Related Experiment Video
Updated: Jun 28, 2025

Multimodal Protocol for Assessing Metacognition and Self-Regulation in Adults with Learning Difficulties
Published on: September 27, 2020
The limitations of automatically generated curricula for continual learning.
Anna Kravchenko1, Rhodri Cusack2
1Faculty of Science, Radboud University, Nijmegen, The Netherlands.
Networks can learn their own training curricula, but may not always select the optimal path, especially in changing environments. This research explores self-directed learning in artificial neural networks for better generalization.
Area of Science:
- Artificial Intelligence
- Machine Learning
- Cognitive Science
Background:
- Artificial neural networks (ANNs) benefit from curriculum learning, where simpler concepts precede complex ones.
- Manual curriculum design is computationally expensive and lacks generalizability.
- Self-directed curriculum learning, where networks monitor their own progress, offers a promising alternative, especially for continual learning.
Purpose of the Study:
- To test the generalizability of self-directed curriculum learning in ANNs.
- To investigate the benefits of curricula and the challenges of continual learning.
- To evaluate task-switching metrics for optimal curriculum selection.
Main Methods:
- A proof-of-principle model was used to train an ANN on two sequential tasks.
- The study examined both static and continual learning conditions.
- Various task-switching metrics were tested to assess curriculum selection effectiveness.
Main Results:
- In some cases, ANNs were unable to autonomously select the optimal curriculum.
- The benefits of certain curricula were only apparent in hindsight, after training completion.
- Continuous learning presented a handicap, impacting curriculum effectiveness.
Conclusions:
- Self-directed curriculum learning is a valuable approach but requires careful consideration of task dynamics.
- The effectiveness of curriculum selection can be context-dependent and may require retrospective evaluation.
- Findings have implications for designing more adaptable ANNs and modeling human cognitive development.
Related Concept Videos
Purposive Learning
Cognitive Learning
E. C. Tolman's theory of purposive behavior emphasizes that much behavior is goal-directed. He argued that to understand behavior, we must look at the entire sequence of actions leading to a goal. For instance, high school students study hard, not just due to past reinforcement but also to achieve the goal of getting into a good college.
Tolman introduced the idea that behavior is influenced by...
Metacognition
Nursing Process for Patient and Caregiver Teaching I: Assessment and Diagnosis
It is critical to determine the patient's learning needs during the assessment. Determination of learning needs compounds data...
Introduction to Learning
In contrast to learned behaviors, unlearned behaviors such as crying, sexual...
Nursing Process for Patient and Caregiver Teaching III: Evaluation and Documentation
Nurses can use several methods to evaluate patient outcomes. For example, oral questions can assess cognitive learning,...

