Definition Quantum Category
	3 Applications to Quantum Computing and Neural Networks
		3.1 Natural Language Processing
	4 Summary
	Literature
Process and  Process Optimization
Processes in a Digital Environment
	1 Introduction
	2 Process Optimization
	3 Three Core Components of Automation
		3.1 Business Process Engine
		3.2 Process Mining
		3.3 Robotic Process Automation
	4 Why They Exist
		4.1 Business Process Engines—Advantages and Challenges
		4.2 Process Mining—Advantages and Challenges
		4.3 Robotic Process Automation—Advantages and Challenges
		4.4 Component Classification
		4.5 Component Combination
	5 Summary and Outlook
	Literature
Process Mining
	1 Introduction
		1.1 Process Mining Brings Full Transparency to Business Processes
		1.2 How to Set Up Process Mining to Reach Full Transparency and Transform it Into Valuable Optimizations
		1.3 Structure of the Chapter
	2 The Process Mining Technology—Definition, Functionality, and Requirements
	3 The Steps of a Process Mining Project
		3.1 Setting the Objectives
			3.1.1 Define the Goals and Fields of Action
			3.1.2 Create an Open and Technology-Independent Solution Space
		3.2 Set Up an Evaluable Process Model
			3.2.1 Generate Event Logs
			3.2.2 Verify Process Model
		3.3 Analyze the Process and Gain Deep Process Knowledge
		3.4 Utilize the Process Knowledge Gained and Optimize the Process
			3.4.1 Adjustment of the General Process Flow
			3.4.2 Optimization of Individual Work Steps
		3.5 Measure the Effectiveness of the Optimizations Performed with Process Mining
	4 Organizational Aspects Within Process Mining Projects
		4.1 Roles and Responsibilities
		4.2 Change and Stakeholder Management
	5 Summary
	Literature
Hyperautomation (Automated Decision-Making as Part of RPA)
	1 Introduction
		1.1 Initial Situation
		1.2 Structure of the Article
	2 General Idea
	3 RPA and UiPATH
	4 Machine Learning/Data Analysis
		4.1 From Data Collection to Full Automation
		4.2 Natural Language Processing for Data Extraction
		4.3 Features Engineering and Model Training
			4.3.1 Label
			4.3.2 Features—Correlation
			4.3.3 Features—Object and Object.Type
			4.3.4 Model Training—Random Forests
	5 Practical Application: Machine Learning UiPATH
		5.1 Python Integration in UiPath
		5.2 Example in UiPath
	6 Summary
	Literature
RPA Use Case—“IFRS 9/SPPI”
	1 Introduction
		1.1 Initial Situation
		1.2 Structure of the Article
	2 Business Requirements in the Context of IFRS 9 and SPPI
	3 Challenges in Practice
	4 General Classification as well as Usability and Advantages of RPA for This Use Case
	5 Digitalization as a Driver for Increased Efficiency in Accounting and Controlling
	6 Summary
	Literature
Open Source
Open-Source Software
	1 Introduction
		1.1 Structure of the Chapter
	2 Open Source—The Community Idea
		2.1 A Short History of Open Source (Software)
		2.2 Further Examples
		2.3 Legal Variants
		2.4 Main Players
	3 The Success of Open Source (Software)
		3.1 Why is Open Source so Successful?
			3.1.1 Reduction in Initial Costs
			3.1.2 Agility
			3.1.3 No Dependencies on a Single Entity
		3.2 Where is Open Source Successful?
		3.3 Peer Production vs. Vendor-Driven Standard Software
		3.4 Community Editions
		3.5 Open-Source Maturity Model
	4 Summary
	Literature
Summary
	1 Ethics Do Matter
	2 Process Optimization and Automation
	3 Digital Transformation is an Irreversible Process
Index