We argue that the type of meta-learning proposed by Binz et al. generates models with low interpretability and falsifiability that have limited usefulness for neuroscience research. An alternative approach to meta-learning based on hyperparameter optimization obviates these concerns and can generate empirically testable hypotheses of biological computations.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1017/S0140525X24000219 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
The reinforcement metalearner as a biologically plausible meta-learning framework.
Behav Brain Sci
September 2024
Institute of Cognitive Sciences and Technologies, CNR, Rome, https://www.istc.cnr.it/en/people/massimo-silvetti.
We argue that the type of meta-learning proposed by Binz et al. generates models with low interpretability and falsifiability that have limited usefulness for neuroscience research. An alternative approach to meta-learning based on hyperparameter optimization obviates these concerns and can generate empirically testable hypotheses of biological computations.
View Article and Find Full Text PDFAn integrative effort: Bridging motivational intensity theory and recent neurocomputational and neuronal models of effort and control allocation.
Psychol Rev
July 2023
Behavioural Science Institute, Radboud University.
An increasing number of cognitive, neurobiological, and computational models have been proposed in the last decade, seeking to explain how humans allocate physical or cognitive effort. Most models share conceptual similarities with motivational intensity theory (MIT), an influential classic psychological theory of motivation. Yet, little effort has been made to integrate such models, which remain confined within the explanatory level for which they were developed, that is, psychological, computational, neurobiological, and neuronal.
View Article and Find Full Text PDFUnbiased Model-Agnostic Metalearning Algorithm for Learning Target-Driven Visual Navigation Policy.
Comput Intell Neurosci
December 2021
State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China.
As deep reinforcement learning methods have made great progress in the visual navigation field, metalearning-based algorithms are gaining more attention since they greatly improve the expansibility of moving agents. According to metatraining mechanism, typically an initial model is trained as a metalearner by existing navigation tasks and becomes well performed in new scenes through relatively few recursive trials. However, if a metalearner is overtrained on the former tasks, it may hardly achieve generalization on navigating in unfamiliar environments as the initial model turns out to be quite biased towards former ambient configuration.
View Article and Find Full Text PDFDeep reinforcement learning (DRL) recently has attained remarkable results in various domains, including games, robotics, and recommender system. Nevertheless, an urgent problem in the practical application of DRL is fast adaptation. To this end, this article proposes a new and versatile metalearning approach called fast task adaptation via metalearning (FTAML), which leverages the strengths of the model-based methods and gradient-based metalearning methods for training the initial parameters of the model, such that the model is able to efficiently master unseen tasks with a little amount of data from the tasks.
View Article and Find Full Text PDFDorsal anterior cingulate-brainstem ensemble as a reinforcement meta-learner.
PLoS Comput Biol
August 2018
Department of Experimental Psychology, Ghent University, Ghent, Belgium.
Optimal decision-making is based on integrating information from several dimensions of decisional space (e.g., reward expectation, cost estimation, effort exertion).
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!