Related Experiment Video
Updated: Jun 25, 2025

Aversive Associative Learning and Memory Formation by Pairing Two Chemicals in Caenorhabditis elegans
Published on: June 23, 2022
Associative Learning of an Unnormalized Successor Representation
1Department of Psychology, New York University Abu Dhabi, Abu Dhabi, United Arab Emirates niels.verosky@gmail.com.
Hebbian learning, a form of associative learning, is an unnormalized version of the successor representation. These two models converge when states are equally frequent, offering new insights into temporal association and representation learning.
Area of Science:
- Cognitive Science
- Computational Neuroscience
- Machine Learning
Background:
- The successor representation (SR) links temporal associations in models like the temporal context model.
- Prior research indicates the SR's broad applicability across spatial, visual, and abstract relational tasks.
Purpose of the Study:
- To investigate the deeper relationship between the successor representation and purely associative learning.
- To demonstrate that Hebbian temporal associations represent an unnormalized form of the SR.
Main Methods:
- Theoretical analysis comparing Hebbian learning and the successor representation.
- Examination of conditions under which the two representations converge or correlate.
Main Results:
- Hebbian temporal associations are identified as an unnormalized form of the successor representation.
- Convergence occurs when all states are equally frequent.
- High correlation is observed even with non-uniform state distributions.
Conclusions:
- Hebbian learning and the successor representation are fundamentally linked.
- The findings deepen our understanding of temporal association and representation learning in artificial and biological systems.
Related Concept Videos
Associative Learning
Classical conditioning, also known...
Higher Mental Functions of Brain: Learning and Memory
Woodward–Hoffmann Selection Rules and Microscopic Reversibility
Long-term Potentiation
Observational Learning
Real-World Application of Classical Conditioning
Higher-order, or second-order, conditioning occurs when a neutral stimulus becomes associated with an already established conditioned stimulus through repeated pairings. For instance, if a dog has been...

