Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Aligning Generalisation Between Humans and Machines

Filip Ilievski, Barbara Hammer, Frank van Harmelen, Benjamin Paassen, Sascha Saralajew, Ute Schmid, Michael Biehl, Marianna Bolognesi, Xin Luna Dong, Kiril Gashteovski, Pascal Hitzler, Giuseppe Marra, Pasquale Minervini, Martin Mundt, Axel-Cyrille Ngonga Ngomo, Alessandro Oltramari, Gabriella Pasi, Zeynep G. Saribatur, Luciano Serafini, John Shawe-Taylor, Vered Shwartz, Gabriella Skitalinskaya, Clemens Stachl, Gido M. van de Ven, Thomas Villmann

TL;DR

The paper argues that humans and machines generalise in fundamentally different ways, which can impede safe and effective human-AI teaming. It proposes a three-dimensional framework—notions, methods, and evaluation of generalisation—to map and compare human and AI generalisation across both cognitive science and AI literature. It inventories three machine-generalisation families (statistical, knowledge-informed, and instance-based) and discusses how each aligns with or diverges from human generalisation, highlighting evaluation challenges and the need for interdisciplinarity. It also outlines emerging directions, including generalisation theory for foundation models, neurosymbolic approaches, continual learning, and evaluation practices, to foster cognitively aligned, robust human-AI collaboration.

Abstract

Recent advances in AI -- including generative approaches -- have resulted in technology that can support humans in scientific discovery and forming decisions, but may also disrupt democracies and target individuals. The responsible use of AI and its participation in human-AI teams increasingly shows the need for AI alignment, that is, to make AI systems act according to our preferences. A crucial yet often overlooked aspect of these interactions is the different ways in which humans and machines generalise. In cognitive science, human generalisation commonly involves abstraction and concept learning. In contrast, AI generalisation encompasses out-of-domain generalisation in machine learning, rule-based reasoning in symbolic AI, and abstraction in neurosymbolic AI. In this perspective paper, we combine insights from AI and cognitive science to identify key commonalities and differences across three dimensions: notions of, methods for, and evaluation of generalisation. We map the different conceptualisations of generalisation in AI and cognitive science along these three dimensions and consider their role for alignment in human-AI teaming. This results in interdisciplinary challenges across AI and cognitive science that must be tackled to provide a foundation for effective and cognitively supported alignment in human-AI teaming scenarios.

Aligning Generalisation Between Humans and Machines

TL;DR

The paper argues that humans and machines generalise in fundamentally different ways, which can impede safe and effective human-AI teaming. It proposes a three-dimensional framework—notions, methods, and evaluation of generalisation—to map and compare human and AI generalisation across both cognitive science and AI literature. It inventories three machine-generalisation families (statistical, knowledge-informed, and instance-based) and discusses how each aligns with or diverges from human generalisation, highlighting evaluation challenges and the need for interdisciplinarity. It also outlines emerging directions, including generalisation theory for foundation models, neurosymbolic approaches, continual learning, and evaluation practices, to foster cognitively aligned, robust human-AI collaboration.

Abstract

Recent advances in AI -- including generative approaches -- have resulted in technology that can support humans in scientific discovery and forming decisions, but may also disrupt democracies and target individuals. The responsible use of AI and its participation in human-AI teams increasingly shows the need for AI alignment, that is, to make AI systems act according to our preferences. A crucial yet often overlooked aspect of these interactions is the different ways in which humans and machines generalise. In cognitive science, human generalisation commonly involves abstraction and concept learning. In contrast, AI generalisation encompasses out-of-domain generalisation in machine learning, rule-based reasoning in symbolic AI, and abstraction in neurosymbolic AI. In this perspective paper, we combine insights from AI and cognitive science to identify key commonalities and differences across three dimensions: notions of, methods for, and evaluation of generalisation. We map the different conceptualisations of generalisation in AI and cognitive science along these three dimensions and consider their role for alignment in human-AI teaming. This results in interdisciplinary challenges across AI and cognitive science that must be tackled to provide a foundation for effective and cognitively supported alignment in human-AI teaming scenarios.

Paper Structure

This paper contains 23 sections, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Parallels in Generalisation by Humans and Machines
  3. Notions of Generalisation
  4. Generalisation as a process
  5. Generalisation as a product
  6. Generalisation as an operator
  7. Alignment of human and machine notions of generalisation
  8. Machine Methods for Generalisation
  9. Statistical generalisation methods in AI
  10. Knowledge-informed generalisation methods in AI
  11. Instance-based translation in AI
  12. Aligning machine generalisation methods and human expectations
  13. Evaluation of Generalisation
  14. Measuring distributional shifts
  15. Determining under- and overgeneralisation
  16. ...and 8 more sections

Figures (2)

  • Figure 1: Comparison of the strengths of humans and statistical ML machines, illustrating the complementary ways they generalise in human-AI teaming scenarios. Humans excel at compositionality, common sense, abstraction from a few examples, and robustness. Statistical ML excels at large-scale data and inference efficiency, inference correctness, handling data complexity, and the universality of approximation. Overgeneralisation biases remain challenging for both humans and machines. Collaborative and explainable mechanisms are key to achieving alignment in human-AI teaming. See Table \ref{['tab:summary']} for a complete overview of the properties of machine methods, including instance-based and analytical machines.
  • Figure 2: Illustrative examples of human generalisation and its inspiration of rule-based (top), example-based (middle), and statistical ML approaches (bottom).