Summary: Humans excel at adapting to new situations, while machines often stumble. A new interdisciplinary study reveals that the root lies in how humans and AI approach “generalization,” the process of transferring knowledge to new problems.
Humans rely on abstraction and conceptual frameworks, whereas AI systems apply statistical or rule-based methods, each with limits. Bridging these approaches could pave the way for more flexible, human-centered AI systems that adapt better to the complexities of everyday life.
Key Facts
- Different Meanings: “Generalization” carries different definitions in cognitive science and AI research.
- Human vs. AI: Humans generalize through abstraction; AI uses domain-specific processes.
- Shared Framework: Researchers propose a unified framework to better align human and machine reasoning.
Source: Bielefeld University
How do humans manage to adapt to completely new situations and why do machines so often struggle with this?
This central question is explored by researchers from cognitive science and artificial intelligence (AI) in a joint article published in the journal “Nature Machine Intelligence”. Among the authors are Professor Dr. Barbara Hammer and Professor Dr. Benjamin Paaßen from Bielefeld University.
“If we want to integrate AI systems into everyday life, whether in medicine, transportation, or decision-making, we must understand how these systems handle the unknown,” says Barbara Hammer, head of the Machine Learning Group at Bielefeld University.
“Our study shows that machines generalize differently than humans and this is crucial for the success of future human–AI collaboration.”
Differences between humans and machines
The technical term “generalization” refers to the ability to draw meaningful conclusions about unknown situations from known information, that is, to flexibly apply knowledge to new problems.
In cognitive science, this often involves conceptual thinking and abstraction. In AI research, however, generalization serves as an umbrella term for a wide variety of processes: from machine learning beyond known data domains (“out-of-domain generalization”) to rule-based inference in symbolic systems, to so-called neuro-symbolic AI, which combines logic and neural networks.
“The biggest challenge is that ‘Generalization’ means completely different things for AI and humans,” explains Benjamin Paaßen, junior professor for Knowledge Representation and Machine Learning in Bielefeld.
“That is why it was important for us to develop a shared framework. Along three dimensions: What do we mean by generalization? How is it achieved? And how can it be evaluated?”
Significance for the future of AI
The publication is the result of interdisciplinary collaboration among more than 20 experts from internationally leading research institutions, including the universities of Bielefeld, Bamberg, Amsterdam, and Oxford. The project began with a joint workshop at the Leibniz Center for Informatics at Schloss Dagstuhl, co-organized by Barbara Hammer.
The project also highlights the importance of bridging cognitive science and AI research. Only through a deeper understanding of their differences and commonalities will it be possible to design AI systems that can better reflect and support human values and decision-making logics.
The research was conducted within the collaborative project SAIL – Sustainable Life-Cycle of Intelligent Socio-Technical Systems. SAIL investigates how AI can be designed to be sustainable, transparent, and human-centered throughout its entire life cycle.
Funding: The project is funded by the Ministry of Culture and Science of the State of North Rhine-Westphalia.
About this artificial intelligence research news
Author: Jörg Heeren
Source: Bielefeld University
Contact: Jörg Heeren – Bielefeld University
Image: The image is credited to Neuroscience News
Original Research: Closed access.
“Aligning generalization between humans and machines” by Barbara Hammer et al. Nature Machine Intelligence
Abstract
Aligning generalization between humans and machines
Recent advances in artificial intelligence (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 generalize. In cognitive science, human generalization commonly involves abstraction and concept learning.
By contrast, AI generalization encompasses out-of-domain generalization in machine learning, rule-based reasoning in symbolic AI, and abstraction in neurosymbolic AI.
Here we combine insights from AI and cognitive science to identify key commonalities and differences across three dimensions: notions of, methods for, and evaluation of generalization.
We map the different conceptualizations of generalization 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 support effective and cognitively supported alignment in human–AI teaming scenarios.
