Contents

pp v-vii

Contributors

pp viii-x

Acknowledgments

pp xi-xii

Part I - Introduction

pp 1-2

1 - Metacognitive AI

pp 3-22

Part II - Taxonomy of Metacognitive Approaches

pp 23-24

2 - An Architectural Approach to Metacognition

pp 25-43

3 - Metacognitive AI through Error Detection and Correction Rules

pp 44-55

4 - Mutual Trust in Human–AI Teams Relies on Metacognition

pp 56-80

Part III - Neuro-Symbolic Models in AI

pp 81-82

5 - Learning Where and When to Reason in Neurosymbolic Inference

pp 83-94

6 - Assessment of Competency of Learning Agents via Inference of Temporal Logic Formulas

pp 95-108

Part IV - Metacognition with LLMS

pp 109-110

7 - Metacognitive Intervention for Accountable LLMs through Sparsity

pp 111-120

8 - Metacognitive Insights into ChatGPT’s Arithmetic Reasoning

pp 121-132

Part V - Metacognition in Learning Agents

pp 133-134

9 - Uncertainty Quantification’s Role in Metacognition

pp 135-147

10 - The Role of Predictive Uncertainty and Diversity in Embodied AI and Robot Learning

pp 148-182

Part VI - Assured Machine Learning in High-Stakes Domains

pp 183-184

11 - Toward Certifiably Trustworthy Deep Learning at Scale

pp 185-211

12 - Metacognition with Neural Network Verification and Repair Using Veritex

pp 212-228

Part VII - Metacognition as a Solution to Handle Failure

pp 229-230

13 - Reasoning about Anomalous Object Interaction Using Plan Failure as a Metacognitive Trigger

pp 231-244

14 - Tractable Probabilistic Reasoning for Trustworthy AI

pp 245-256

Part VIII - Applications of Metacognitive AI

pp 257-258

15 - Robust and Compositional Concept Grounding for Image Generative AI

pp 259-263

16 - mLINK: Machine Learning Integration with Network and Knowledge

pp 264-275

17 - Military Applications of Artificial Intelligence Metacognition

pp 276-296