mlsanityData to 7 Oct 2026

Position: Artificial Intelligence Needs Meta Intelligence -- the Case for Metacognitive AI

arXiv 2605.15567 ↗First posted 15 May 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

This position paper argues for metacognition as a general design principle for creating more accurate, secure, and efficient AI. The metacognitive solution involves systems monitoring their own states and judiciously allocating resources depending on each problem instance's difficulty or cost of mistakes. Drawing inspiration both from past work on resource-rational AI and from well-documented metacognitive strategies in psychology and cognitive science, we identify specific challenges in embedding these strategies into AI design and highlight open theoretical and implementation problems. We showcase these principles through a tangible example of improved learning efficiency, effectiveness, and security in a Federated Learning (FL) case study. We show how these principles can be translated into practice with a novel software framework developed specifically to allow the community to design, deploy, and experiment with metacognition-enabled AI applications.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 36 of its 102 references that are in the map. First posted on arXiv 15 May 2026.

Builds on

36 of 102 references.

References · 36

  1. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2020Scaling Laws for Neural Language Modelscited by 6,411
  4. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  5. 2017On Calibration of Modern Neural Networkscited by 4,860
  6. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  7. 2022Emergent Abilities of Large Language Modelscited by 2,483
  8. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  9. 2019Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shiftcited by 1,274
  10. 2021Generalized Out-of-Distribution Detection: A Surveycited by 796
  11. 2018Byzantine-Robust Distributed Learning: Towards Optimal Statistical Ratescited by 698
  12. 2018The Hidden Vulnerability of Distributed Learning in Byzantiumcited by 324
24 more
  1. 2019Robust Aggregation for Federated Learningcited by 314
  2. 2020SPINN: Synergistic Progressive Inference of Neural Networks over Device and Cloudcited by 109
  3. 2023The Metacognitive Demands and Opportunities of Generative AIcited by 85
  4. 2024Metacognitive Capabilities of LLMs: An Exploration in Mathematical Problem Solvingcited by 82
  5. 2021Fast and Slow Learning of Recurrent Independent Mechanismscited by 36
  6. 2022Trading Off Privacy, Utility and Efficiency in Federated Learningcited by 33
  7. 2017Failing to Learn: Autonomously Identifying Perception Failures for Self-driving Carscited by 28
  8. 2013Reasoning About Beliefs and Actions Under Computational Resource Constraintscited by 18
  9. 2016Introspective Perception: Learning to Predict Failures in Vision Systemscited by 17
  10. 2022Few-Shot Calibration of Set Predictors via Meta-Learned Cross-Validation-Based Conformal Predictioncited by 17
  11. 2024Imagining and building wise machines: The centrality of AI metacognitioncited by 12
  12. 2024Metacognitive AI: Framework and the Case for a Neurosymbolic Approachcited by 9
  13. 2023Rule-Based Error Detection and Correction to Operationalize Movement Trajectory Classificationcited by 8
  14. 2024Error Detection and Constraint Recovery in Hierarchical Multi-Label Classification without Prior Knowledgecited by 7
  15. 2024Failures in Perspective-taking of Multimodal AI Systemscited by 6
  16. 2015Towards Machine Waldcited by 4
  17. 2022Gluing Neural Networks Symbolically Through Hyperdimensional Computingcited by 4
  18. 2025Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel Environmentscited by 4
  19. 2024Metal Price Spike Prediction via a Neurosymbolic Ensemble Approachcited by 3
  20. 2025Multiple Distribution Shift -- Aerial (MDS-A): A Dataset for Test-Time Error Detection and Model Adaptationcited by 2
  21. 2018Seeing in the dark with recurrent convolutional neural networkscited by 1
  22. 2025Probabilistic Foundations for Metacognition via Hybrid-AIcited by 1
  23. 2025Error Detection and Correction for Interpretable Mathematics in Large Language Modelscited by 1
  24. 2025You Only Need Your Transformer 25% of the Time: Meaning-First Execution for Eliminating Unnecessary Inferencecited by 1

Built on by

Not cited yet.