What is Artificial Intelligence?: A Conversation between an AI Engineer and a Humanities Researcher

✍ Scribed by Suman Gupta, Peter H Tu

• Publisher: WSPC (Europe)
• Year: 2020
• Tongue: English
• Leaves: 289
• Category: Library

✦ Synopsis

This book engages with the title question: what is artificial intelligence (AI)? Instead of reiterating received definitions or surveying the field from a disciplinary perspective, the question is engaged here by putting two standpoints into conversation. The standpoints are different in their disciplinary groundings — i.e. technology and the humanities — and also in their approaches — i.e. applied and conceptual. Peter is an AI engineer: his approach is in terms of how to make AI work. Suman is a humanities researcher: his approach is in terms of what people and academics mean when they say 'AI'.

A coherent argument, if not a consensus, develops by putting the two standpoints into conversation. The conversation is presented in 32 short chapters, in turn by Suman and Peter. There are two parts: Part 1, Questioning AI, and Part 2, AI and Government Policy. The first part covers issues such as the meaning of intelligence, automation, evolution, artificial and language. It outlines some of the processes through which these concepts may be technologically grounded as AI. The second part addresses policy considerations that underpin the development of AI and responds to the consequences. Themes taken up here include: rights and responsibilities; data usage and state-level strategies in the USA, UK and China; unemployment and policy futures.

Readership: General informed readers. Readers with knowledge of AI and its underpinning principles will be interested. In particular, those with a disciplinary background in computing and technological development, policy studies, current affairs, philosophy and discourse analysis. Appropriate for undergraduate students and upwards, or those whose interest is more conceptual than the general user of technology.

✦ Table of Contents

Contents
About the Authors
Part I: Overview of Highway Location and Alignment Optimization Problem
Chapter 1: Introduction
1.1 Background
1.2 Highway Alignment Optimization Problems
1.3 Organization of the Monograph
Chapter 2: Highway Cost and Constraints
2.1 Initial Construction Costs
2.2 Highway Maintenance Costs
2.3 User Costs
2.4 Environmental and Socio-economic Impacts
2.5 Highway Constraints
2.5.1 Geometric Constraints
2.5.2 Geographical and Environmental Constraints
Chapter 3: Review of Artificial Intelligence-based Models for Optimizing Highway Location and Alignment Design
3.1 Genetic Algorithms-based Optimization Models
3.1.1 Highway Alignment Optimization (HAO) Models by the UMD Research Team
3.1.2 Other Models that Use Genetic Algorithms for Optimizing Highway Alignments
3.2 Other Artificial Intelligence-based Optimization Models
3.2.1 Swarm Intelligence-basedModels
3.2.2 Neighborhood Search Heuristic
3.3 Other Methods Tried with AIs for Optimizing Highway Alignments
3.3.1 Distance Transform
3.3.2 Multi-objective Optimization
3.3.3 Multiple Path Optimization
3.3.4 Other Objectives Considered
3.3.5 Summary
Part II: Highway Alignment Optimization with Genetic Algorithms
Chapter 4: Modeling Highway Alignments with GAs
4.1 Representation of Highway Alignments with GAs
4.2 Modeling Horizontal Alignments
4.2.1 Horizontal Alignment Generation Procedure
4.3 Modeling Vertical Alignments
4.4 Modeling Highway Endpoints
4.4.1 Determination of Highway Endpoints
4.4.2 Highway Endpoint Determination Procedure
4.4.3 GA Operators for Endpoint Generation
4.5 Modeling Highway Structures
4.5.1 Small Highway Bridges for Grade Separation
4.5.2 Structures for Highway Junction Points with Existing Roads
Chapter 5: Highway Alignment Optimization Formulation
5.1 Objective Function
5.2 Constraints
5.3 Integrating GAs and Geographic Information System
5.4 Highway Cost Formulation
5.4.1 Highway Agency Cost
5.4.1.1 Length-dependent cost (CL)
5.4.1.2 Right-of-way cost (CR)
5.4.1.3 Earthwork cost (CE)
5.4.1.4 Structure cost (CS)
5.4.1.5 Maintenance cost (CM)
5.4.2 User Cost
5.4.2.1 Travel time cost (CT )
5.4.2.2 Vehicle operating cost (CV )
5.4.2.3 Accident cost (CA)
5.4.3 Penalty and Environmental Costs
5.4.3.1 Penalty cost (CP)
5.4.3.2 Environmental cost (CEN)
5.4.4 Life-cycle Cost
Chapter 6: Constraint Handling for Evolutionary Algorithms
6.1 Direct Constraint Handling
6.1.1 Elimination Method
6.1.2 Repairing Method
6.1.3 Preserving Method
6.1.4 Decoding Method
6.1.5 Locating the Boundary of the Feasible Regions
6.2 Indirect Constraint Handling (Penalty Approaches)
6.2.1 Death Penalty
6.2.2 Static Penalty
6.2.3 Dynamic Penalty
6.2.4 Adaptive Penalty
6.3 Handling Infeasible Solutions of GA-based Highway Alignment Optimization
Chapter 7: Highway Alignment Optimization Through Feasible Gates
7.1 Research Motivation of Feasible Gates
7.2 Feasible Gates for Horizontal Alignments
7.2.1 User-defined Horizontal Feasible Bounds
7.2.2 Representation of Horizontal Feasible Gates
7.2.3 Horizontal Feasible Gate Determination Procedure
7.2.4 User-defined Constraints for Guiding Feasible Alignments
7.3 Feasible Gates for Vertical Alignments
7.3.1 Road Elevation Determination Procedure
7.4 Example Study
7.5 Summary
Chapter 8: Prescreening and Repairing in Highway Alignment Optimization
8.1 ResearchMotivation of Prescreening and Repairing
8.2 Prescreening and Repairing for Alignments Violating Design Constraints
8.2.1 P&R Basic Concept
8.2.2 Determination of Design Constraint Violations
8.3 Example Study
8.4 Summary
Part III: Optimizing Simple Highway Networks: An Extension of Genetic Algorithms-based Highway Alignment Optimization
Chapter 9: Overview of Discrete Network Design Problems
9.1 Bi-level Discrete Network Design Problems
9.1.1 Upper-level DNDP
9.1.2 Lower-level DNDP
9.2 Comparison of Highway Alignment Optimization and Discrete Network Design Problems
Chapter 10: Bi-level Highway Alignment Optimization within a Small Highway Network
10.1 Bi-level HAO Concept
10.2 Upper Level of Bi-level HAO
10.2.1 Highway Agency Cost for Bi-level HAO
10.2.2 Highway User Cost for Bi-level HAO
10.2.3 Penalty and Environmental Costs
10.3 Lower Level of Bi-level HAO
10.3.1 User and System Optimal Traffic Assignment Problems
10.3.2 Determination of Traffic Reassignment
10.4 Bi-level HAO Model Structure
10.5 Inputs Required for Lower Level of Bi-level HAO
10.5.1 Highway Network and O/D Trip Matrix
10.5.2 Travel Time Functions
10.6 Summary and FutureWork
Chapter 11: Bi-level HAO Model Application Example
11.1 Example Description
11.2 Optimized Alternatives
11.3 Summary
Part IV: Highway Alignment Optimization Model Applications and Extensions
Chapter 12: HAO Model Application in Maryland Brookeville Bypass Project
12.1 Project Description
12.2 Data and Application Procedure
12.2.1 Horizontal Map Digitization
12.2.2 Vertical Map Digitization
12.2.3 Tradeoffs in Map Representation for Environmental Issues
12.2.4 Description of Model Inputs and Outputs
12.3 Optimization Results
12.3.1 Optimized Alignments with Different Numbers of PI’s
12.3.2 Goodness Test
12.4 Sensitivity of Optimized Alignments to OtherMajor Input Parameters
12.4.1 Sensitivity to the Model’s Objective Function
12.4.2 Sensitivity to Design Speed
12.4.3 Sensitivity to Elevation Resolution
12.4.4 Sensitivity to Cross-section Spacing
Chapter 13: HAO Model Application in US 220 Project in Maryland
13.1 Project Description
13.2 Projection Preparation
13.2.1 Spatial Data
13.2.2 Project Segmentation
13.2.3 Geometric Design Specification
13.3 Optimization Results
Chapter 14: HAO Model Application to Maryland ICC Project
14.1 Project Description
14.1.1 Overview of the ICC Study
14.1.2 Description of HAO Model Application to ICC Project
14.2 Input Data Preparation
14.2.1 Road Network
14.2.2 Traffic Information
14.2.3 GIS Map Preparation
14.2.4 Important Input Parameters
14.3 Optimization Results
14.3.1 Determination of Traffic Reassignments
14.3.2 Optimized Alignments
14.3.3 Goodness Test
Chapter 15: Related Developments and Extensions
15.1 Related Developments
15.2 Extensions
Appendix A: Notation Used in the Monograph
Appendix B: Traffic Inputs to theHAOModel for the ICC Case Study
References
Index