Learning environment-specific learning rates DOI Creative Commons
Jonas Simoens, Tom Verguts, Senne Braem

et al.

PLoS Computational Biology, Journal Year: 2024, Volume and Issue: 20(3), P. e1011978 - e1011978

Published: March 22, 2024

People often have to switch back and forth between different environments that come with problems volatilities. While volatile require fast learning (i.e., high rates), stable call for lower rates. Previous studies shown people adapt their rates, but it remains unclear whether they can also learn about environment-specific instantaneously retrieve them when revisiting environments. Here, using optimality simulations hierarchical Bayesian analyses across three experiments, we show use rates switching two We even observe a signature of these the volatility both is suddenly same. conclude humans flexibly associate environments, offering important insights developing theories meta-learning context-specific control.

Language: Английский

Interpretable SHAP Model Combining Meta-learning and Vision Transformer for Lithology Classification Using Limited and Unbalanced Drilling Data in Well Logging DOI
Youzhuang Sun, Shanchen Pang, Zhiyuan Zhao

et al.

Natural Resources Research, Journal Year: 2024, Volume and Issue: 33(6), P. 2545 - 2565

Published: Aug. 19, 2024

Language: Английский

Citations

5
AutomaticAI – A hybrid approach for automatic artificial intelligence algorithm selection and hyperparameter tuning DOI
Zoltan Czako, Gheorghe Sebestyen, Anca Hângan

et al.

Expert Systems with Applications, Journal Year: 2021, Volume and Issue: 182, P. 115225 - 115225

Published: May 26, 2021

Language: Английский

Citations

30
Organizing memories for generalization in complementary learning systems DOI Open Access
Weinan Sun,

Madhu Advani,

Nelson Spruston

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2021, Volume and Issue: unknown

Published: Oct. 15, 2021

ABSTRACT Memorization and generalization are complementary cognitive processes that jointly promote adaptive behavior. For example, animals should memorize a safe route to water source generalize features allow them find new sources, without expecting paths exactly resemble previous ones. Memory aids by allowing the brain extract general patterns from specific instances were spread across time, such as when humans progressively build semantic knowledge episodic memories. This process depends on neural mechanisms of systems consolidation, whereby hippocampal-neocortical interactions gradually construct neocortical memory traces consolidating hippocampal precursors. However, recent data suggest consolidation only applies subset memories; why certain memories consolidate more than others remains unclear. Here we introduce novel network formalization highlights an overlooked tension between transfer generalization, resolve this postulating it generalization. We specifically show unregulated can be detrimental in unpredictable environments, whereas optimizing for generates high-fidelity, dual-system supporting both theory generalization-optimized produces transfers some components neocortex leaves dependent hippocampus. It thus provides normative principle reconceptualizing numerous puzzling observations field insight into how behavior benefits learning specialized memorization

Language: Английский

Citations

28
Statistical Learning in Vision DOI Open Access
József Fiser, Gábor Lengyel

Annual Review of Vision Science, Journal Year: 2022, Volume and Issue: 8(1), P. 265 - 290

Published: June 21, 2022

Vision and learning have long been considered to be two areas of research linked only distantly. However, recent developments in vision changed the conceptual definition from a signal-evaluating process goal-oriented interpreting process, this shift binds learning, together with resulting internal representations, intimately vision. In review, we consider various types (perceptual, statistical, rule/abstract) associated past decades argue that they represent differently specialized versions fundamental which must captured its entirety when applied complex visual processes. We show why generalized version statistical can provide appropriate setup for such unified treatment vision, what computational framework best accommodates kind plausible neural scheme could feasibly implement framework. Finally, list challenges field faces fulfilling promise being right vehicle advancing our understanding entirety.

Language: Английский

Citations

19
Learning environment-specific learning rates DOI Creative Commons
Jonas Simoens, Tom Verguts, Senne Braem

et al.

PLoS Computational Biology, Journal Year: 2024, Volume and Issue: 20(3), P. e1011978 - e1011978

Published: March 22, 2024

People often have to switch back and forth between different environments that come with problems volatilities. While volatile require fast learning (i.e., high rates), stable call for lower rates. Previous studies shown people adapt their rates, but it remains unclear whether they can also learn about environment-specific instantaneously retrieve them when revisiting environments. Here, using optimality simulations hierarchical Bayesian analyses across three experiments, we show use rates switching two We even observe a signature of these the volatility both is suddenly same. conclude humans flexibly associate environments, offering important insights developing theories meta-learning context-specific control.

Language: Английский

Citations

4