Engineering Artificially Intelligent Systems: A Systems Engineering Approach to Realizing Synergistic Capabilities (Lecture Notes in Computer Science)

Preface
Organization
Addendum
Contents
Motivations for and Initiatives on AI Engineering
	1 Introduction: Urgencies for AI Engineering
	2 AIE Status Summary
	3 Overview of the LNCS Book
	4 Chapter Synopses
	5 A Tabular Review
	References
Architecting Information Acquisition to Satisfy Competing Goals
	1 Motivation for IBSM: Requirements and Constraints
	2 Design Considerations
		2.1 Human-On-The-Loop Vice Human-In-The-Loop
		2.2 Partition the System into Orthogonal Components
		2.3 Probabilistic World Model
		2.4 Partitioned Components Are Interconnected by Bidirectional Interfaces
		2.5 Data, Information, and Knowledge
		2.6 Mission Value and Competing, Interdependent Goals
	3 IBSM Architecture
		3.1 Goal Lattice
		3.2 Situation Information Expected Value Network
		3.3 Information Instantiator
		3.4 Applicable Function Table
		3.5 Sensor Scheduler
		3.6 Communications Manager
	4 IBSM Operational Narrative
	5 Machine Learning in IBSM
	6 Conclusion
	References
Trusted Entropy-Based Information Maneuverability for AI Information Systems Engineering
	1 Introduction
	2 Information Power and Maneuverability
	3 Machine Learning Combat Power Study
		3.1 Example 1: Data Transfer
		3.2 Example 2: Data Trust
	4 Discussion
	5 Conclusions
	References
BioSecure Digital Twin: Manufacturing Innovation and Cybersecurity Resilience
	1 Introduction
	2 Problem
	3 Applying Cyber-Informed Engineering (CIE) to Effectively Develop and Deploy a Digital Twin for BioPharma Manufacturing
	4 Biopharma System Level Security Gaps
	5 Digital Twin R&D, Testbeds, and Benefits
	6 Alignment to U.S. Government Cybersecurity Goals for Critical Infrastructure
	7 Project Impact, Outcomes, Dissemination
	8 Responding to the Current Coronavirus and Preparing for the Next Potential Pandemic
	9 Conclusion
	References
Finding the Path Toward Design of Synergistic Human-Centric Complex Systems
	1 Introduction
	2 State of the Art in Systems Engineering for Complex Systems and Human Integration
		2.1 Systems Engineering and Role of Artificial Intelligence Techniques for Complex Systems
		2.2 Challenges of Human Interaction with Complex Systems
		2.3 Perspectives on Synergy
	3 Engineering Synergy Between Human and Complex Systems
		3.1 Systems Engineering for Human-Centric Design
		3.2 Complex System View from a Human Prespective
		3.3 Synergistic Design of Human-Centered Complex Systems
	4 Conclusions and Path Forward
	References
Agent Team Action, Brownian Motion and Gambler's Ruin
	1 Introduction
		1.1 Agent/Team Behavior in Brief
	2 Stochastic Processes
	3 Gambler's Ruin
		3.1 Some Analysis
		3.2 Effects of Parameters on Absorbing Probabilities
	4 Team Behavior
		4.1 Drift
		4.2 Diffusion
	5 Discussion
	6 Conclusion and Future Work
	References
How Deep Learning Model Architecture and Software Stack Impacts Training Performance in the Cloud
	1 Introduction
	2 Set-Up
		2.1 Benchmarks
		2.2 Changes to the Reference Implementations
		2.3 Benchmarking Software Stack
		2.4 Infrastructure
	3 Benchmark Results
		3.1 GPU Instances
		3.2 Performance Implications of GPU Drivers and CUDA Libraries
	4 Conclusion
	References
How Interdependence Explains the World of Teamwork
	1 Introduction
	2 Interdependence as an Integrative Framework for Teamwork
		2.1 The Challenge
		2.2 What Criteria Define Joint Activity That Is Teamwork?
		2.3 What Makes Teamwork a Special Kind of Joint-Work Activity?
		2.4 Where Does Interdependence Come from?
		2.5 Where Does the Skeletal Plan Come from?
		2.6 How Does Teamwork Relate to Taskwork?
		2.7 What Kinds of Support Are Needed to Facilitate Teamwork?
		2.8 How Does Teamwork Continually Adjust Over Time?
	3 How Does the Framework Help Us to Understand the Broader World of Human-Machine Teamwork?
		3.1 How Does Interdependence Relate to Situation Awareness?
		3.2 How Does Interdependence Relate to the “Levels of Automation” and “Adjustable Autonomy” Approaches?
		3.3 How Are Trust Decisions Made?
	4 Discussion
		4.1 How Does an Interdependence-Centric Framework Help Researchers Generalize Results?
		4.2 How Does an Interdependence-Centric Framework Help Explain Experimental Results?
	5 Conclusion
	References
Designing Interactive Machine Learning Systems for GIS Applications
	1 Introduction
	2 Interactive Machine Learning in Practice
		2.1 Airfield Change Detection (ACD)
		2.2 Geographic Region Annotation Interface Toolkit (GRAIT)
		2.3 Digital Map Editing (SmartMaps)
	3 Design Considerations in Interactive Machine Learning
		3.1 Uncertainty Models
		3.2 Constraining the Problem
		3.3 User Preferences
		3.4 Cognitive Feedback
	4 Challenges in Automated Map Labeling
	5 Conclusion
	References
Faithful Post-hoc Explanation of Recommendation Using Optimally Selected Features
	1 Introduction
	2 Related Work
	3 LIME Algorithm
	4 Proposed Method
		4.1 Item Recommendation for Target User
		4.2 Interpretation of Recommended Item by LIME
		4.3 Generation of Explanation
		4.4 Providing an Explanation
	5 Experiments
		5.1 Dataset
		5.2 Recommendation Algorithms
		5.3 Baseline Methods
		5.4 Objective Evaluation - Recall of Explanation Model
		5.5 Subjective Human Evaluation - Explanation Evaluation
	6 Conclusion
	References
Risk Reduction for Autonomous Systems
	1 Introduction
		1.1 Society’s Current Acceptance of Control Systems
		1.2 More Recent Problems
		1.3 Ethical Guidelines
	2 Critical Questions
	3 Which Laws Apply to Machine-Made Decisions and Consequent Actions?
		3.1 Unmanned Air Vehicles (UAVs) and Drones
		3.2 Lethal Autonomous Weapon Systems (LAWS)
	4 Are Theories of Human Decision-Making Applicable to Non-human Systems?
		4.1 Models of Human Decision-Making
		4.2 Common Architecture for Humans and Automated Decisions and Actions
	5 Minimising Risks in a Non-deterministic System’s Supply Chain?
	6 Minimising Risks in a Non-deterministic System’s Supply Chain?
		6.1 Regulator
		6.2 Marketing and Design Specifier Roles
		6.3 System Specifier for Supply Chain
		6.4 Integrator and Design Authority
		6.5 Manufacturer
		6.6 Owner/Maintainer/Driver
		6.7 Updates to Requirements
	7 Discussion and Conclusions
	References
Agile Systems Engineering in Building Complex AI Systems
	1 Introduction
	2 Consumer Analytics Scenario and Agile Process
	3 Agile and Scrum: The State-of-the-Art
	4 Scrum for Machine Learning: A Necessity
		4.1 Model-Based Analytics
		4.2 Agility in Analytics System Development
		4.3 Agility in Machine Learning Model Development
		4.4 Machine and Deep Learning for NLP
	5 Agile ETL and System Implementation
	6 Validation and Feedback in Agile Process
	7 Conclusions
	References
Platforms for Assessing Relationships: Trust with Near Ecologically-Valid Risk, and Team Interaction
	1 Introduction
	2 Requirements of Platforms to Allow for the Study of Human-Machine Teaming
		2.1 Requirement 1: Human Perception of Risk and Vulnerability
		2.2 Requirement 2: Machines and Humans as Equally Critical to the Mission
		2.3 Requirement 3: Ability to Manipulate Team Structure and Roles
		2.4 Requirement 4: Allow for Objective Measurement of Trust
		2.5 Requirement 5: Leverage Human Expectations and Experience
	3 Bot Behavior
	4 Pilot Test of PAR-TNER
		4.1 Participants
		4.2 Equipment and Setup
		4.3 Procedures
		4.4 Questionnaire
	5 Results
	6 Conclusions
	References
Design Principles for AI-Assisted Attention Aware Systems in Human-in-the-Loop Safety Critical Applications
	1 Introduction
	2 Related Work
	3 Human-Machine Challenges
		3.1 Attention and Situational Awareness
		3.2 Classifier Interaction
	4 Functional Architecture
	5 System Design Principles
		5.1 Determine Operator Focus of Attention
		5.2 Track Operator Target Awareness
		5.3 Track Operator Target Intention
		5.4 Track Operator Target Interaction
		5.5 Balance Operator Attention
		5.6 Intervene Operator
	6 Discussion
	7 Conclusions
	References
Interdependence and Vulnerability in Systems: A Review of Theory for Autonomous Human-Machine Teams
	1 Introduction
		1.1 What is the Problem?
		1.2 Interdependence Defined
		1.3 The Effects of Interdependence
		1.4 Positive Effects of Interdependence
		1.5 Negative Effects of Interdependence
	2 Interdependence and Vulnerability
	3 Convergence Processes
	4 Conclusion
	References
Principles of an Accurate Decision and Sense-Making for Virtual Minds
	1 Cognition and Virtual Minds
		1.1 Cognitive Contexts and Virtual Collective Mind
		1.2 Acausal Algebras and Cognition
		1.3 Virtual Awareness Versus Virtual Sensations
		1.4 Virtual Matter and Virtual Mind
		1.5 Cross Constructions and Hopf Algebras
	2 Hypercomplex Representation of Decision
		2.1 Decision, Subjectivity and Conflictuality
		2.2 Double Complex Representations of Causality
		2.3 Quantum Superposition of States
	3 Quantum Physics and Semantic of Sense-Making
		3.1 Bilinearity and Sense-Making
		3.2 Triality and Incompatibility
		3.3 Sense-Making and Subtractive Arithmetic
		3.4 Quasi Additive Structures and Anti-superposition
		3.5 Cross Constructions and Clockwise Orientations
		3.6 Sketch of the Decision Process
		3.7 Square Annihilation of Counter Models
	4 Conclusion
	References
Author Index