What AI Can Do: Strengths and Limitations of Artificial Intelligence

Cover
Half Title
Title Page
Copyright Page
Table of Contents
Editors
Contributors
Section I: Nature and Culture of the Algorithm
1. AI Ethics as a Form of Research Ethics
1.1 Introduction
1.2 Technology as a Means
1.2.1 Values – Metaethical Considerations
1.2.2 Technology and Values
1.2.3 The Ambivalence of Technology
1.3 How to Ask Ethical Questions in AI Research
1.3.1 Bad Scientific Practice I: Methodological Aspects
1.3.2 Bad Scientific Practice II: Ethical Aspects
1.4 Discussion
1.5 Conclusion and Outlook
References
2. Going through the Challenges of Artificial Intelligence: Gray Eminences, Algocracy, Automated Unconsciousness
2.1 The Intervention of Gray Eminences in the Digital World
2.2 Algocracy as the Sterilization of Politics
2.3 Discussion: The Ethics of AI in the Mirror of its Weaknesses
2.4 Who Must Adapt to Whom? A Matter of Strategic Enveloping
2.5 Conclusion
References
3. AI, Ethics, and Coloniality: A Feminist Critique
3.1 Introduction
3.2 Structural Violence and the Decolonial Turn
3.2.1 Decolonial Feminism
3.3 Decoloniality as Praxis: A Brief Genealogy
3.4 Power and Knowledge: The Coloniality of Technology, Gender, Race/Ethnicity, and Nature
3.5 Decolonizing AI: Geopolitics and Body Politics
3.6 AI and Ethics: Between Geopolitics and Body Politics
3.6.1 Geopolitics
3.6.2 Body Politics
3.7 Conclusion
References
4. A Cultural Vision of Algorithms: Agency, Practices, and Resistance in Latin America
4.1 The Judges' Question
4.2 Studies on Technology and Culture in Latin America
4.3 Algorithmic Cultures
4.4 Agency and Play
4.5 Collectives and Resistance
4.6 Who Knows How Algorithms Work?
References
5. From Deepfake to Deeptruth: Toward a Technological Resignification with Social and Activist Uses
5.1 Introduction
5.2 The Use of Synthetic Media for Denunciation and Social Activism
5.3 A Theoretical Framework to Study Content Created with Synthetic Media
5.3.1 Techno-Aesthetics
5.3.2 Post-Truth
5.3.3 Discourse
5.3.4 Criticism
5.4 Case Study: Javier Valdez's Synthetic Video
5.4.1 Technical Procedures
5.4.2 Ethical and Legal Transparency
5.4.3 Discourse
5.4.4 Context
5.4.5 Critical Sense
5.5 Conclusions: The Resignification from Deepfake to Deeptruth
References
6. The Neurocomputational Becoming of Intelligence: Philosophical Challenges
6.1 The Shadow of Phenomenology on Contemporary Philosophy
6.2 The Philosophical Reductionism of Immediacy
6.3 The Neurocognitive Model of the Mind
6.4 The Threat of the Transhumanism
6.5 Posthuman Transcendental Structure
6.6 AI Frame Problems
6.7 Non-Propositional Metacognition
6.8 AI Historical and Social Context
6.9 Discussion: What Would it Mean to Survive for an AI?
6.10 Conclusions
References
Section II: Knowledge Areas Facing Al
7. A Cluster Analysis of Academic Performance in Higher Education through Self-Organizing Maps
7.1 Introduction
7.2 What Are Self-Organizing Maps (SOMs)?
7.2.1 Structure of SOMs
7.2.2 Competitive Learning of SOMs
7.2.3 Designing of SOMs
7.2.4 SOMs Quality Measures
7.3 How SOMs Can be Used to Analyze Academic Performance in Higher Education?
7.3.1 Description of the Participants
7.3.2 Description of the Variables
7.3.3 Data Analysis
7.3.3.1 Preparation of the Data
7.3.3.2 Designing the SOM Model
7.3.4 Patterns between Prior Academic Achievement, SES, and Academic Performance in Higher Education
7.3.5 Grouping Students Based on their Prior Academic Achievement, SES, and Academic Performance
7.4 What are the Takeaway Messages of this Chapter?
References
8. Artificial Intelligence as a Way to Improve Educational Practices
8.1 Introduction
8.2 What is a Predictive Algorithm Based on Artificial Intelligence (AI)?
8.3 AI Applied in Education to Improve the Learning Experience
8.4 Limitations in the Educational Field for AI Prediction Algorithms
8.5 AI-Based Adaptive Learning: Case Study
8.6 Final Lecture
References
9. Using AI for Educational Research in Multimodal Learning Analytics
9.1 Introduction
9.2 Learning Analytics
9.3 Multimodal Learning Analytics
9.4 Applications of Artificial Intelligence in MMLA
9.4.1 Analyzing Body Postures in Oral Presentations
9.4.2 Assessing Spoken Interactions in Remote Active Learning Environments
9.4.3 Analyzing Physical Environment Factors in Distance Learning
9.5 Limitations and Future of AI in MMLA
9.6 Closing Remarks
References
10. Artificial Intelligence in Biomedical Research and Clinical Practice
10.1 Introduction to Artificial Intelligence in Healthcare
10.2 Digital Health, Person-Centered Attention Paradigm, and Artificial Intelligence
10.3 Regulatory Affairs Implementing Artificial Intelligence in Medical Devices
10.4 Radiomics and AI, the Most Extended Application
10.5 New Drugs' Development and AI
10.6 Challenges to Developing and Implementing AI in Medical Areas
10.7 Conclusion
References
11. The Dark Side of Smart Cities
11.1 Introduction
11.2 What Kinds of Urban Planning Problems is AI Being Used on?
11.3 What Kinds of Problems and Resistances Have Been Called Out to the Implementation of AI in Urban Planning?
11.4 What Can We Learn from the AI that is Already Implemented in Urban Planning?
11.5 Conclusions
References
12. The Control of Violence between the Machine and the Human
12.1 Introduction
12.2 Human and Algorithmic Implications for the State Decision to Control Violence
12.3 The Colonizing Private Sector of Violence Control
12.4 Human Function and Mechanical Function in the Penal Decision and in the Algorithmic Interpretation
12.5 Conclusions
References
13. AI in Music: Implications and Consequences of Technology Supporting Creativity
13.1 Why Puccini for Artificial Intelligence? An Introduction
13.2 These Hopeful and Intelligent Machines
13.2.1 Knowledge Representation and Reasoning
13.2.2 Machine Learning
13.3 Music-Making with AI
13.4 Solving Puccini
13.4.1 Solving with KRR
13.4.2 Solving with ML
13.5 Implications and Consequences for Art Music Endeavors
13.5.1 Acceptance
13.5.2 Efficiency
13.5.3 Trustworthy AI
13.5.4 Search Space
13.6 Encore: Another Muse for Music Creation
Acknowledgments
References
Section III: Future Scenarios and Implications for the Application of AI
14. Classification Machine Learning Applications for Energy Management Systems in Distribution Systems to Diminish CO[sub(2)] Emissions
14.1 Introduction
14.2 Energy Management Systems
14.3 What is Machine Learning?
14.4 Driving Energy Management Systems with Machine Learning
14.5 Classification and Trends of Machine Learning Applications in Energy Management Systems
14.5.1 Costumer-Oriented ML Applications
14.5.2 System-Oriented ML Applications
14.5.3 Comparison between ML Strategies in the EMS Context
14.6 Challenges of Machine Learning Driving Energy Management Systems
14.7 Wrap-Up
Acknowledgement(s)
Disclosure Statement
Funding
References
15. Artificial Intelligence for Construction 4.0: Changing the Paradigms of Construction
15.1 Introduction
15.2 Construction 4.0
15.3 Information Management in Construction 4.0
15.4 On-Site Surveillance
15.5 Quality Control
15.6 Design Optimization
15.7 Other Artificial Intelligence Technologies
15.7.1 Smart Robotics
15.7.2 Digital Twins
15.7.3 Artificial Intelligence of Things (AIoT)
15.7.4 Blockchain
15.7.5 Architectural Design
15.7.6 Document Management
15.8 Conclusions
References
16. A Novel Deep Learning Structure for Detecting Human Activity and Clothing Insulation
16.1 Classification Algorithms for Computer Vision
16.1.1 Convolutional Neural Networks
16.1.1.1 Convolutional Layer
16.1.1.2 Pooling and Flattening Layer
16.1.1.3 Activation Functions
16.1.1.4 Fully Connected Layer and Loss Functions
16.1.2 CNN Regularization
16.1.3 Object Recognizers
16.1.3.1 R-CNN
16.1.3.2 Fast R-CNN
16.1.3.3 Faster R-CNN
16.1.3.4 Mask R-CNN
16.1.3.5 YOLO
16.1.3.6 YOLOv2
16.1.3.7 YOLOv3
16.1.3.8 YOLOv4
16.1.3.9 YOLOv5
16.1.3.10 PP-YOLO
16.1.3.11 YOLOX
16.1.3.12 YOLOv6
16.1.3.13 YOLOv7
16.1.4 Recurrent Neural Networks (RNNs)
16.1.4.1 LSTM Networks
16.2 Clothing Classification
16.2.1 Methods Used
16.2.2 Clothing Detection Implementation
16.3 Activity Recognition
16.3.1 Deep Learning Model for HAR
16.3.1.1 Data Acquisition
16.3.1.2 Pre-Processing
16.3.1.3 Feature Extraction and Classification
16.4 Case Study: Thermal Comfort
16.4.1 Clothing Level and Metabolic Rate Estimations
16.5 Discussion
References
17. Building Predictive Models to Efficiently Generate New Nanomaterials with Antimicrobial Activity
17.1 Introduction
17.1.1 The Process to Generate a New Material with Certain Desired Properties
17.1.1.1 Synthesis Route
17.1.1.2 Nanomaterial Characterization
17.1.1.3 Antibacterial Assays
17.2 Artificial Intelligence: The Tool for Complex Problems
17.2.1 Machine Learning Algorithms
17.2.2 Supervised Machine Learning and Nanomaterial Development
17.2.2.1 Phase 1: Understanding the Problem
17.2.2.2 Phase 2: Understanding the Data
17.2.2.3 Phase 3: Prepare the Data
17.2.3 Combine Data from Different Sources
17.2.3.1 Phase 4: Building the Model
17.2.3.2 Phase 5: Error Analysis
17.2.4 Define Evaluation Criteria
17.2.4.1 Phase 6: Deployment
17.3 Our Application
17.4 The Future
References
18. Neural Networks for an Automated Screening System to Detect Anomalies in Retina Images
18.1 Introduction
18.2 Artificial Intelligence and Machine Learning
18.3 Artificial Neural Networks for Retinal Abnormalities Detection
18.3.1 Data
18.3.2 Preprocessing and Feature Extraction
18.3.3 Classification
18.4 The Future
References
19. Artificial Intelligence for Mental Health: A Review of AI Solutions and Their Future
19.1 Introduction
19.2 Mental Health Overview
19.3 Part 1. Cognitive Behavioral Therapy (CBT)
19.4 Part 2. How Does Artificial Intelligence Provide Mental Health Support?
19.4.1 Mental Health Support through Natural Language Processing
19.4.2 Mental Health Support through Computer Vision
19.4.3 Mental Health Support through AI Robotics
19.5 Part 3. Current Applications of AI
19.5.1 Chatbots
19.5.1.1 Wysa – Mental Health Chatbot
19.5.2 Virtual AI Therapists
19.5.3 Robot AI Therapists
19.5.3.1 Commercial Therapy Robots
19.5.3.2 Humanoid Therapy Robots
19.6 Concerns and Ethical Implications of AI in Mental Health
19.7 Conclusions
References
20. What AI Can Do for Neuroscience: Understanding How the Brain Represents Word Meanings
20.1 Introduction
20.2 Overview of the Key Ideas
20.3 Models of Semantic Representation
20.3.1 Distributional Semantic Models (DSMs)
20.3.2 Embodied Semantic Model
20.4 Case Study: The CEREBRA Model
20.5 Hybrid Approach to Ground NLP Applications
20.6 Challenges for Building More Brain-Like AIs
20.7 Conclusion
References
21. Adversarial Robustness on Artificial Intelligence
21.1 Recent Advances in Artificial Intelligence
21.2 Model Vulnerabilities
21.2.1 Adversarial Robustness
21.3 Limitations of Current Models
21.4 Designing Robust Models for Evasion Attacks
21.4.1 Adversarial Examples
21.4.2 Perturbations Sets
21.4.3 Training Robust Models
21.5 Recent Advances on Robust Models
21.6 Discussion
References
Index